Tasmota 🚀 - Sonoff T1 3CH US Ghost Switching

Please use weblog 4 for more debug information. When you use that, it will tell you in the console the source of the command that triggers your relay. It will be shown as SRC:

ascillato on 12 Mar 2019

ok, will enable weblog 4 and send result when random switching happens again, thanks

Bloodyagent on 12 Mar 2019

Hi, I have same issue. At the first time I thought was some interference on 315MHz receiver. So, seeing data sheet of SYN470R, I lift pin 11, in this way receiver is in shutdown. After that I put pin 10(DO line) at GND to avoid floating pin. After that I tried all my 7 sonoff without glass cover for 2 days and worked fine! no ghost switch.
So, I remount glass cover and ghost switching returns!
So, if you try to long press one of touch button you hear the first beep and after that quickly 2 beep. In original firmware with this operation you can pair with phone, with tasmota seems do nothing fortunately.
So, the problem is touch sensitivity. I try to find a solution but the are no name on IC between touch button and IC. I hope that someone know its name. I do reverse engineering on that and I draw schematic connection and I'm searching on web a datasheet with pin compatibility.
If is useful I can share my schematic on this part of circuit. If I have time tomorrow I'll see some signal coming from button with oscilloscope because maybe it's possibile to adjust sensitivity adjusting input RC network.

beppeserafini on 12 Mar 2019

Ok, just had random switching again, here is output

20:53:35 APP: Button2 immediate
20:53:35 SRC: Button
20:53:35 MQT: stat/TestBench/RESULT = {"POWER2":"ON"}
20:53:35 MQT: stat/TestBench/POWER2 = ON
20:53:35 CFG: Saved to flash at F4, Count 136, Bytes 3584

Thanks

Bloodyagent on 12 Mar 2019

Hi, I have same issue. At the first time I thought was some interference on 315MHz receiver. So, seeing data sheet of SYN470R, I lift pin 11, in this way receiver is in shutdown. After that I put pin 10(DO line) at GND to avoid floating pin. After that I tried all my 7 sonoff without glass cover for 2 days and worked fine! no ghost switch.
So, I remount glass cover and ghost switching returns!
So, if you try to long press one of touch button you hear the first beep and after that quickly 2 beep. In original firmware with this operation you can pair with phone, with tasmota seems do nothing fortunately.
So, the problem is touch sensitivity. I try to find a solution but the are no name on IC between touch button and IC. I hope that someone know its name. I do reverse engineering on that and I draw schematic connection and I'm searching on web a datasheet with pin compatibility.
If is useful I can share my schematic on this part of circuit. If I have time tomorrow I'll see some signal coming from button with oscilloscope because maybe it's possible to adjust sensitivity adjusting input RC network.

Thanks for this, I suspected this and did remove the cover at some point, but didn't really take note, think you may be right in that it didn't switch while cover was off. I didn't pursue this because its only happening with the 3CH versions and wasn't an issue with original firmware. I may just disable the middle button and turn the 3CH into a 2CH to see if this fixes it.

Bloodyagent on 12 Mar 2019

Just not sure why, if its a sensitivity problem, is it a non issue with original firmware and with the 2ch and 1ch versions?

Bloodyagent on 12 Mar 2019

I have also another sonoff t1 and I use it for 1 month without any ghost switch using last Tasmota firmware with the last core version. When I received others sonoff I saw some disconnection and looking on the web I installed last Tasmota with 2.30 core and disconnections disappears. Some minutes ago sonoff that worked for a month had a ghost switch. So, now I flash last core version again and I want to see if ghost switch disapper. I’ll inform you

beppeserafini on 12 Mar 2019

Yes I moved to core 2.3.0 because of disconnects and haven't tested on 2.4.0

I've removed the white rectangular pads on switches....still getting ghost switching!

Let me know. Thanks

Bloodyagent on 12 Mar 2019

Did you remove all rectangular pads and did you have ghost switch without glass?

beppeserafini on 13 Mar 2019

Removed pads. Still ghost switching with cover on. Cover off no switching pads or no pads

Bloodyagent on 13 Mar 2019

Ok, just had random switching again, here is output

20:53:35 APP: Button2 immediate
20:53:35 SRC: Button
20:53:35 MQT: stat/TestBench/RESULT = {"POWER2":"ON"}
20:53:35 MQT: stat/TestBench/POWER2 = ON
20:53:35 CFG: Saved to flash at F4, Count 136, Bytes 3584

Thanks

Bloodyagent on 13 Mar 2019

Ok so have compared the boards off the 1 ch and 3 ch switches. They are identical in every way. There are no missing components on the 1ch as opposed to the 3 channel. Only difference I can see is relay count. So why is it the that when I configure the 3ch as a 2ch the bottom touch button is disabled and not the middle similar to the 2ch. Was hoping to make the 3ch exactly as the 2ch to see if removing the middle touch switch would stop the ghost switching.

Bloodyagent on 13 Mar 2019

Can you see if in 1ch and 2ch versione there are some mark on ic touch component? In 3ch there is not mark on this.
Thank you!
Inked40544554-ff83d8d6-6028-11e8-8cf4-c951b75897d9_LI

beppeserafini on 13 Mar 2019

Chips blank ok all boards. I Googled and there are a few 16pin touch chips out there. Looks to be a 4 ch touch chip

Bloodyagent on 13 Mar 2019

Today I had one ghost switch on button 2 with core 2.42. Do you have ghost switch only on middle button or also others? Now I try to put a paper sheet between cover and plastic pad. I'll see...
I found some chip but no one that I found is compatible with schematic I drawn by inspection(I think is correct).
If you found component with 16 pin that can fit please tell me. Or if you write down a list I can check if they can be fit.
1WqXCpu2TceY3FYOoPuz0A_thumb_95b

beppeserafini on 13 Mar 2019

I'm using about 20 Sonoff T1-2-3 in my home since one year with Tasmota and I found that ghost switching only happen with Mqtt retain and/or RF interference.
In my case I solved disabling Mqtt retain, also try wifi config 4 or 5.

ToRvaLDz on 14 Mar 2019

Hi @ToRvaLDz

This is on the test bench without mqtt configured. Only happens on my 3CH units, all of them.

I erased the flash, reflashed with 6.4.1 - 2.3.0, joined to wireless, chose sonoff T1 3CH from the config menu, restarted and left them on the testbench with no load and no mqtt, they all ghost switching.

Bloodyagent on 14 Mar 2019

haven't tried wifi config 4 or 5. Will try...thanks

Bloodyagent on 14 Mar 2019

@Bloodyagent without Mqtt configured I had another issue, the touch is not responding every time ( look at thie video https://youtu.be/TUrKbRcctf8 I opened another issue for that #5461), also I noticed that if mqtt is not available it restart itself after some time, this could probably make a relay toggle.

ToRvaLDz on 14 Mar 2019

@ToRvaLDz ghost switching you had like us? So, beep and then other two beep quickly?

beppeserafini on 14 Mar 2019

@beppeserafini The beep indicate the long press, the only thing I understood is that the touch is not responding while tasmota is looking for an mqtt connection if not available and that if wifi is not connected tasmota do a reboot causing a relay toggle. For ghost switching, that is a button/relay that activate/deactivate itself, the only cause I found are: mqtt retain or wificonfig != 3 or 4.

ToRvaLDz on 14 Mar 2019

OK, so this is strange. I remove the 1k resistor from the middle touchpad thereby totally disabling the middle sensor touchpad. The ghost switching is now only happening on the middle sensor!!

Bloodyagent on 14 Mar 2019

So its definitely interference of some type. Maybe a lowpass filter on the input pins of the IC? Not sure how the capacitive touch IC works? I see there is a space right next to the 1k resistor for a capacitor, not sure if this is meant to be an RC filter of some type?
tempFileForShare_20190314-190132

Bloodyagent on 14 Mar 2019

This looks close?
Edit. This is the chip for sure.

http://en.wincomtech.com/prod_view.aspx?TypeId=65&Id=162&FId=t3:65:3

Bloodyagent on 14 Mar 2019

Just a stupid question: are you trying it with a serial console connected?

ToRvaLDz on 14 Mar 2019

Nope. Can we get more debug data that way?

Bloodyagent on 14 Mar 2019

No, just because I noticed a ghost switching issue while it is not mounted in the wall socket

ToRvaLDz on 14 Mar 2019

@Bloodyagent component is compatible with schematic I drawn! I never seen bad data sheet like this!
There are only one thing that seems wrong with it. So, since pin1,2 are connected to gnd, the outputs behavior should be like this:

but in this way seems impossible to recognize long press because outputs toggle only the button is pressed again. There are others configurations that permit to detect long press but maybe is inverted in datasheet. I have to check this with oscilloscope.
However, looking datasheet, since description of detection is not described(but I think to know that), there is a capacitor that determines touch sensitivity. It's capacitor that I also drawn on pin 7. Datasheet suggests value from 4.7nF to 22nF. Translating from Chinese seems we have to reduce it in order to reduce sensitivity. If I have time tomorrow unmount cap from board and try to measure and substitute it with smaller value.

beppeserafini on 14 Mar 2019

All chinese documentation is bad🤣 Either way dude you a genius. I'm very rusty with digital electronics, I dont even own a scope anymore😫

Bloodyagent on 14 Mar 2019

Pin7 is chip select? Not sure how that's going to affect touch sensitivity?

Bloodyagent on 14 Mar 2019

Its C6? Ok just measured it. I get 22nf. Which would agree with your theory.

Bloodyagent on 14 Mar 2019

Ok. I replaced it with a 4.7nf (That's all I have at home at the moment) Its reduced sensitivity to the point now where it only works with cover off. So looks like you are right!! I may just have to increase it to 10nf, so will have to wait for tomorrow. For now I will leave it on testbench overnight. Lets see if it fixes the ghost issue!!

Bloodyagent on 14 Mar 2019

I have been using 1 & 2 channel Sonoff switches for over a year without any ghost switching problems. I was using various versions of Tasmota so felt I should update all to 6.4.1. Big mistake now all 1 & 2 channel switches have random ghost switching. Therefore I strongly suspect this is some form of bug in 6.4.1.

AndyGoldie on 14 Mar 2019

Hi Andy
I not have any issue with 1 and 2 ch. Just the 3ch. @beppeserafini seems to have nailed it with the touch sensitivity cap. So if you handy with a solder iron you can try change to a lower value. We still trying to figure out the correct value. I will also go back to an older firmware and try that. Thanks!

Bloodyagent on 14 Mar 2019

Yes I am. I design electronic boards daily. 😄
Until now, since I put paper sheet between pad and cover, I didn't get ghost switches from that Sonoff. Doing that mean increase dielectric material and increase distance between "capacitor" armors, that would be decrease button sensitivity. So I think if it's mounted 22nF, a value about 15nF or 18nF could be good for us.

_Pin 7 isn't chip select. It is called Csel. It's not explained how it works but generally pin like that work like a constant time (series resistor inside) and it is compared with signals measured on pad pin._

beppeserafini on 15 Mar 2019

Its C6? Ok just measured it. I get 22nf. Which would agree with your theory.

Yes it is

beppeserafini on 15 Mar 2019

@beppeserafini

Until now, since I put paper sheet between pad and cover, I didn't get ghost switches from that Sonoff. Doing that mean increase dielectric material and increase distance between "capacitor" armors, that would be decrease button sensitivity.

Nice finding :+1:

ascillato on 15 Mar 2019

I walked into my study this morning and saw that the modified sonoff hadn't switched all night! So thought wow, its fixed. But then as I was about to start typing this it switched. I don't know if its because I'm sitting right here next to it and its picking up my body (lol) or if its because I placed my cell phone right next to it. I will get some higher value caps and try them with the units in the wall and see if its resolved. Will also try the paper idea. Anyway thanks for the help and assistance. This has been quite educational! Either way it seems we have 2 options now to resolve, and its clearly a hardware issue so I am going to close this now.

Bloodyagent on 15 Mar 2019

adios amigos:) for now

Bloodyagent on 15 Mar 2019

Can someone explain properly where to find theese capacitor in the board and possible values to use to lower the sensitivity? We have about 30 Sonoff and about 4 of them suffer from ghost switching, it's really annoing. It is quite strange that those who suffer from this problem are positioned in the bathrooms.

simoneluconi on 1 Apr 2019

It is C6 capacitor (0603 case). You can replace it with value near to 16nF. I mount 16,4nF with no sensitivity degradation but no ghost switching.

beppeserafini on 1 Apr 2019

🎉1

I can also confirm this has resolved the ghosting issue.

Bloodyagent on 1 Apr 2019

Are you sure that ghost switching you have is not about RF interference? In this case you can follow also my solution in first reply I did

beppeserafini on 1 Apr 2019

How could it be RF Interference? If no remote is programmed the switch couldn't switch by itself.

simoneluconi on 1 Apr 2019

If you look on web you can see some video showing this issue on others sonoff products that are using same RF chip.

beppeserafini on 1 Apr 2019

Should also point out that for that i can ear the sonoff also beeps, so it detect a long press on the button, not some "ghost press" or i believe "RF Interference"

simoneluconi on 3 Apr 2019

If you ear beeps it's a sensitivity problem that you can resolve changing capacitor. I mount on my sonoff devices two 8.2nF ceramic capacitors in parallels (mount the second on top of the first)

beppeserafini on 3 Apr 2019

👍1

any other thoughts? I have ghosting issues on a 3ch T1 US, have put two 8.2nf ceramic caps on there, still having ghosting issues.

xtremedays on 26 Apr 2019

You can try with lower cap value, I.e. 15nF.
If you think is a RF interference and you aren’t using RF remote controller you can follow pcb modifications that I introduced in my first post https://github.com/arendst/Sonoff-Tasmota/issues/5449#issuecomment-472136413

beppeserafini on 26 Apr 2019

👍1

i've removed the RF chip, so doubt its that. Does it have to be a ceramic capacitor?

xtremedays on 26 Apr 2019

Yes, datasheet suggest X5R if I remember correctly.

beppeserafini on 26 Apr 2019

Have a look on to Ghost / Random Switching on Sonoff Devices (MQTT Retained)

HobbytronicsPK on 13 May 2019

Just curious where this has left off. It appears by comments here that CAPs did not necessarily fix the problem, so I have not tried it. I did followed the MQTT Retained instructions in link above, but I still get ghost switching on both of my T1 3CH. The T1 starts with beep.... beep.beep. Then the light changes state. It is most often the center switch, but the top and bot switch also trigger on occasion.

REALogics on 23 May 2019

With log turned ON, the ghost switch happens here:

`02:27:03 WIF: Checking connection...
02:27:03 WIF: Connected
02:27:05 APP: Button1 multi-press 1
02:27:16 SRC: Button
02:27:16 MQT: sonoff-5682/tele/STATE = {"Time":"2019-05-25T02:27:16","Uptime":"10T01:17:00","Vcc":3.202,"SleepMode":"Dynamic","Sleep":50,"LoadAvg":19,"POWER1":"OFF","POWER2":"ON","POWER3":"ON","Wifi":{"AP":1,"SSId":"XXXX","BSSId":"A0:63:91:D7:XX:XX","Channel":9,"RSSI":78,"LinkCount":2,"Downtime":"0T00:02:05"}}
02:27:16 MQT: sonoff-5682/stat/RESULT = {"POWER1":"OFF"}
02:27:16 MQT: sonoff-5682/stat/POWER1 = OFF (retained)
02:27:17 SRC: MQTT
02:27:17 RSL: Received Topic sonoff-5682/cmnd/POWER1, Data Size 3, Data OFF
02:27:17 RSL: Group 0, Index 1, Command POWER, Data OFF
02:27:17 MQT: sonoff-5682/tele/STATE = {"Time":"2019-05-25T02:27:17","Uptime":"10T01:17:01","Vcc":3.186,"SleepMode":"Dynamic","Sleep":50,"LoadAvg":19,"POWER1":"OFF","POWER2":"ON","POWER3":"ON","Wifi":{"AP":1,"SSId":"XXXX","BSSId":"A0:63:91:D7:XX:XX","Channel":9,"RSSI":72,"LinkCount":2,"Downtime":"0T00:02:05"}}
02:27:17 MQT: sonoff-5682/stat/RESULT = {"POWER1":"OFF"}
02:27:17 MQT: sonoff-5682/stat/POWER1 = OFF (retained)
02:27:17 CFG: Saved to flash at F4, Count 304, Bytes 3584
02:27:23 WIF: Checking connection...
02:27:23 WIF: Connected'

REALogics on 25 May 2019

Did you try to keep sonoff without glass cover?
In my opinion is a sensitivity problem about touch

beppeserafini on 25 May 2019

I was trying to avoid modifying hardware, thinking it could be done in firmware. After looking at the chip document, it appears WTC6104 chip sensitivity is set by a capacitor (unlike the MPR121 set by firmware). Since we are all putting tasmota firmware on these devices, the next question for me is... does this happen with Sonoff T1's with stock factory binary. I don't have one to test. But if stock binary does not have the issue, then there may still be something about tasmota firmware causing this. I am studying the tasmota firmware at the moment.

By the way, with Home Assistant PI turned off (not controlling Sonoffs), I still had a T1 US button change state over night.

REALogics on 28 May 2019

Me too when the problem appeared I suspected that was a firmware bug but I didn’t have anymore sonoff to try. After that I saw no ghost switching without glass cover and I looked for hardware bugs..

beppeserafini on 28 May 2019

Im using T1 3 gang usa for 2 weeks. 11 days with stock FW and last 2 days with Tasmota. I havent noticed this problem on stock and noticed it 3 times in last 2 days with tasmota.

I have configured OpenHab to turn on some other lights on long press so even if i dont hear beep, pause, beep beep, I can see in the morning which channel ghosted a long press.

There's always a chance this is only accident, that somehow i missed ghost clicks on the stock fw, but I'd say 11 days (0) to 2 days (3)... Slim chance. I bet on some tasmota error or at least hardware error byoassed in original fw and not in tasmota.

psiho on 30 May 2019

This is my scenario.

T1 1 gang: 3 un
T1 3 gang: 1 un

One T1 1 gang I’m running esp easy firmware. Others T1’s I’m running Tasmota firmware.

First T1 1 gang I improved were flashed with esp easy firmware and I got ghost switch at few hours later.

When received other T1’s I was deicied flash with Tasmota.

T1’s 1 gang running last Tasmota firmware start ghost switch after few hours later also.

As I had two T1 1 gang with different firmwares version with ghost switching, I was deciding test T1 3 gang with stock firmware and leave it for two weeks, results no ghost switching. After two weeks, flashed same T1 3 gang with Tasmota, result ghost switch appear again.

After that, I was deicied running all T1’s 1 and 3 gang without glass cover, no ghost switching for three days tested. After three days, remount glass cover but now with an different assembly approach leaving glass cover for last one mounting it very carefully , results no ghost switching for few tested days, including T1 running esp easy firmware.

All T1’s running Tasmota firmware I had pasted below commands at console

SwitchRetain Off
ButtonRetain On
ButtonRetain Off
Powerretain On

Part a time I thinked could be touch sensibility but when running with stock firmware no ghost switch appear.

All times I had ghost switching I heard beep, pause beep beep.

sorokaalex on 31 May 2019

Hi

What version of Tasmota are you running?

I had the ghost switching problem with version 6.5

Once I downgraded the version to 6.3 all ghost problems stop with no modification to the devices required.

Best regards

Andy Goldie
On 31 May 2019, at 22:37, sorokaalex <[email protected]notifications@github.com> wrote:

This is my scenario.

T1 1 gang: 3 un
T1 3 gang: 1 un

One T1 1 gang I’m running esp easy firmware. Others T1’s I’m running Tasmota firmware.

First T1 1 gang I improved were flashed with esp easy firmware and I got ghost switch at few hours later.

When received other T1’s I was deicied flash with Tasmota.

T1’s 1 gang running last Tasmota firmware start ghost switch after few hours later also.

As I had two T1 1 gang with different firmwares version with ghost switching, I was deciding test T1 3 gang with stock firmware and leave it for two weeks, results no ghost switching. After two weeks, flashed same T1 3 gang with Tasmota, result ghost switch appear again.

After that, I was deicied running all T1’s 1 and 3 gang without glass cover, no ghost switching for three days tested. After three days, remount glass cover but now with an different assembly approach leaving glass cover for last one mounting it very carefully , results no ghost switching for few tested days, including T1 running esp easy firmware.

All T1’s running Tasmota firmware I had pasted below commands at console

SwitchRetain Off
ButtonRetain On
ButtonRetain Off
Powerretain On

Part a time I thinked could be touch sensibility but when running with stock firmware no ghost switch appear.

—
You are receiving this because you commented.
Reply to this email directly, view it on GitHubhttps://github.com/arendst/Sonoff-Tasmota/issues/5449?email_source=notifications&email_token=AIO5ZCRWMFKEPEURT5FVST3PYGLCJA5CNFSM4G5OGQN2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODWWN2GQ#issuecomment-497868058, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AIO5ZCWJVZSPLFWMSW2XDLDPYGLCJANCNFSM4G5OGQNQ.

AndyGoldie on 1 Jun 2019

Hi What version of Tasmota are you running? I had the ghost switching problem with version 6.5 Once I downgraded the version to 6.3 all ghost problems stop with no modification to the devices required. Best regards Andy Goldie

I’m running 6.5 firmware on all Tasmota devices. If one more ghost switch appear I’ll try downgrade firmware version.

sorokaalex on 1 Jun 2019

@AndyGoldie

May I request that you try your tests with 6.5.0.13 available at http://thehackbox.org/tasmota/020300/sonoff.bin??

If your results are similar, please open a new issue providing the details of your stock firmware tests, Tasmota 6.3, and Tasmota 6.5.0.13.. Perhaps these clues will result in Theo being able to uncover what causes the instability.

Thanks

Mike

meingraham on 1 Jun 2019

I have this issue and it is not restricted to 3 channel versions. I am currently seeing it on 1s, 2s and 3s. 3s are most prone to it, however I think this is straight up a function of having 3 pads.

Currently I have 2 x 3 channel, 2 x 1 channel and 3 x 2 channels installed and all have ghosted. Mtqq server is offline, the only solution is to have the faceplate off. I tried the paper behind the glass option without success.

I will try the firmware above. Testing wise a 3ch will ghost approximately once every 6 to 8 hours.

I have 1 3ch which still has stock firmware. It has had zero ghost switches.

kyemacdonald on 1 Jun 2019

With the 6.5.0 firmware, I set "SetOption13=1" to disable multi-button action (enables single action only). The Beep -- Beep - Beep still occurs once or twice a day BUT the lights do not switch as a result.

REALogics on 3 Jun 2019

I've got setoption13 1 on all my switches and they still switch the lights. This was one of the things I tried.

kyemacdonald on 3 Jun 2019

I've updated with the firmware listed above and set weblog 4, as well as SwitchRetain Off; ButtonRetain On; ButtonRetain Off; PowerOnState 3; PowerRetain On; SetOption13 1

For the purposes of this test the MQTT server field is left intentionally blank.

Information page output

Sonoff T1 3CH Module
Sonoff Master Bedroom 1
Program Version 6.5.0.14(d0502c1-sonoff)
Build Date & Time 2019-06-02T19:01:37
Core/SDK Version 2_3_0/1.5.3(aec24ac9)
Uptime 0T00:03:23
Flash write Count 231 at 0xF8000
Boot Count 20
Restart Reason Software/System restart
Friendly Name 1 Sonoff Master Bedroom 1
Friendly Name 2 Sonoff Master Bedroom 2
Friendly Name 3 Sonoff Stairwell Light Upper

AP1 SSId (RSSI) Skyenet (100%)
Hostname sonoffmasterbedroom-6947
IP Address 10.1.1.52
Gateway 10.1.1.254
Subnet Mask 255.255.255.0
DNS Server 10.1.1.254
MAC Address C4:4F:33:9E:DB:23

MQTT Host
MQTT Port 1883
MQTT User mqtt
MQTT Client DVES_9EDB23
MQTT Topic sonoffmasterbedroom
MQTT Group Topic sonoffs
MQTT Full Topic sonoffmasterbedroom/cmnd/
MQTT Fallback Topic cmnd/DVES_9EDB23_fb/

Emulation None
mDNS Discovery Disabled

ESP Chip Id 10410787
Flash Chip Id 0x144051
Flash Size 1024kB
Program Flash Size 1024kB
Program Size 524kB
Free Program Space 476kB
Free Memory 14kB

Initial boot output:
00:00:00 CFG: Loaded from flash at F9, Count 230
00:00:00 SRC: Restart
00:00:00 Project sonoff Sonoff Master Bedroom 1 Version 6.5.0.14(d0502c1-sonoff)-2_3_0
00:00:00 WIF: Checking connection...
00:00:00 WIF: Attempting connection...
00:00:00 WIF: Patch issue 2186
00:00:00 WIF: Connecting to AP1 Skyenet in mode 11N as sonoffmasterbedroom-6947...
00:00:01 WIF: Checking connection...
00:00:01 WIF: Attempting connection...
00:00:02 WIF: Checking connection...
00:00:02 WIF: Attempting connection...
00:00:03 WIF: Checking connection...
00:00:03 WIF: Attempting connection...
00:00:04 WIF: Checking connection...
00:00:04 WIF: Connected
00:00:04 HTP: Web server active on sonoffmasterbedroom-6947 with IP address 10.1.1.52
00:00:06 HTP: Main Menu
00:00:08 HTP: Console
00:00:10 APP: Boot Count 20
00:00:10 CFG: Saved to flash at F8, Count 231, Bytes 3584
00:00:12 APP: (UTC) Mon Jun 03 01:24:24 2019, (DST) Sun Mar 31 02:00:00 2019, (STD) Sun Oct 27 03:00:00 2019
02:24:32 CMD: backlog SwitchRetain Off; ButtonRetain On; ButtonRetain Off; PowerOnState 3; PowerRetain On
02:24:32 SRC: WebConsole from 10.1.1.45
02:24:32 RSL: Received Topic /backlog, Data Size 83, Data SwitchRetain Off; ButtonRetain On; ButtonRetain Off; PowerOnState 3; PowerRetain On
02:24:32 RSL: Group 0, Index 1, Command BACKLOG, Data SwitchRetain Off; ButtonRetain On; ButtonRetain Off; PowerOnState 3; PowerRetain On
02:24:32 SRC: Backlog
02:24:32 RSL: Received Topic /SwitchRetain, Data Size 3, Data Off
02:24:32 RSL: Group 0, Index 1, Command SWITCHRETAIN, Data Off
02:24:32 RSL: sonoffmasterbedroom/stat/RESULT = {"SwitchRetain":"OFF"}
02:24:32 SRC: Backlog
02:24:32 RSL: Received Topic /ButtonRetain, Data Size 2, Data On
02:24:32 RSL: Group 0, Index 1, Command BUTTONRETAIN, Data On
02:24:32 RSL: sonoffmasterbedroom/stat/RESULT = {"ButtonRetain":"ON"}
02:24:33 SRC: Backlog
02:24:33 RSL: Received Topic /ButtonRetain, Data Size 3, Data Off
02:24:33 RSL: Group 0, Index 1, Command BUTTONRETAIN, Data Off
02:24:33 RSL: sonoffmasterbedroom/stat/RESULT = {"ButtonRetain":"OFF"}
02:24:33 SRC: Backlog
02:24:33 RSL: Received Topic /PowerOnState, Data Size 1, Data 3
02:24:33 RSL: Group 0, Index 1, Command POWERONSTATE, Data 3
02:24:33 RSL: sonoffmasterbedroom/stat/RESULT = {"PowerOnState":3}
02:24:33 SRC: Backlog
02:24:33 RSL: Received Topic /PowerRetain, Data Size 2, Data On
02:24:33 RSL: Group 0, Index 1, Command POWERRETAIN, Data On
02:24:33 RSL: sonoffmasterbedroom/stat/RESULT = {"PowerRetain":"ON"}
02:24:36 WIF: Checking connection...
02:24:36 WIF: Connected
02:24:45 CMD: SetOption13 1
02:24:45 SRC: WebConsole from 10.1.1.45
02:24:45 RSL: Received Topic /SetOption13, Data Size 1, Data 1
02:24:45 RSL: Group 0, Index 13, Command SETOPTION, Data 1
02:24:45 RSL: sonoffmasterbedroom/stat/RESULT = {"SetOption13":"ON"}
02:24:56 WIF: Checking connection...
02:24:56 WIF: Connected
02:25:11 CMD: weblog 4
02:25:11 SRC: WebConsole from 10.1.1.45
02:25:11 RSL: Received Topic /weblog, Data Size 1, Data 4
02:25:11 RSL: Group 0, Index 1, Command WEBLOG, Data 4
02:25:11 RSL: sonoffmasterbedroom/stat/RESULT = {"WebLog":4}
02:25:16 WIF: Checking connection...
02:25:16 WIF: Connected
02:25:36 WIF: Checking connection...
02:25:36 WIF: Connected

kyemacdonald on 3 Jun 2019

I am reopening this issue as I am still experiencing ghosting after lowering C6 capacitor to 16nf.

Only happens on 3 gang units
Is not an issue with original sonoff firmware
Is not related to mqtt retain setting
Capacitor C6 controls touch sensitivity, but lowering the value does not seem to resolve completely
Seems sonoff somehow rectified the hardware design issue through firmware

Bloodyagent on 3 Jun 2019

Morning all,

So I had ghost switching on a 3 gang last night with the firmware above. I didn't catch it in the weblog so I have configured a logging server to catch the next one.

Unlike bloodyagent I am definitely getting them on 2ch switches as well as 3ch.

kyemacdonald on 3 Jun 2019

Managed to catch the change on a 2 channel. This one is connected to MQTT server.

23:18:08 WIF: Checking connection...
23:18:08 WIF: Connected
23:18:28 WIF: Checking connection...
23:18:28 WIF: Connected
23:18:48 WIF: Checking connection...
23:18:48 WIF: Connected
23:19:08 WIF: Checking connection...
23:19:08 WIF: Connected
23:19:28 WIF: Checking connection...
23:19:28 WIF: Connected
23:19:48 WIF: Checking connection...
23:19:48 WIF: Connected
23:20:08 WIF: Checking connection...
23:20:08 WIF: Connected
23:20:26 MQT: sonoffliving/tele/HASS_STATE = {"Version":"6.5.0.14(d0502c1-sonoff)","BuildDateTime":"2019-06-02T19:01:37","Core":"2_3_0","SDK":"1.5.3(aec24ac9)","Module":"Sonoff T1 2CH","RestartReason":"Software/System restart","Uptime":"0T11:45:01","WiFi LinkCount":1,"WiFi Downtime":"0T00:00:04","MqttCount":1,"BootCount":21,"SaveCount":286,"IPAddress":"10.1.1.53","RSSI":"100","LoadAvg":19}
23:20:28 WIF: Checking connection...
23:20:28 WIF: Connected
23:20:38 MQT: sonoffliving/tele/STATE = {"Time":"2019-06-03T23:20:38","Uptime":"0T11:45:13","SleepMode":"Dynamic","Sleep":50,"LoadAvg":19,"POWER1":"OFF","POWER2":"OFF","Wifi":{"AP":1,"SSId":"Skyenet","BSSId":"1C:F2:9A:CA:FE:A3","Channel":1,"RSSI":100,"LinkCount":1,"Downtime":"0T00:00:04"}}
23:20:48 WIF: Checking connection...
23:20:48 WIF: Connected
23:20:49 APP: Button1 immediate
23:20:49 SRC: Button
23:20:49 MQT: sonoffliving/tele/STATE = {"Time":"2019-06-03T23:20:49","Uptime":"0T11:45:24","SleepMode":"Dynamic","Sleep":50,"LoadAvg":19,"POWER1":"ON","POWER2":"OFF","Wifi":{"AP":1,"SSId":"Skyenet","BSSId":"1C:F2:9A:CA:FE:A3","Channel":1,"RSSI":100,"LinkCount":1,"Downtime":"0T00:00:04"}}
23:20:49 MQT: sonoffliving/stat/RESULT = {"POWER1":"ON"}
23:20:49 MQT: sonoffliving/stat/POWER1 = ON (retained)
23:20:50 CFG: Saved to flash at FA, Count 287, Bytes 3584
23:21:08 WIF: Checking connection...
23:21:08 WIF: Connected
23:21:15 HTP: Console
23:21:28 WIF: Checking connection...
23:21:28 WIF: Connected
23:21:48 WIF: Checking connection...
23:21:48 WIF: Connected
23:22:08 WIF: Checking connection...
23:22:08 WIF: Connected

kyemacdonald on 4 Jun 2019

Additional on the switch that triggered.

Sonoff T1 2CH Module
Sonoff Living 1
Program Version 6.5.0.14(d0502c1-sonoff)
Build Date & Time 2019-06-02T19:01:37
Core/SDK Version 2_3_0/1.5.3(aec24ac9)
Uptime 0T11:50:15
Flash write Count 287 at 0xFA000
Boot Count 21
Restart Reason Software/System restart
Friendly Name 1 Sonoff Living 1
Friendly Name 2 Sonoff Living 2

AP1 SSId (RSSI) Skyenet (100%)
Hostname sonoffliving-2245
IP Address 10.1.1.53
Gateway 10.1.1.254
Subnet Mask 255.255.255.0
DNS Server 10.1.1.254
MAC Address 84:0D:8E:67:C8:C5

MQTT Host 10.1.1.99
MQTT Port 1883
MQTT User mqtt
MQTT Client DVES_67C8C5
MQTT Topic sonoffliving
MQTT Group Topic sonoffs
MQTT Full Topic sonoffliving/cmnd/
MQTT Fallback Topic cmnd/DVES_67C8C5_fb/

Emulation None
mDNS Discovery Disabled

ESP Chip Id 6801605
Flash Chip Id 0x144051
Flash Size 1024kB
Program Flash Size 1024kB
Program Size 524kB
Free Program Space 476kB
Free Memory 13kB

kyemacdonald on 4 Jun 2019

@Bloodyagent

Have you tried increasing debounce with buttondebounce command? (https://github.com/arendst/Sonoff-Tasmota/wiki/Commands#control)

ascillato on 4 Jun 2019

Denounce set to 1000 to test

kyemacdonald on 4 Jun 2019

@kyemacdonald have you reduced size of C6 capacitor? This definitely helps but does not eliminate completely.
@ascillato no have not tried, will give it a go, thanks.

Bloodyagent on 4 Jun 2019

Debounce to 1000 is too much. Try first 100 please

ascillato on 4 Jun 2019

I apologize. In my earlier post when I said I "SetOption13 1"

I failed to indicated that at the same time I set "ButtonDebounce 500".

Since setting these two option settings last week, the 3x beeps still occur once or twice a day, but the states of the buttons/lights do not change.

I have 2x "T1 US 3CH" sonoff switches.

REALogics on 4 Jun 2019

I don’t remember if my final cap value was 16nF. Maybe a little bit lower. However in my home I’m using 7 sonoff T1 3 gang for 4 months with this setting and I NEVER had ghost switching.
So in my opinion in original firmware there is some algorithm to bypass hardware issue but adjusting capacitance we fix it

beppeserafini on 4 Jun 2019

@Bloodyagent

Have you tried increasing debounce with buttondebounce command?

ascillato on 4 Jun 2019

@Bloodyagent No I haven't swapped the capacitor as yet as the stock firmware doesn't have this issue. So from my perspective, at this stage anyway, a software solution should be possible.

@ascillato I've gone to the extreme of 1000 to see if that removes the problem completely. If it does then I will lower the number until it starts happening again. It does make the switches pretty horrible to use set that high as you have to wait a noticeable length of time before you can activate the next button in the gang.

kyemacdonald on 5 Jun 2019

@kyemacdonald

Please, use a small value like 100 for debounce. Thanks

ascillato on 5 Jun 2019

@ascillato - done. Let me know if there is any other testing setting that would be useful.

kyemacdonald on 5 Jun 2019

That debounce time should be enough and the response time is not noticeable (on button release). If you want faster response time (on button push), disable multipress with setoption commands.

ascillato on 5 Jun 2019

Switch triggered with debounce set to 100. Upping to 200

03:54:28 WIF: Checking connection...
03:54:28 WIF: Connected
03:54:31 APP: Button2 immediate
03:54:31 SRC: Button
03:54:31 MQT: sonoffliving/tele/STATE = {"Time":"2019-06-05T03:54:31","Uptime":"0T00:51:47","SleepMode":"Dynamic","Sleep":50,"LoadAvg":19,"POWER1":"OFF","POWER2":"ON","Wifi":{"AP":1,"SSId":"Skyenet","BSSId":"1C:F2:9A:CA:FE:A3","Channel":1,"RSSI":100,"LinkCount":1,"Downtime":"0T00:00:04"}}
03:54:31 MQT: sonoffliving/stat/RESULT = {"POWER2":"ON"}
03:54:31 MQT: sonoffliving/stat/POWER2 = ON (retained)
03:54:32 CFG: Saved to flash at F9, Count 376, Bytes 3584
03:54:48 WIF: Checking connection...
03:54:48 WIF: Connected
03:55:08 WIF: Checking connection...
03:55:08 WIF: Connected
03:55:17 HTP: Main Menu
03:55:17 HTP: Console

kyemacdonald on 5 Jun 2019

Switch triggered with debounce set to 200. Upping to 500.

07:16:09 WIF: Checking connection...
07:16:09 WIF: Connected
07:16:29 WIF: Checking connection...
07:16:29 WIF: Connected
07:16:38 APP: Button1 immediate
07:16:38 SRC: Button
07:16:38 MQT: sonoffliving/tele/STATE = {"Time":"2019-06-05T07:16:38","Uptime":"0T04:13:54","SleepMode":"Dynamic","Sleep":50,"LoadAvg":19,"POWER1":"ON","POWER2":"OFF","Wifi":{"AP":1,"SSId":"Skyenet","BSSId":"1C:F2:9A:CA:FE:A3","Channel":1,"RSSI":100,"LinkCount":1,"Downtime":"0T00:00:04"}}
07:16:38 MQT: sonoffliving/stat/RESULT = {"POWER1":"ON"}
07:16:38 MQT: sonoffliving/stat/POWER1 = ON (retained)
07:16:38 CFG: Saved to flash at F6, Count 379, Bytes 3584
07:16:49 WIF: Checking connection...
07:16:49 WIF: Connected
07:17:09 WIF: Checking connection...
07:17:09 WIF: Connected

kyemacdonald on 5 Jun 2019

@Bloodyagent

Have you tried increasing debounce with buttondebounce command?

ascillato2 on 5 Jun 2019

Closing this issue as there is no feedback. When you have time, please share your results. Thanks.

ascillato2 on 5 Jun 2019

Buttondebounce set has 500 has removed ghost press on all my switches.

I believe there is a difference in the sensitivity between individual switches. 500 is likely too high for most of my switches and I will test lowering it, however 1 of my switches (the logs above) was particularly sensitive.

kyemacdonald on 6 Jun 2019

Great findings. Thanks a lot for sharing. :+1:

ascillato on 6 Jun 2019

Sorry for delayed feedback. Have work kids and wife to deal with! Have changed debounce to 200, so far so good.

Bloodyagent on 6 Jun 2019

Why does this issue get closed when there is a known problem with no conclusive resolve? The fact that "SetOption13 1" and ButtonDebounce 500" has prevented false state changes on my two T1 SS 3CH only suggests that firmware needs investigation. I still get an occasional 3 beeps with "SetOption13 1", but why? And a "ButtonDebounce 500" setting is unrealistic. I will set 100ms debounce tonight.
Changing capacitors internally is only an investigative approach, but not an answer for the masses. I would dig into the firmware myself, if it were not for other pressing activities. This issue should stay open until a viable fix is identified, and firmware updated if necessary.

REALogics on 6 Jun 2019

An update.

Buttondebounce 500 has not removed ghost switching entirely. It has reduced the frequency though.

Severely is definitely different between each piece of hardware.

kyemacdonald on 7 Jun 2019

I have same problem.
I hear 1 beep the fast 2 beeps and light is turning on
Increasing buttondebounce helps but I need to put over 500 and its unrealistic to use it with so slow response time

Mikuchck on 7 Jun 2019

👍1

Same issue here in my Sonoff T1 3CH US.
Changed the settings to SwitchRetain Off; ButtonRetain On; ButtonRetain Off; PowerOnState 3; PowerRetain On
removed /var/lib/mosquitto/mosquitto.db, restarted the service, but the ghost switching still happens.

ericalvim on 8 Jun 2019

@ascillato - would you like to re-open this or would you rather I created a new issue?

Debounce doesn't solve the issue at best it lowers the frequency.

I've just done a full reset on the switch that is doing it the most and set it up with full logging. I will upload the logs when it switches.

I was wondering if it would be possible to have a firmware that ignored the first press, waited 40 and then resampled the sensor. Would this be possible?

kyemacdonald on 8 Jun 2019

No ghosting on all 3 of my T1 3CH US units for 2 days now. So looking promising.
I did the following:

Replaced the 22nf c6 cap with a 16nf
Changed debounce to 200ms.
Seems to do the trick

Bloodyagent on 8 Jun 2019

👍2

Ok took 20 minutes for the first ghost press to occur

06:50:54 WIF: Connected
06:51:14 WIF: Checking connection...
06:51:14 WIF: Connected
06:51:14 APP: Button1 multi-press 1
06:51:23 SRC: Button
06:51:23 RSL: stat/sonoff/RESULT = {"POWER1":"ON"}
06:51:23 RSL: stat/sonoff/POWER1 = ON
06:51:23 CFG: Saved to flash at F4, Count 16, Bytes 3584
06:51:34 WIF: Checking connection...
06:51:34 WIF: Connected

Also it has caused a reset of the device. So it has gone back to a sonoff basic.

Redoing with setoption1 on, SetOption13 on & SetOption32 100

kyemacdonald on 8 Jun 2019

@Bloodyagent - surely modding the hardware shouldn't be necessary though. The stock firmware isn't doing this so there must be a software solution.

kyemacdonald on 8 Jun 2019

@kyemacdonald

I agree but it's a very simple procedure and seems a good temporary solution. However the firmware still needs to be looked at and the issue should be reopened.

Bloodyagent on 8 Jun 2019

Lowering the cap to 16nf does reduce the switch sensitivity considerably, so it's not ideal.

Bloodyagent on 8 Jun 2019

Sigh. This switch is doing it all the time. Basically an hour since the last trigger. Upped debounce from 200 to 500.

I'm going to swap the control board with a spare next to see if it's the control board or maybe the power supply from the relay block.

07:55:05 RSL: tele/sonoff/STATE = {"Time":"2019-06-08T07:55:05","Uptime":"0T00:44:59","SleepMode":"Dynamic","Sleep":50,"LoadAvg":19,"POWER1":"ON","POWER2":"ON","Wifi":{"AP":1,"SSId":"Skyenet","BSSId":"1C:F2:9A:CA:FE:A3","Channel":1,"RSSI":96,"LinkCount":1,"Downtime":"0T00:00:04"}}
07:55:10 WIF: Checking connection...
07:55:10 WIF: Connected
07:55:15 APP: Button2 immediate
07:55:15 SRC: Button
07:55:15 RSL: stat/sonoff/RESULT = {"POWER2":"OFF"}
07:55:15 RSL: stat/sonoff/POWER2 = OFF
07:55:16 CFG: Saved to flash at F5, Count 31, Bytes 3584
07:55:30 WIF: Checking connection...
07:55:30 WIF: Connected
07:55:50 WIF: Checking connection...

kyemacdonald on 8 Jun 2019

Just tried the latest firmware based on the 2.5.2 core. Still ghost switches.

12:57:41 WIF: Attempting connection...
12:57:42 WIF: Checking connection...
12:57:42 WIF: Connected
12:58:02 WIF: Checking connection...
12:58:02 WIF: Connected
12:58:13 APP: Button1 immediate
12:58:13 SRC: Button
12:58:13 RSL: stat/sonoff/RESULT = {"POWER1":"ON"}
12:58:13 RSL: stat/sonoff/POWER1 = ON
12:58:13 CFG: Saved to flash at F9, Count 38, Bytes 3584
12:58:22 WIF: Checking connection...
12:58:22 WIF: Attempting connection...
12:58:22 WIF: Connecting to AP1 Skyenet in mode 11N as sonoff-2245...
12:58:23 WIF: Checking connection...
12:58:23 WIF: Attempting connection...
12:58:24 WIF: Checking connection...
12:58:24 WIF: Attempting connection...
12:58:25 WIF: Checking connection...

kyemacdonald on 8 Jun 2019

As I do not have a US T1 I'll start monitoring a UK T1 3CH and see if it ghost switches...

arendst on 8 Jun 2019

Observed the reset to Basic Too. It's the result of numerous restarts due to Exception 9 caused by ets_timer functions.

Need to investigate why this happens on T1's as I never observed it on anything else.

arendst on 8 Jun 2019

I've just realised I don't think I've ever witnessed a light turn off. Only ever ghost switches from off to on.

kyemacdonald on 8 Jun 2019

I also tested the oldest firmware I could see last night and it still switched.

Program Version | 6.3.0.4(15148f2-sonoff)
-- | --
2018-11-06T22:03:19

kyemacdonald on 8 Jun 2019

Still investigating the exception but it seems to be related to LWIP2 and syslog using UDP.

For the ghost switching on a 3 channel T1 you might want to try the template below:

{"NAME":"T1 3CH NoPU","GPIO":[90,255,255,255,23,22,91,92,21,56,0,0,0],"FLAG":0,"BASE":30}

This will disable the internal button pull-up resistors which might not be needed anyway as the touch sensor chip is directly connected to the ESP8286.

arendst on 9 Jun 2019

Right. Just fixed the Exception 9 issue caused by enabling syslog and NTP re-sync message. It's in the latest dev release.

Pls. test ghost switching using the tamplate shown above and report back.

arendst on 9 Jun 2019

For a 2ch what do I need to change in that template? One of my 2ch is by far the worst for switching.

kyemacdonald on 9 Jun 2019

Ok so I've deployed the latest dev release to 2 of my switches. One is a 3ch and one is a 2ch. I have configured the 3ch to use the template above. The 2ch is using the standard module template and both reporting to syslog server.

The 2 channel was given a factory reset and has had only SwitchRetain Off; ButtonRetain On; ButtonRetain Off; PowerOnState 3; PowerRetain On; setoption1 1; setoption13 1;

The 3ch also has debounce set to 400. Because it's in the master bedroom and my wife is going to murder me if it keeps turning on in the middle of the night :P

kyemacdonald on 9 Jun 2019

I've had to remove the remote server logging from the 2ch as it is crashing constantly.

Also I caught the 2ch switching from on to off this evening, so it is bi-directional.

kyemacdonald on 9 Jun 2019

Try this for the t1 2CH

{"NAME":"T1 2CH NoPU","GPIO":[90,255,255,255,0,22,91,0,21,56,0,0,0],"FLAG":0,"BASE":29}

Are you sure the 2CH has the dev version with commit d3eaf74? You might have downloaded an earlier one as it takes some time before the available firmwares have been compiled.

arendst on 9 Jun 2019

This is the from the info page.
Sonoff Living 1
Program Version 6.5.0.15(76919b2-sonoff)
Build Date & Time 2019-06-08T23:01:51
Core/SDK Version 2_4_2/2.2.1(cfd48f3)
Uptime 0T00:16:08

I'm guessing that isn't the right version as the 3ch just had a fit as well.

kyemacdonald on 9 Jun 2019

Currently the correct ones are available at http://thehackbox.org/tasmota/

I also update the 3CH template allowing to status led to function the same as the original one:

{"NAME":"T1 3CH NoPU","GPIO":[90,255,255,255,23,22,91,92,21,56,0,0,0],"FLAG":0,"BASE":30}

arendst on 9 Jun 2019

So the following three templates provide support for T1 1CH, 2CH and 3CH with buttons having no internal pull-up resistors enabled:

{"NAME":"T1 1CH NoPU","GPIO":[90,255,255,255,0,0,0,0,21,56,0,0,0],"FLAG":0,"BASE":28}
{"NAME":"T1 2CH NoPU","GPIO":[90,255,255,255,0,22,91,0,21,56,0,0,0],"FLAG":0,"BASE":29}
{"NAME":"T1 3CH NoPU","GPIO":[90,255,255,255,23,22,91,92,21,56,0,0,0],"FLAG":0,"BASE":30}

arendst on 9 Jun 2019

Thanks,

I think I'm doing something wrong as every time I flash I end up with version 76919b2-sonoff.

Normally I do a minimal and then a sonoff.bin but after 3 goes I changed it to point at the released OTA option and still get that version.

Screenshot from 2019-06-09 23-14-43

kyemacdonald on 9 Jun 2019

Change OTA Url to http://thehackbox.org/tasmota/sonoff.bin including the bin

arendst on 9 Jun 2019

Yep just realised that. Running the update now.

kyemacdonald on 9 Jun 2019

All are now updated to the correct version and all have been setup with the suitable templates and remote logging. I've also set buttondebounce 40 on all bar the bedroom light.

kyemacdonald on 9 Jun 2019

Great! Fingers crossed...

arendst on 9 Jun 2019

😄1

Just incase i've stuffed something obvious up here is the status 0 on all the switches.
status 0 outputs.txt

kyemacdonald on 9 Jun 2019

Didn't take one of them long......

14:38:41 WIF: Checking connection...
14:38:41 WIF: Connected
14:39:01 WIF: Checking connection...
14:39:01 WIF: Connected
14:39:05 APP: Button1 immediate
14:39:05 SRC: Button
14:39:05 MQT: sonoffkitchen/tele/STATE = {"Time":"2019-06-09T14:39:05","Uptime":"0T00:13:52","SleepMode":"Dynamic","Sleep":50,"LoadAvg":19,"POWER1":"ON","POWER2":"OFF","POWER3":"OFF","Wifi":{"AP":1,"SSId":"Skyenet","BSSId":"1C:F2:9A:CA:FE:A3","Channel":11,"RSSI":92,"LinkCount":1,"Downtime":"0T00:00:06"}}
14:39:05 MQT: sonoffkitchen/stat/RESULT = {"POWER1":"ON"}
14:39:05 MQT: sonoffkitchen/stat/POWER1 = ON (retained)
14:39:05 CFG: Saved to flash at F7, Count 304, Bytes 3584
14:39:21 WIF: Checking connection...
14:39:21 WIF: Connected
14:39:41 WIF: Checking connection...

kyemacdonald on 9 Jun 2019

Grrrrrr.

arendst on 9 Jun 2019

ESP-APP.log
ESP-CFG.log
ESP-CMD.log
ESP-HTP.log
ESP-LOG.log
ESP-MQT.log
ESP-RSL.log
ESP-SRC.log
ESP-WIF.log

kyemacdonald on 9 Jun 2019

Other than the one that happened straight after the setup I haven't had another ghost switch.

kyemacdonald on 10 Jun 2019

24hrs and so far no ghost switching!

kyemacdonald on 10 Jun 2019

🎉1

Bollocks bollocks bollocks

00:06:45 WIF: Checking connection...
00:06:45 WIF: Connected
00:06:46 APP: Button2 immediate
00:06:46 SRC: Button
00:06:46 MQT: sonoffkitchen/tele/STATE = {"Time":"2019-06-11T00:06:46","Uptime":"1T09:41:33","SleepMode":"Dynamic","Sleep":50,"LoadAvg":19,"POWER1":"OFF","POWER2":"ON","POWER3":"OFF","Wifi":{"AP":1,"SSId":"Skyenet","BSSId":"1C:F2:9A:CA:E9:AB","Channel":6,"RSSI":86,"LinkCount":6,"Downtime":"0T00:08:17"}}
00:06:46 MQT: sonoffkitchen/stat/RESULT = {"POWER2":"ON"}
00:06:46 MQT: sonoffkitchen/stat/POWER2 = ON (retained)
00:06:46 CFG: Saved to flash at FA, Count 357, Bytes 3584
00:07:05 WIF: Checking connection...
00:07:05 WIF: Connected
00:07:07 HTP: Console

kyemacdonald on 11 Jun 2019

Can anyone confirm the ghost switching also appears on 240V powered T1's?

arendst on 11 Jun 2019

I'm in Australia. So 240v mains supply.

kyemacdonald on 11 Jun 2019

I can confirm the ghost switching appears on 240v T1's I have 4 x 3 gangs and it is ghost switching on all of them.

I have just updated the firmware to the latest dev and applied the template.

Also from Australia.

justflashit on 11 Jun 2019

OK.

What strikes me from your last two ghosts is that at the same time the teleperiod message comes along. I'm starting to suspect the same problem we had with Sonoff Basics being underpowered by design. As the latests Tasmota softwares are doing more background tasks it takes more CPU cycles using more power.

We had a specific sonoff-basic.bin firmware with most sensors disabled which solved it for the Sonoff Basic hardware. Since then the sleep code was also revised lowering CPU load too.

I noticed the sonoff-basic.bin is not on hackbox but if you compile yourself you might want to give it try.

In the meantime try to increase sleep to 100 and see what happens.

arendst on 11 Jun 2019

Damn just had a an issue

19:32:57 WIF: Checking connection...
19:32:57 WIF: Connected
19:33:17 WIF: Checking connection...
19:33:17 WIF: Connected
19:33:24 APP: Button2 immediate
19:33:24 SRC: Button
19:33:24 MQT: stat/Test/RESULT = {"POWER2":"ON"}
19:33:24 MQT: stat/Test/POWER2 = ON
19:33:25 CFG: Saved to flash at F7, Count 228, Bytes 3584
19:33:26 MQT: tele/Test/STATE = {"Time":"2019-06-11T19:33:26","Uptime":"0T01:45:16","SleepMode":"Dynamic","Sleep":50,"LoadAvg":19,"POWER1":"OFF","POWER2":"ON","POWER3":"OFF","Wifi":{"AP":1,"SSId":"Zero","BSSId":"EC:08:6B:67:2F:9E","Channel":3,"RSSI":100,"LinkCount":1,"Downtime":"0T00:00:04"}}
19:33:37 WIF: Checking connection...
19:33:37 WIF: Connected
19:33:57 WIF: Checking connection...
19:33:57 WIF: Connected

justflashit on 11 Jun 2019

I'm in SA so ghost switching on 240v
Only 3 Ch
Always on never off.

Bloodyagent on 11 Jun 2019

I've just compiled the latest sonoff-basic. Do you want me to use the template still?

kyemacdonald on 11 Jun 2019

Here is my firmware is anyone else wants to use it. Use at own risk obviously - https://drive.google.com/open?id=1jFO9Io4difXtTW4LGVS5v9qWGcz1uC5c

Status 0 output
2:20:56 RSL: Group 0, Index 1, Command STATUS, Data 0
12:20:56 MQT: sonoffliving/stat/STATUS = {"Status":{"Module":0,"FriendlyName":["Sonoff Living 1","Sonoff Living 2"],"Topic":"sonoffliving","ButtonTopic":"0","Power":0,"PowerOnState":3,"LedState":1,"LedMask":"FFFF","SaveData":1,"SaveState":1,"SwitchTopic":"0","SwitchMode":[0,0,0,0,0,0,0,0],"ButtonRetain":0,"SwitchRetain":0,"SensorRetain":0,"PowerRetain":1}}
12:20:56 MQT: sonoffliving/stat/STATUS1 = {"StatusPRM":{"Baudrate":115200,"GroupTopic":"sonoffs","OtaUrl":"http://thehackbox.org/tasmota/sonoff.bin","RestartReason":"Software/System restart","Uptime":"0T00:03:51","StartupUTC":"2019-06-11T11:17:05","Sleep":50,"CfgHolder":4617,"BootCount":14,"SaveCount":71,"SaveAddress":"F5000"}}
12:20:56 MQT: sonoffliving/stat/STATUS2 = {"StatusFWR":{"Version":"6.5.0.15(basic)","BuildDateTime":"2019-06-11T21:12:47","Boot":7,"Core":"2_5_2","SDK":"2.2.2-dev(c0eb301)"}}
12:20:57 MQT: sonoffliving/stat/STATUS3 = {"StatusLOG":{"SerialLog":2,"WebLog":4,"SysLog":4,"LogHost":"10.1.1.200","LogPort":514,"SSId":["Skyenet",""],"TelePeriod":300,"Resolution":"558180C0","SetOption":["000AA029","280500000100060000000000000000000000","00000200"]}}
12:20:57 MQT: sonoffliving/stat/STATUS4 = {"StatusMEM":{"ProgramSize":469,"Free":532,"Heap":28,"ProgramFlashSize":1024,"FlashSize":1024,"FlashChipId":"144051","FlashMode":3,"Features":["00000809","0F082380","041483A0","21021704","010001C0"]}}
12:20:57 MQT: sonoffliving/stat/STATUS5 = {"StatusNET":{"Hostname":"sonoffliving-2245","IPAddress":"10.1.1.53","Gateway":"10.1.1.254","Subnetmask":"255.255.255.0","DNSServer":"10.1.1.254","Mac":"84:0D:8E:67:C8:C5","Webserver":2,"WifiConfig":4}}
12:20:57 MQT: sonoffliving/stat/STATUS6 = {"StatusMQT":{"MqttHost":"10.1.1.99","MqttPort":1883,"MqttClientMask":"DVES_%06X","MqttClient":"DVES_67C8C5","MqttUser":"mqtt","MqttCount":1,"MAX_PACKET_SIZE":1000,"KEEPALIVE":30}}
12:20:57 MQT: sonoffliving/stat/STATUS7 = {"StatusTIM":{"UTC":"Tue Jun 11 11:20:57 2019","Local":"Tue Jun 11 12:20:57 2019","StartDST":"Sun Mar 31 02:00:00 2019","EndDST":"Sun Oct 27 03:00:00 2019","Timezone":"+01:00","Sunrise":"04:46","Sunset":"20:53"}}
12:20:57 MQT: sonoffliving/stat/STATUS10 = {"StatusSNS":{"Time":"2019-06-11T12:20:57"}}
12:20:57 MQT: sonoffliving/stat/STATUS11 = {"StatusSTS":{"Time":"2019-06-11T12:20:57","Uptime":"0T00:03:52","SleepMode":"Dynamic","Sleep":50,"LoadAvg":19,"POWER1":"OFF","POWER2":"OFF","Wifi":{"AP":1,"SSId":"Skyenet","BSSId":"1C:F2:9A:CA:FE:A3","Channel":11,"RSSI":96,"LinkCount":1,"Downtime":"0T00:00:06"}}}

kyemacdonald on 11 Jun 2019

Just start using the default template.

arendst on 11 Jun 2019

All set to the default template.

kyemacdonald on 11 Jun 2019

The switches are noticeably faster to respond to touch inputs on this firmware

kyemacdonald on 11 Jun 2019

The switches are noticeably faster to respond to touch inputs on this firmware

The sonoff-basic.bin runs much lighter on both load and power consumption - This is usually an issue where the engineering of the pc board did not allow for sufficient heat dissipation.

The same has been observed on some wemos d1 mini boards which have either voltage regulators unable to match the specification required to run at higher cpu loads or do not have sufficient thermal dissipation - The pc board plays an important role in getting the heat away from the ESP8266/85 chip and in a lot of cases they are manufactured to run simple node-mcu type scripts which are not necessarily very active in the main loop.

@arendst requested the sonoff-basic.bin development binaries to be made available again so that has also been done.

andrethomas on 11 Jun 2019

I have loaded the sonoff basic dev earlier today
Set the it back to the module type 30

Set the sleep to 100

Just recently experienced another ghost switching

19:41:57 WIF: Checking connection...
19:41:57 WIF: Connected
19:42:17 WIF: Checking connection...
19:42:17 WIF: Connected
19:42:22 APP: Button1 immediate
19:42:22 SRC: Button
19:42:22 MQT: stat/Test/RESULT = {"POWER1":"ON"}
19:42:22 MQT: stat/Test/POWER1 = ON
19:42:22 CFG: Saved to flash at F9, Count 260, Bytes 3584
19:42:37 WIF: Checking connection...
19:42:37 WIF: Connected
19:42:57 WIF: Checking connection...
19:42:57 WIF: Connected
19:43:03 HTP: Console

justflashit on 12 Jun 2019

Can anyone confirm the ghost switching also appears on 240V powered T1's?

Can confirm on T1 EU 1CH 230V

fjen on 12 Jun 2019

24hrs no ghost switching

kyemacdonald on 12 Jun 2019

And just like a curse one switched. I didn't catch the log unfortunately, but the switch went from on to off accompanied by the beeps.

kyemacdonald on 13 Jun 2019

@kyemacdonald

I you can set up a remote syslog server on a Raspberry Pi or PC, you can set up the syslog host on the device and turn on SysLog logging. This way you'll have the logs for whenever the issue occurs.

meingraham on 13 Jun 2019

@meingraham I had one setup until yesterday when my dishwasher dying killed my server. I had replaced it but forgot to setup rsyslog.

kyemacdonald on 13 Jun 2019

Stupid dishwasher :wink:

And quit posting that all's well. You're jinxing yourself :wink:

meingraham on 13 Jun 2019

😄1

Is there anything else I can do to help troubleshoot this further.

justflashit on 13 Jun 2019

Different switch this time.

Going to deploy the template from before and see if that makes any difference.

7:19:27 WIF: Connected
07:19:47 WIF: Checking connection...
07:19:47 WIF: Connected
07:20:07 WIF: Checking connection...
07:20:07 WIF: Connected
07:20:27 WIF: Checking connection...
07:20:27 WIF: Connected
07:20:41 APP: Button2 immediate
07:20:41 SRC: Button
07:20:41 MQT: sonoffkitchen/tele/STATE = {"Time":"2019-06-13T07:20:41","Uptime":"0T00:58:34","SleepMode":"Dynamic","Sleep":50,"LoadAvg":19,"POWER1":"ON","POWER2":"ON","POWER3":"OFF","Wifi":{"AP":1,"SSId":"Skyenet","BSSId":"1C:F2:9A:CA:FE:A3","Channel":11,"RSSI":94,"LinkCount":7,"Downtime":"0T00:00:27"}}
07:20:41 MQT: sonoffkitchen/stat/RESULT = {"POWER2":"ON"}
07:20:41 MQT: sonoffkitchen/stat/POWER2 = ON (retained)
07:20:41 CFG: Saved to flash at F6, Count 409, Bytes 3584
07:20:47 WIF: Checking connection...

kyemacdonald on 13 Jun 2019

Hi!

First I want to thank you for an amazing work that you @arendst and the community have done creating this incredible piece of software called tasmota!

I wanted to chime in and provide some additional facts. I have the same issue as @kyemacdonald
with the Sonoff T1 US 3 gang. A curious fact is that it only happens to switches that I recently bought, the ones I bough last year have been working flawlessly without any hiccups. Same firmware and configuration.

When I switched from the standard to the basic firmware as recommended here the beeps and ghosting frequency reduced , but haven't gone away. I've also tried with the modified template with no success.

Is there any test or specific debug that you would like me to provide to help with the debugging?

sergiorademacher on 13 Jun 2019

I'm a bit out if ideas other than power issue.

The fact that some of you hear a beep means the button is pushed and held down at least several seconds as the RF microcontroller started RF detection phase. This also means it's not a tasmota software error as both tasmota AND the RF microcontroller detected the button push. (The RF controller is NOT under tasmota control. Neither are the buttons; they are being fed to a microcontroller before send to the ESP8266 running Tasmota).

As the original software does not do ghost switching the power issue seems to occur by Tasmota software cpu load.

Also my "old" UK and EU T1's do not experience this issue.

arendst on 13 Jun 2019

Just to add that when I am next to the switch I also hear it beep pause beep beep when it ghost switches.

Would it help if we donated you a new 3 gang us T1 switch for you to have and test in either hardware or money to buy one ?

justflashit on 13 Jun 2019

One thing that makes me question the power supply as culprit is the fact they don't ghost switch if the glass plate is removed.

kyemacdonald on 13 Jun 2019

I recently purchased my first US T1 3CH switch and it has been ghosting twice per day every sense I set it up. I have been following along with this thread and testing in parallel. Nothing has changed it's behavior. The times I have experienced the ghosting first hand I have heard the "beep pause beep beep". It's sounds like this must be a hardware issue and not tasmota software. It seems it's pretty obvious the touch button is getting engaged and held for the needed number of seconds to get the beep code. Since I know nothing about the RF capabilities of these switches, does anyone know if an incoming RF signal can create the same beeps? I'm wondering if there could be an issue with these picking up stray RF signals and acting on them inappropriately?

geokscott on 13 Jun 2019

I have a 315Mhz transmitter paired with the T1 3 gang us switch
When you use the transmitter to turn on or off the switch it will not beep at all
in the console you just see the following

01:20:58 APP: Button1 immediate
01:20:58 SRC: Button
01:20:58 MQT: stat/Test/RESULT = {"POWER1":"OFF"}
01:20:58 MQT: stat/Test/POWER1 = OFF
01:20:59 CFG: Saved to flash at F4, Count 321, Bytes 3584

justflashit on 13 Jun 2019

My 2x T1 US 3CH have not ghost switched since "SetOption13 1" and "ButtonDebounce 200" several weeks ago, but I still get beep... beep.beep a few times a day. Using default settings, the ghost switching was always paired with the beeps, and happened a few times a day.

I never tried my switches with original sonoff firmware. Possibly someone saved their original sonoff binary and can reload on suspect switches for a few days... just to verify what others have suggested... that it does not happen with sonoff firmware.

I know little about the code other than I've managed to built it with VSCode::PlatformIO. Can I assume the origin of button detection is coming out of "support_button.ino" and not "support_switch.ino"?

Maybe there is additional debug logging that could be incorporated temporarily to provide additional clues?

REALogics on 13 Jun 2019

One thing that makes me question the power supply as culprit is the fact they don't ghost switch if the glass plate is removed.

@kyemacdonald Good idea. I will try that and I'll report back. Maybe is a contact issue? I will try to swap the glass plates between one without issues.

sergiorademacher on 13 Jun 2019

This is happening to my Sonoff T1 EU V1.0.
With my limited skills I've boiled it down to the RC filtering + debouncing done by the back unmarked chip causing the GPIO0 to be pulled down for a random amount of time.

I've tested this with the board removed and with custom firmware to detect any GPIO change. So this isn't a firmware problem but appears to be an hardware one.

Since I haven't tested this switch with the original FW I cannot be certain that there isn't some calibration done prior to boot that's missing.

Hope this helps :\

EDIT: Also, I'm noticing the reset button pulling the GPIO0 down for a few microseconds but it isn't connected to GPIO04 (relay3?) as written in the documentation. Is this also happening to you?

jpmrodrigues on 13 Jun 2019

I have 3 unflashed switches on hand. I will wire one up to check they aren't doing it.

kyemacdonald on 13 Jun 2019

I can also hookup a stock 3gang T1 tonight.

justflashit on 14 Jun 2019

I've had a stock one running for 36hrs with no switching or beeping of anykind. However it is noticeably slower to respond to touch than a tasmota version, so I'm not sure if there is a way of maybe ignoring some touch inputs or even underclocking the processor via firmware.

kyemacdonald on 15 Jun 2019

Seems sonoff have tweaked the firmware to reduce the sensitivity. My units are all fine with c6 reduced to 16nf and buttondebounce to 200ms.

Bloodyagent on 15 Jun 2019

Ghost switching is now much much rarer for me. The worst offending switch is down to 1 in 36 hours. I've gone from debouce 40 to debounce 100 with the template above about 24 hrs ago without a switch so far.

kyemacdonald on 15 Jun 2019

I swapped the glass plates between and old (no ghost switching) and a new sonoff. Unfortunately the new (and offending) sonoff kept ghosts switching.

I changed the buttondebounce to 100 with the stock template, will report back.

EDIT: Still getting ghost switching. Will try with debouncing + template.
EDIT2: Still ghost switching with the no pull up template.

sergiorademacher on 15 Jun 2019

I have been running the a stock sonoff T1 3 gang for 36hours no ghosting
On the T1 3 gang with the above template and the dev basic bin It has ghosted 4 times in the same 36hours.

I know what I am about to say is not a fix and we are all after a fix but from what I can tell it seems like the ghost issue is the button being held down for over 6 seconds continuously. Holding it for the first 4 seconds it beeps for rf pairing then if you hold it for another 2 seconds you get the beep beep for delete rf pairing. Which is whatg I think I a hearing when the ghost switch is happening

Is it possible to have an option in Tasmota that you can enable so if the button is pressed for more than say 4 seconds, tasmota will ignore trying to change state of the switch?
Obviously you would need to disable this option if you wanted to rf pair or delete and rf pair.

As I mention I know this is not a fix just a work around.

justflashit on 16 Jun 2019

👍2

In addition to the no switching I'm also getting no beeps on the stock T1 3ch.

From previous comments of @arendst he said that the RF chip being triggered bypassed the tasmota firmware all together, that being the case the stock firmware shouldn't be able to influence the double beeps. As it does there must be a way the ESP8266 can interact somewhere along the Button --> Microprocessor --> RF path.

Potentially something like a reset / state change being sent to the microprocessor on a regular basis? My experience of the ghost switching is that there is a minimum time before they occur. ie I don't think I have witnessed a ghost switch at under 2 hrs.

kyemacdonald on 16 Jun 2019

I have been running the a stock sonoff T1 3 gang for 36hours no ghosting
On the T1 3 gang with the above template and the dev basic bin It has ghosted 4 times in the same 36hours.

I know what I am about to say is not a fix and we are all after a fix but from what I can tell it seems like the ghost issue is the button being held down for over 6 seconds continuously. Holding it for the first 4 seconds it beeps for rf pairing then if you hold it for another 2 seconds you get the beep beep for delete rf pairing. Which is whatg I think I a hearing when the ghost switch is happening

Is it possible to have an option in Tasmota that you can enable so if the button is pressed for more than say 4 seconds, tasmota will ignore trying to change state of the switch?
Obviously you would need to disable this option if you wanted to rf pair or delete and rf pair.

As I mention I know this is not a fix just a work around.

How about testing the SetOption32 parameter? That makes sense with what you said.
I changed the settings to SwitchRetain Off; ButtonRetain On; ButtonRetain Off; PowerOnState 3; PowerRetain On
But the ghost switching continues.
And then I set "SetOption13 1" and "ButtonDebounce 200" and "SetOption32 100"

After that, I think there's no more ghost switching going on.
I'm still monitoring. Using 6.5.0 version.

ericalvim on 16 Jun 2019

@ericalvim I've tried setoption32 100 and it doesn't seem to have an effect. I think that option is only relevant for resetting tasmota back to defaults rather than switching. I still leave it in though just in case....

This is my config atm:
backlog SwitchRetain Off; ButtonRetain On; ButtonRetain Off; PowerOnState 3; PowerRetain On; ButtonDebounce 200; SetOption13 1; setoption19 On; sleep 100; setoption32 100;

kyemacdonald on 16 Jun 2019

Having tasmota wait for over 4 seconds to decide the button was NOT pressed makes the response equally slow so that's not an option.

What could be done is triggering the button at the time it is released by the user.

With SetOption13 1 it detects the button at the time the user touches the button, waits ButtondeDebounce time and checks the button state again. If still touched it performs it's action. This provides the most reactive response.

With SetOption13 0 it will action at the moment it is released which will provide a more sluggish response.
With SetOption1 1 it will perform the hold function if held longer than SetOption32 / 10. With no ButtonTopic set it won't do anything.

As users hear a beep and sometimes double beep it means the button is held for over a longer period of time. You might want to try and see what happens with the following settings SetOption13 0 and SetOption1 1

arendst on 16 Jun 2019

To rule out the power supply someone might want to scope the power supply of the device running the stock firmware vs tasmota and look for any significant difference.

arendst on 16 Jun 2019

I have 8 tasmota switches installed, all of which ghost and 3 stock switches wired up on my bench which don't.

I could stick a multimeter on them, but nothing more sensitive than that.

kyemacdonald on 16 Jun 2019

@arendst Thanks for the update.

I have had it on SetOption13 1 and the response was fantastic?

I have changed the settings to your suggestion on SetOption13 0 and SetOption1 1 and as you mention is much more sluggish for the switch to activate like 1 second once I let go it activates. I will test this and see what happens with the ghosting.

What I have also noticed and I have not mentioned is the following with the stock Sonoff

The switch activates once you let go of the button.
If you hold the switch until you hear the first beep (around 4 seconds) and let go it changes state.
If you hold the switch in until you hear the beep pause beep beep (around 6 seconds) and let go it does not change state.

Sorry as for can we get an option for the for 4 seconds I should of worded it differently
Have an option that when enables could be the following:
If button press equal to or less than 4 seconds change state
else do nothing

Will see what I can also setup to data log the voltage of a Stock Sonoff

Also I really appreciate the time and effort you have put into tasmota and think it is a fantastic product.

justflashit on 16 Jun 2019

Just to chime in, i have recently installed 7 of these sonoff t1 1,2 and 3 switches. the single switches will beep and double beep when placing the galss plate back on after making changes to the board or installing. But after that its fine.

I have a 2 gang that has just ghost switched for the first time in 3 days since install.
I have checked the capacitor on the board c6 and its 28uf. I dont have any others to test, so ill have to wait.

But when it ghost switches, it will only beep, then double beep. they dont turn on or off the relay.

Just a note i tried to load the latest dev firmware from the hackbox, but after doing so. my rules dont work.
currently i have the switches setup as generic, so i can use rules. as rules wont work with the stock t1 1,2,3 gang templates.
they also didn't work when i used a generic setup on the latest dev firmware, so i have gone back to 6.5.0.

its definitely a sensitivity issue. You may be onto something about a power supply issue and the tasmota firmware using more CPU.
Is there a way we can get a bin firmware specificaly for the t1, that disables anything not needed, like extra sensors. etc?

wookie666 on 16 Jun 2019

Ghost trigger with SetOption13 0 and SetOption1 1.

14:50:38 WIF: Connected
14:50:41 APP: Button2 multi-press 1
14:50:50 SRC: Button
14:50:50 MQT: sonoffhallwaykitchen/tele/STATE = {"Time":"2019-06-16T14:50:50","Uptime":"0T11:53:38","SleepMode":"Dynamic","Sleep":100,"LoadAvg":9,"POWER1":"OFF","POWER2":"ON","Wifi":{"AP":1,"SSId":"Skyenet","BSSId":"1C:F2:9A:CA:E9:AB","Channel":6,"RSSI":96,"LinkCount":1,"Downtime":"0T00:00:04"}}
14:50:50 MQT: sonoffhallwaykitchen/stat/RESULT = {"POWER2":"ON"}
14:50:50 MQT: sonoffhallwaykitchen/stat/POWER2 = ON (retained)
14:50:51 CFG: Saved to flash at FA, Count 108, Bytes 3584
14:50:58 WIF: Checking connection...
14:50:58 WIF: Connected
14:51:18 WIF: Checking connection...
14:51:18 WIF: Connected

kyemacdonald on 16 Jun 2019

And again. I think those options were a step in the wrong direction unfortunately.

15:04:08 SRC: Button
15:04:08 MQT: sonoffhallwaykitchen/tele/STATE = {"Time":"2019-06-16T15:04:08","Uptime":"0T12:06:56","SleepMode":"Dynamic","Sleep":100,"LoadAvg":9,"POWER1":"ON","POWER2":"ON","Wifi":{"AP":1,"SSId":"Skyenet","BSSId":"1C:F2:9A:CA:E9:AB","Channel":6,"RSSI":100,"LinkCount":1,"Downtime":"0T00:00:04"}}
15:04:08 MQT: sonoffhallwaykitchen/stat/RESULT = {"POWER1":"ON"}
15:04:08 MQT: sonoffhallwaykitchen/stat/POWER1 = ON (retained)
15:04:08 CFG: Saved to flash at F9, Count 109, Bytes 3584
15:04:08 SRC: MQTT

kyemacdonald on 16 Jun 2019

The sonoff-basic.bin firmware has all sensors disabled.

In addition you might want to disable the following:

timers (timers 0)
rules (rule1 0, rule2 0, rule3 0)
emulation (emulation 0)
webserver (webserver 0)
mqtt (setoption3 0)

arendst on 16 Jun 2019

Before I turn off the webserver, what is the best way to re-access the switches afterwards? Is it a case of needing to reset them via long button press?

kyemacdonald on 17 Jun 2019

update:

Sonoff-basic (latest) + buttondebounce 200 + new template + SetOption13 0 + SetOption1 1 = no ghost switching for 12 hours. Still getting the beeps though.

Looks promising but it's too early to celebrate. Will see.

sergiorademacher on 17 Jun 2019

I don't consistently see anything interesting preceding the rogue "APP: Button2 immediate".

I most often see this kind of thing:

08:36:41 WIF: Connected
08:37:02 WIF: Checking connection...
08:37:02 WIF: Connected
08:37:22 WIF: Checking connection...
08:37:22 WIF: Connected
08:37:42 WIF: Checking connection...
08:37:42 WIF: Connected
08:38:03 WIF: Checking connection...
08:38:03 WIF: Connected
08:38:24 WIF: Checking connection...
08:38:24 WIF: Connected
08:38:35 APP: Button2 immediate  // (sometimes Button1)
08:38:35 SRC: Button

I'll join in the testing now too, since I have 2x 3CH T1's and 2x 2CH T1's that do this (none of my T1's do).

alphaprawn on 17 Jun 2019

I happened to be right next to a switch when it ghosted. It actioned the state change before the single beep. Also during the time from the state change till the 2nd beep the switch was unresponsive on both button.

kyemacdonald on 17 Jun 2019

To be clear, could this possibly have anything to do with the ghost switching reportedly being fixed by adding a low pass filter to these boards?

alphaprawn on 17 Jun 2019

Mine does the same thing. But my single press action doesn't get actioned.
Only my long press from setoption32 and the rf remote pairing and delete

On Mon., 17 Jun. 2019, 7:39 pm kyemacdonald, notifications@github.com
wrote:

I happened to be right next to a switch when it ghosted. It actioned the
state change before the single beep. Also during the time from the state
change till the 2nd beep the switch was unresponsive on both button.

—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
https://github.com/arendst/Sonoff-Tasmota/issues/5449?email_source=notifications&email_token=AEDNT7EDRBUZD6XJ6R5BWU3P25LU5A5CNFSM4G5OGQN2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODX2T26I#issuecomment-502611321,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AEDNT7GRZVNLTNJ4E74RXPDP25LU5ANCNFSM4G5OGQNQ
.

wookie666 on 17 Jun 2019

For what it is worth, I reported false triggering to SONOFF. Their response was:

Press and hold the touch button of T1 until you hear two beep sounds. Press any button on the RF Remote. You’ll hear a beep sound when the code is cleared successfully.

I have no clue what this does... and do not have an RF remote. Possibly they are clearing code on a second micro (not ESP8266) embedded in the T1?

REALogics on 17 Jun 2019

They think you have an RF remote paired and it is causing interference.
They have no idea basically

On Mon., 17 Jun. 2019, 10:27 pm REALogics, notifications@github.com wrote:

For what it is worth, I reported false triggering to SONOFF. There
response was:

Press and hold the touch button of T1 until you hear two beep sounds.
Press any button on the RF Remote. You’ll hear a beep sound when the code
is cleared successfully.

I have no clue what this does... and do not have an RF remote. Possibly
they are clearing code on a second micro (not ESP8266) embedded in the T1?

—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
https://github.com/arendst/Sonoff-Tasmota/issues/5449?email_source=notifications&email_token=AEDNT7H2YWW354RRJBRJRDTP257K7A5CNFSM4G5OGQN2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODX3AEAY#issuecomment-502661635,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AEDNT7DAUSU5SAQ6VHGNIQ3P257K7ANCNFSM4G5OGQNQ
.

wookie666 on 17 Jun 2019

So far 36 hours with no ghost switching, only the annoying beeps.

sergiorademacher on 17 Jun 2019

Had a single T1 touch turn itself on last night. Funny enough it seems to
be one of the units that go into pairing mode when you put the glass cover
plate back on. Maybe it's too sensitive.

On Tue., 18 Jun. 2019, 3:34 am srademac, notifications@github.com wrote:

So far 36 hours with no ghost switching, only the annoying beeps.

—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
https://github.com/arendst/Sonoff-Tasmota/issues/5449?email_source=notifications&email_token=AEDNT7H7YVO3OXAEY5AKIL3P27DIRA5CNFSM4G5OGQN2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODX342KA#issuecomment-502779176,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AEDNT7FDMOLQEK745DPSEG3P27DIRANCNFSM4G5OGQNQ
.

wookie666 on 17 Jun 2019

I have had 3 ghosting switch events occur since changing to the new suggested configuration on the test T1 3 gang. Nothing on the stock sonoff.

justflashit on 17 Jun 2019

Seems as though its an issue on stock itead firmware also. So i would bet its mostly hardware.

How are those who changed the c6 capacitor on the rear faring now?

https://support.itead.cc/support/discussions/topics/11000025142

wookie666 on 17 Jun 2019

I have the experience with the T1 1 Gang. I had 2 of the T1 gang that did the ghosting already.

I flashed 6 of the T1 1 Gang and all of them showed the ghosting when I was busy with the configuring of the software and thought was something that is was connected with the usb flashing tool.

saboaua on 17 Jun 2019

Seems like people are seeing a fix with a low pass filter, but only if
using the right cap. See night reavers last comment of 20 days ago In #2192.

On Tue, 18 Jun 2019 at 8:52 AM, saboaua notifications@github.com wrote:

I have the experience with the T1 1 Gang. I had 2 of the T1 gang that did
the ghosting already.

I flashed 6 of the T1 1 Gang and all of them showed the ghosting when I
was busy with the configuring of the software and thought was something
that is was connected with the usb flashing tool.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
https://github.com/arendst/Sonoff-Tasmota/issues/5449?email_source=notifications&email_token=AMIBUYDHOFL2HUYK37WSWU3P272QRA5CNFSM4G5OGQN2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODX4NHJQ#issuecomment-502846374,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AMIBUYHB3BKFE3RKFB54NVTP272QRANCNFSM4G5OGQNQ
.

alphaprawn on 17 Jun 2019

Anyway, not sure if there's anything that can actually be done about it in tasmota.

alphaprawn on 17 Jun 2019

C6 cap to 16nf and debounce to 200ms still working on my side. No ghost switching for at least a week now.

Bloodyagent on 17 Jun 2019

With that setup do you see a change is the sensitivity of the touch ? Also do you have to hold the switch longer to register a touch?

wookie666 on 17 Jun 2019

Yes and yes, only slightly, but definitely can feel the difference.

Bloodyagent on 17 Jun 2019

@Bloodyagent have you heard any beeps in the past week ?

justflashit on 17 Jun 2019

Just looked at the board. I'm so not confident swapping that c6 capacitor....

kyemacdonald on 18 Jun 2019

Ok. Took 4 days, but my stock non flashed 3 channel switch just ghost switched.

So at this point my working theory has to be a hardware issue.....

kyemacdonald on 18 Jun 2019

Since I've set up to "SetOption13 1" and "ButtonDebounce 200" and "SetOption32 100" the ghost switch has gone. About 3 days without it.
No hardware intervention was performed.

I also changed some settings in my environment (Home Assistant + Mosquitto mqtt broker).

Added in /etc/mosquitto/mosquitto.conf:

persistence false

And followed the instructions about the cleansession parameter. See in: https://mosquitto.org/man/mosquitto-conf-5.html

I'll buying some others sonoffs T1 US to cover my house.

ericalvim on 18 Jun 2019

Just tried SetOption13 1" , "ButtonDebounce 200" and "SetOption32 100". Still getting the beeps.

Edit: and also ghost switching.

sergiorademacher on 18 Jun 2019

Yes, seems that the interference is both high and low freq at times.
@Bloodyagent did you also add a resistor along with the cap?

alphaprawn on 18 Jun 2019

Ok I have just replaced C6 on my test T1 3gang with a 16nF (2 x 8nF ceramic caps).
Still running dev basic with the above template.

justflashit on 18 Jun 2019

@alphaprawn. No resistor just replaced 22nf surface mount c6 with discreet 16nf.

Bloodyagent on 18 Jun 2019

@justflashit. No beeps for at least a week now since I changed debounce to 200ms. I only have 1 x 3ch running at the moment. Have 2 more 3ch units that I need to connect and test. I also have 3 x 2ch and 3x 1ch units whuch are running fine with no hardware or software mod.

Bloodyagent on 18 Jun 2019

Just tried SetOption13 1" , "ButtonDebounce 200" and "SetOption32 100". Still getting the beeps.

Edit: and also ghost switching.

After almost 4 days without any event here, a ghost switch triggered. Grrrrrrr!
I think it's not just a software solution. ☹

ericalvim on 18 Jun 2019

😕1

I have 5 sonoff basics, none have ever ghost switched, I have 14 Sonoff T1 US switches, a mix of one, two and three gangs. Its always the same ones that ghost switch, not all of them do it. I have all the software suggested fixes applied. I am wondering if its not defective devices, since only some of them do it. Maybe interference from high voltage circuit, in the defective devices (If it is a defect - in hardware)

Collin7 on 18 Jun 2019

Make sure to clear all RF codes in T1 US-3C

wongnam on 18 Jun 2019

How do you clear the RF Codes?
Out of the 14 T1 switches, two of the 3 gang and 2 of the one gang's constantly ghost switch. All the other two gang and one gangs are not ghost switching at all and never have. So four out of 14 ghost switch. All are running the same version or Tasmota, all have the same MQTT persistence settings according to the recommendations.
I am using openhab, it appears the ghost switching in not sending MQTT message. Openhab does not detect the state change.

Collin7 on 18 Jun 2019

Tocuh wood. I uploaded sonoff-basic.bin 6.5.0 to the affected switches
that have ghost switched

Setoption13 1
Setoption1 1
Switchdebounce 100
Buttondebounce 100
Setoption32 4
Timers 0

And no ghost switching in over 24 hours.

On Wed., 19 Jun. 2019, 1:54 am Collin, notifications@github.com wrote:

I have 5 sonoff basics, none have ever ghost switched, I have 14 Sonoff T1
US switches, a mix of one, two and three gangs. Its always the same ones
that ghost switch, not all of them do it. I have all the software suggested
fixes applied. I am wondering if its not defective devices, since only some
of them do it. Maybe interference from high voltage circuit, in the
defective devices (If it is a defect - in hardware)

—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
https://github.com/arendst/Sonoff-Tasmota/issues/5449?email_source=notifications&email_token=AEDNT7GJOCU3HCIAEWOZT5TP3EAMBA5CNFSM4G5OGQN2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODX7DFXI#issuecomment-503198429,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AEDNT7BEDY425XR76IOM2N3P3EAMBANCNFSM4G5OGQNQ
.

wookie666 on 18 Jun 2019

👀1

Out of desperation, cut the RF antenna. Didn't work. Pulled out all the
sonoffs in the house late at night, couldn't stand it anymore.

On Wed, Jun 19, 2019 at 7:45 AM wookie666 notifications@github.com wrote:

Tocuh wood. I uploaded sonoff-basic.bin 6.5.0 to the affected switches
that have ghost switched

Setoption13 1
Setoption1 1
Switchdebounce 100
Buttondebounce 100
Setoption32 4
Timers 0

And no ghost switching in over 24 hours.

On Wed., 19 Jun. 2019, 1:54 am Collin, notifications@github.com wrote:

I have 5 sonoff basics, none have ever ghost switched, I have 14 Sonoff
T1
US switches, a mix of one, two and three gangs. Its always the same ones
that ghost switch, not all of them do it. I have all the software
suggested
fixes applied. I am wondering if its not defective devices, since only
some
of them do it. Maybe interference from high voltage circuit, in the
defective devices (If it is a defect - in hardware)

—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
<
https://github.com/arendst/Sonoff-Tasmota/issues/5449?email_source=notifications&email_token=AEDNT7GJOCU3HCIAEWOZT5TP3EAMBA5CNFSM4G5OGQN2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODX7DFXI#issuecomment-503198429
,
or mute the thread
<
https://github.com/notifications/unsubscribe-auth/AEDNT7BEDY425XR76IOM2N3P3EAMBANCNFSM4G5OGQNQ

.

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/arendst/Sonoff-Tasmota/issues/5449?email_source=notifications&email_token=AMIBUYAU7CZVBCFWH6VMZ6DP3E3L7A5CNFSM4G5OGQN2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODX7X4KQ#issuecomment-503283242,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AMIBUYA2YKCGRSNU2DMZC4LP3E3L7ANCNFSM4G5OGQNQ
.

alphaprawn on 18 Jun 2019

😕1

Ok I am 99% sure that this is a hardware issue now and that there is no way for tasmota to stop the beeps. I have 3 unflashed switches wired up and they are demonstrating the beeping behaviour at similar frequency to my flashed switches.

As such @arendst would it be possible to add an option to tasmota to have a long press button ignore? This would have to be mixed with setoption13 0 to function and the buttondebounce setting.

But something along the lines of setoptionxx 10 (Number of 0.1 seconds to hold button before input is ignored). Have the default set to 0 = disabled.

This would mean that the switches will still beep, but the state of the switches wouldn't change on button presses exceeding setoptionxx time. But any switch input between the buttondebounce setting and the setoptionxx time would action the switch.

I strongly believe this problem is due to poor manufacturing control, likely around the capacitance of the c6 capacitor, mixed with the design specified capacitance of the c6 being too close to the line. I see different switching rates between my switches but eventually all of them will demonstrate the behaviour. I have not done @Bloodyagent mod, but I have reduced the frequency of the ghosts by layering electrical tape on the inside of the glass, but then I lose the pretty glow :P

Tasmota flashed devices:
3ch Master Bedroom - Ghost frequency ~36hr
3ch Kitchen - Ghost frequency ~24hr
2ch Lounge - Ghost Frequency ~6 hr
2ch Hall 1 - Ghost Frequency ~ 4hr
2ch Hall 2 - Ghost Frequency ~ 48hr
1ch Hall 3 - Ghost frequency - not yet observed - install 4 days
1ch Dining - Ghost Frequency - 24hr
1ch Stair - Ghost frequency - 24 hr

Stock unflashed devices. These are much harder to catch because they don't change state only beep. So I have them setup in my home office where I am within earshot for about 12hrs a day. In total I would have been near them approx 36hrs.
3ch - caught twice
2ch - caught once
2ch - caught once

kyemacdonald on 19 Jun 2019

👍2

I tried with no luck, I´m still getting beeps. So far no ghosting though.

Tocuh wood. I uploaded sonoff-basic.bin 6.5.0 to the affected switches that have ghost switched Setoption13 1 Setoption1 1 Switchdebounce 100 Buttondebounce 100 Setoption32 4 Timers 0 And no ghost switching in over 24 hours.
…

sergiorademacher on 19 Jun 2019

You'll get beeps. That's part of the unit itself. Nothing to do with
tasmota. But the settings should stop single presses from being triggered.
Long press will still be triggered depending on setoption32 settings.

On Wed., 19 Jun. 2019, 11:39 am srademac, notifications@github.com wrote:

I tried with no luck, I´m still getting beeps. So far no ghosting though.

Tocuh wood. I uploaded sonoff-basic.bin 6.5.0 to the affected switches
that have ghost switched Setoption13 1 Setoption1 1 Switchdebounce 100
Buttondebounce 100 Setoption32 4 Timers 0 And no ghost switching in over 24
hours.
… <#m_-6938698976107866656_>

—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
https://github.com/arendst/Sonoff-Tasmota/issues/5449?email_source=notifications&email_token=AEDNT7DHFJJHC4ZPPXCPIILP3GE33A5CNFSM4G5OGQN2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODYANSSI#issuecomment-503372105,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AEDNT7EJ2VA3WNMNOVI5MI3P3GE33ANCNFSM4G5OGQNQ
.

wookie666 on 19 Jun 2019

As @kyemacdonald suggested I will introduce an option along with SetOption13 0 to disable hold functionality.

This will trigger the relay on touch RELEASE and will not act when active for a longer period of time.

arendst on 19 Jun 2019

👍2

That will prevent false triggers. For those getting the beeps. Just remove
the buzzer

On Wed., 19 Jun. 2019, 7:32 pm Theo Arends, notifications@github.com
wrote:

As @kyemacdonald https://github.com/kyemacdonald suggested I will
introduce an option along with SetOption13 0 to disable hold
functionality.

This will trigger the relay on touch RELEASE and will not act when active
for a longer period of time.

—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
https://github.com/arendst/Sonoff-Tasmota/issues/5449?email_source=notifications&email_token=AEDNT7EIFM2QVYL7BKW6KNLP3H4MHA5CNFSM4G5OGQN2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODYBI7UY#issuecomment-503484371,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AEDNT7GUHXALDMMIGN6XV73P3H4MHANCNFSM4G5OGQNQ
.

wookie666 on 19 Jun 2019

How do you clear the RF Codes?
Out of the 14 T1 switches, two of the 3 gang and 2 of the one gang's constantly ghost switch. All the other two gang and one gangs are not ghost switching at all and never have. So four out of 14 ghost switch. All are running the same version or Tasmota, all have the same MQTT persistence settings according to the recommendations.
I am using openhab, it appears the ghost switching in not sending MQTT message. Openhab does not detect the state change.

If it happens on your switch that sounds like hardware, please remove the glass mask to see if it is still there, I suspect the electrostatic glass surface causes random switching.

BTW to clear 315MHz codes:

Press and hold the touch button of T1 until you hear two beep sounds. Press any button on the RF Remote. You’ll hear a beep sound when the code is cleared successfully.

wongnam on 19 Jun 2019

Thank you @wongnam
I do not have the remotes for these units.
I will test by removing the glass and see what that does.

Collin7 on 19 Jun 2019

@Collin7 With the glass panels removed the switches won't ghost. As soon as you put them back though the problem will return.

kyemacdonald on 19 Jun 2019

With the latest fix it provides command SetOption40 0..250 allowing 0 (disabled) or up to 250 * 0.1 second hold button control.

When used together with SetOption1 1 and SetOption13 0 and set as SetOption40 10 it will disable any button action if held down longer than 1 second.

arendst on 19 Jun 2019

👍1

@arendst If so, does it have a reset function when pressed 40 seconds?

wongnam on 19 Jun 2019

No as it will activate when the ghost button is active.

To get Reset over 40 seconds functionality you'll have to set SetOption1 0 disabling the max hold control functionality. If you manage to descriminate a ghost button press from a reset or a toggle let me know ;-)

arendst on 19 Jun 2019

The option to configure a valid button press won't work, unfortunately.

From what I could see with the DSO, the issue is the noise produced by the button itself that cannot be distinguished from a button press. That leads to totally random (both time and duration) button presses. For me, this was further enhanced when the ESP board is off the socket and exposed to all of the workbench interferences.

My question regarding stock Sonoff firmware is if anyone knows if it's doing some special signalling for button SNR calibration.

jpmrodrigues on 19 Jun 2019

Thanks @arendst - flashing now.

@jpmrodrigues - the stock switches demonstrate the same behaviour

kyemacdonald on 19 Jun 2019

Thanks @arendst! Trying with the new firmware and settings. Will report back.

sergiorademacher on 19 Jun 2019

Where I can find the last update? http://thehackbox.org/tasmota/release/sonoff.bin? To stop beep I cut off speaker from sonoff board.

mpkoslovsky on 19 Jun 2019

http://thehackbox.org/tasmota/

arendst on 19 Jun 2019

Morning @arendst

Is it possible to adjust the behaviour of the new option? Currently it switches after waiting the setoption40 time period. eg button press, finger removed, wait, state change.

My current settings are:
buttondebounce 40
setoption1 1
setoption13 0
setoption40 10

kyemacdonald on 19 Jun 2019

Hallelujah!!!! @arendst the new option 40 solves the switching problem. Still get the beeps but that's an easy fix.

Log output of a switch that tried to ghost.
01:30:28 WIF: Connected
01:30:48 WIF: Checking connection...
01:30:48 WIF: Connected
01:30:51 APP: Button1 multi-press 1
01:31:08 WIF: Checking connection...
01:31:08 WIF: Connected
01:31:28 WIF: Checking connection...
01:31:28 WIF: Connected

kyemacdonald on 20 Jun 2019

🎉1

Just lowered the setoption40 to 1 to try to increase the response time of the switch. This didn't seem to change how long it took to respond. It's a solid second after finger removal to change state.

vid demo - https://youtu.be/4LHlVOQngWg

buttondebounce is currently set to 40

kyemacdonald on 20 Jun 2019

14 hours, not ghosting here either! Just the annoying beep. Thank you @arendst!!!

@kyemacdonald in my case it takes about 700ms since I touch the plate to activate the relay. I don´t have the behavior of your video.

I´m using sonoff.bin (not the basic) and my settings are: backlog SetOption1 1 ; SetOption13 0 ; SetOption40 10; Buttondebounce 40; setoption32 40; Switchdebounce 40;

Now what would be the best approach to get rid the beep? changing the cap or removing the speaker?

sergiorademacher on 20 Jun 2019

Hello @sergiodemacher I cut off speaker from sonoff. It’s easy to resolve. Now I configured my 3 switches ch 3 like @kyemacdonald posted and I will wait for this night.
Thanks

mpkoslovsky on 20 Jun 2019

@kyemacdonald are you sure it switches after SetOption40 time? To chk raise it to SetOption40 40 and see when it reacts.

In my situation (T1 year old) it reacts as soon as I lift my finger and that is what I would expect from the tasmota code. I also cannot reproduce your SetOption40 1 behaviour as the time is too small to lift my finger before 0.1 second.

Pls remember the T1's button detection circuit consists of the touch sensor (or RF received signal) -> microcontroller (sonoff firmware) -> esp8266 (tasmota). I think the microcontroller firmware of your T1's is different from my old T1.

arendst on 20 Jun 2019

Flashed back to sonoff.bin and reset to factory defaults. Switching time is now heaps better.

kyemacdonald on 20 Jun 2019

thanks @mpkoslovsky! Unfortunately I have to report that I´m still getting ghosts. Maybe the config is messed up with so many changes. I will reset to factory defaults and see how it goes.

sergiorademacher on 20 Jun 2019

Thank you @arendst
So far so good, updated 3 switches to test, since yesterday has not ghosted once.
Im having an issue updating today though, using http://thehackbox.org/tasmota/ as I did yesterday, but after the switch restarts it still shows tasmota release version. Two of the switches I cant access the web interface anymore. I may need to just take them out and flash them again

Collin7 on 20 Jun 2019

Just wanted to say a huge thankyou @arendst

Not had a single ghost on any switches even though I've had plenty of beeps.

kyemacdonald on 20 Jun 2019

Thx. So I guess we can close this one.

arendst on 20 Jun 2019

@sergiorademacher only one switched was in this last night. I will continue monitoring. Thanks again

mpkoslovsky on 20 Jun 2019

👍1

I factory reset the offending switches, reapplied the fix and zero ghost switching. Thank you!!!

sergiorademacher on 21 Jun 2019

👍1

I've got 2x Sonoff T1 (2gang EU) and both of them worked fine for 3 months (installed them at the same time) but one of them started acting up recently. Had it 2 times going ON without anyone touching the sensor.

Last time I was close and saw the status LED was constantly blinking, which indicated that it had no WiFi connection. During that period it had a 'ghost switch' event. Also the uptime was <3h and restart reason was 'external system'. I had to restart my AP in order for it to connect again.

Did anyone also notice a low uptime and this restart reason after a 'ghost switch' event?

And what does 'external system' mean? A power supply outage or something? Maybe my PSU is going bad?

Software: Tasmota 6.3 with Core 2.4.2 SDK 2.2.1

TimelessNL on 16 Jul 2019

@TimelessNL - Please upgrade your device(s). Fixes specifically for the T1 were added in 6.6.

meingraham on 16 Jul 2019

👍1

Yeah, will do that asap.

But can anyone explain what restart reason 'external system' means? I've not switched of the mains or anything so if it's a power issue the PSU may be going bad :/

TimelessNL on 16 Jul 2019

Hello,
I'm here to seek some help as well. Experiencing ghost switching on couple of Sonoff Touch (1C), some of them are ok some of them turns self on occasionally.
(and no it's not because of MQTT :) )

here's one of them (log level 4)
setting: SetOption1 1 ; SetOption13 0 ; SetOption40 10
tasmota 6.6.0.3

12:36:06 WIF: Connected
12:36:15 APP: Button1 multi-press 1
12:36:26 WIF: Checking connection...
12:36:26 WIF: Connected
12:36:46 WIF: Checking connection...
12:36:46 WIF: Connected
12:37:06 WIF: Checking connection...
12:37:06 WIF: Connected
12:37:07 MQT: home/dressroom/light/tele/STATE = {"Time":"2019-08-03T12:37:07","Epoch":1564828627,"Uptime":"0T15:00:20","UptimeSec":54020,"Heap":32,"SleepMode":"Dynamic","Sleep":50,"LoadAvg":19,"POWER":"OFF","Wifi":{"AP":1,"SSId":"muhehe","BSSId":"XE:XE:XB:X7:XF:AA","Channel":11,"RSSI":80,"LinkCount":4,"Downtime":"0T00:00:26"}}
12:37:26 WIF: Checking connection...
12:37:26 WIF: Connected
12:37:46 WIF: Checking connection...
12:37:46 WIF: Connected
12:38:06 WIF: Checking connection...
12:38:06 WIF: Connected
12:38:26 WIF: Checking connection...
12:38:26 WIF: Connected
12:38:46 WIF: Checking connection...
12:38:46 WIF: Connected
12:39:06 WIF: Checking connection...
12:39:06 WIF: Connected
12:39:24 APP: Button1 multi-press 1
12:39:25 SRC: Button
12:39:25 MQT: home/dressroom/light/tele/STATE = {"Time":"2019-08-03T12:39:25","Epoch":1564828765,"Uptime":"0T15:02:38","UptimeSec":54158,"Heap":32,"SleepMode":"Dynamic","Sleep":50,"LoadAvg":19,"POWER":"ON","Wifi":{"AP":1,"SSId":"muhehe","BSSId":"XE:XE:XB:X7:XF:AA","Channel":11,"RSSI":76,"LinkCount":4,"Downtime":"0T00:00:26"}}
12:39:25 MQT: home/dressroom/light/stat/RESULT = {"POWER":"ON"}
12:39:25 MQT: home/dressroom/light/stat/POWER = ON (retained)

Those which behaves ok, has got default setting and setoption13 1, but this one is ghostswitching with 13 1 and 13 0 and 40 10 (but must say, it's better with opt40 10)
Should I higher opt40 a bit?

kriznik on 3 Aug 2019

@kriznik

Please address this to the Tasmota Support Discord Chat. The chat is a better and more dynamic channel for helping you. Github issues are best used for Tasmota _software feature requests and bug reporting_. Troubleshooting and setup assistance is more effective using an interactive forum.

Please check the Contributing Guideline and Policy and the Support Guide.

Thanks.

Support Information

See Wiki for more information.
See Chat for more user experience.
See Community for forum.
See Code of Conduct

meingraham on 3 Aug 2019

Hello,

I have no additional insights or solutions to offer, but perhaps the info below is useful for someone.

I have been fighting sporadic (once every few days) ghost switches on a modified Tuya curtain switch. It has three touch buttons, up, down and stop. Only the down button shows ghosting behavior. It has a BS813-A-1 touch button chip on it.

I am going tot try the SetOption40 solution. I am not sure if the ghost switches manifest themselves as long presses. Is there a way to see that in the logs?

Best.

Edit: Immediately after setting Setoption40 it ghost switched :-)
I realize now that SetOption40 only works for a button, but I am using switches.
Is there a similar option for switch?

MarcG046 on 7 Aug 2019

You can view log messages in the web UI Console or you can set up a remote syslog server. You can also increase the logging level by setting Weblog/Syslog . The normal level is to but you can incrementally go up to 5 as desired.

meingraham on 7 Aug 2019

I have log at level 4 (the maximum).
It shows that the switch was activated at 10.03:06, but I cannot see how long the sensor signal went low for, so I don't know if a feature like SetOption40 for a switch would solve the problem.

Thanks.
Marc

Aug 7 10:02:48 rolluikvoor-4727 ESP-WIF: Controleer verbinding...
Aug 7 10:02:48 rolluikvoor-4727 ESP-WIF: Verbonden
Aug 7 10:03:05 rolluikvoor-4727 ESP-MQT: huis/schakelaars/rolluikvoor/stat/RESULT = {"POWER2":"OFF"}
Aug 7 10:03:05 rolluikvoor-4727 ESP-MQT: huis/schakelaars/rolluikvoor/stat/POWER2 = OFF
Aug 7 10:03:06 rolluikvoor-4727 ESP-SRC: Switch
Aug 7 10:03:06 rolluikvoor-4727 ESP-MQT: huis/schakelaars/rolluikvoor/stat/RESULT = {"POWER1":"ON"}
Aug 7 10:03:06 rolluikvoor-4727 ESP-MQT: huis/schakelaars/rolluikvoor/stat/POWER1 = ON
Aug 7 10:03:06 rolluikvoor-4727 ESP-CFG: Opgeslagen in flash op F9, Aantal 219, Bytes 3584
Aug 7 10:03:08 rolluikvoor-4727 ESP-WIF: Controleer verbinding...
Aug 7 10:03:08 rolluikvoor-4727 ESP-WIF: Verbonden

MarcG046 on 7 Aug 2019

What is the reason you have assigned the buttons in the device as switches?

Have you tried altering the debounce settings?

meingraham on 7 Aug 2019

Good question. Frankly I don't exactly understand how they differ.
(I use it in switch mode 3 and I have SwitchDebounce set to 203. I have tried different debounce values, but it does not seem to make a difference)
The description of their behavior on the Wiki seems identical.
https://github.com/arendst/Sonoff-Tasmota/wiki/Buttons-and-Switches

I have configured it as a button and set option 40 to 2sec. Experiment is running.

MarcG046 on 7 Aug 2019

It did not work. Two more ghost switches today.
I will try to measure the duration of the ghost button presses.

MarcG046 on 8 Aug 2019

@MarcG046

What is the Tuya device?

Please post output from the following commands. When you paste the results in your reply, please enclose it in triple backtick marks `
Status 0
Template
GPIO

meingraham on 8 Aug 2019

It's this one:
https://www.amazon.de/LoraTap-Zeitschaltuhr-Rolladenzeitschaltuhr-Rolladenmotor-Kompatibel/dp/B07L8TF88C

However, I have removed the ESP that was inside it, because I did not know how to flash it.
Instead I connected all electronics inside it to an external WEMOS.

19:37:54 CMD: Status 0
19:37:54 MQT: huis/schakelaars/rolluik2/stat/STATUS = {"Status":{"Module":18,"FriendlyName":["Sonoff","Sonoff2","Sonoff3"],"Topic":"rolluik2","ButtonTopic":"0","Power":1,"PowerOnState":3,"LedState":1,"LedMask":"FFFF","SaveData":1,"SaveState":1,"SwitchTopic":"0","SwitchMode":[3,3,3,0,0,0,0,0],"ButtonRetain":0,"SwitchRetain":0,"SensorRetain":0,"PowerRetain":0}}
19:37:54 MQT: huis/schakelaars/rolluik2/stat/STATUS1 = {"StatusPRM":{"Baudrate":115200,"GroupTopic":"sonoffs","OtaUrl":"http://thehackbox.org/tasmota/release/sonoff-NL.bin","RestartReason":"Software/System restart","Uptime":"0T00:11:49","StartupUTC":"2019-08-08T18:26:05","Sleep":50,"CfgHolder":4617,"BootCount":31,"SaveCount":681,"SaveAddress":"F4000"}}
19:37:54 MQT: huis/schakelaars/rolluik2/stat/STATUS2 = {"StatusFWR":{"Version":"6.6.0.4(sonoff)","BuildDateTime":"2019-08-08T20:20:22","Boot":31,"Core":"2_5_2","SDK":"2.2.2-dev(c0eb301)"}}
19:37:54 MQT: huis/schakelaars/rolluik2/stat/STATUS3 = {"StatusLOG":{"SerialLog":0,"WebLog":2,"SysLog":4,"LogHost":"192.168.2.107","LogPort":514,"SSId":["airport express","Buiten"],"TelePeriod":300,"Resolution":"558180C0","SetOption":["0000C00B","280500000100060014005A00000000000000","00000000"]}}
19:37:54 MQT: huis/schakelaars/rolluik2/stat/STATUS4 = {"StatusMEM":{"ProgramSize":555,"Free":448,"Heap":21,"ProgramFlashSize":1024,"FlashSize":4096,"FlashChipId":"1640EF","FlashMode":3,"Features":["00000413","8FDAE397","043683A0","22B617CD","01001BC0","00000000"]}}
19:37:54 MQT: huis/schakelaars/rolluik2/stat/STATUS5 = {"StatusNET":{"Hostname":"rolluik2-5040","IPAddress":"192.168.2.131","Gateway":"192.168.2.1","Subnetmask":"255.255.255.0","DNSServer":"192.168.2.1","Mac":"CC:50:E3:55:D3:B0","Webserver":2,"WifiConfig":4}}
19:37:54 MQT: huis/schakelaars/rolluik2/stat/STATUS6 = {"StatusMQT":{"MqttHost":"192.168.2.107","MqttPort":1883,"MqttClientMask":"DVES_%06X","MqttClient":"DVES_55D3B0","MqttUser":"mqtt","MqttCount":1,"MAX_PACKET_SIZE":1000,"KEEPALIVE":30}}
19:37:54 MQT: huis/schakelaars/rolluik2/stat/STATUS7 = {"StatusTIM":{"UTC":"Thu Aug 08 18:37:54 2019","Local":"Thu Aug 08 19:37:54 2019","StartDST":"Sun Mar 31 02:00:00 2019","EndDST":"Sun Oct 27 03:00:00 2019","Timezone":"+01:00","Sunrise":"05:33","Sunset":"20:18"}}
19:37:54 MQT: huis/schakelaars/rolluik2/stat/STATUS10 = {"StatusSNS":{"Time":"2019-08-08T19:37:54","Epoch":1565289474}}
19:37:54 MQT: huis/schakelaars/rolluik2/stat/STATUS11 = {"StatusSTS":{"Time":"2019-08-08T19:37:54","Epoch":1565289474,"Uptime":"0T00:11:49","UptimeSec":709,"Heap":21,"SleepMode":"Dynamic","Sleep":50,"LoadAvg":19,"MqttCount":1,"POWER1":"ON","POWER2":"OFF","POWER3":"OFF","Wifi":{"AP":1,"SSId":"airport express","BSSId":"B4:18:D1:E2:6A:AE","Channel":11,"RSSI":80,"LinkCount":1,"Downtime":"0T00:00:04"}}}
19:38:30 CMD: Template
19:38:30 MQT: huis/schakelaars/rolluik2/stat/RESULT = {"NAME":"Generic","GPIO":[255,255,255,255,255,255,255,255,255,255,255,255,255],"FLAG":15,"BASE":18}
19:38:43 CMD: GPIO
19:38:43 MQT: huis/schakelaars/rolluik2/stat/RESULT = {"GPIO0":"0 (Geen)","GPIO1":"0 (Geen)","GPIO2":"0 (Geen)","GPIO3":"0 (Geen)","GPIO4":"91 (Button2n)","GPIO5":"21 (Relais1)","GPIO12":"22 (Relais2)","GPIO13":"92 (Button3n)","GPIO14":"23 (Relais3)","GPIO15":"0 (Geen)","GPIO16":"90 (Button1n)"}

MarcG046 on 8 Aug 2019

I connected all electronics inside it to an external WEMOS - wow! that's drastic.

I think all bets are off since the code may also be looking at other signals besides just the buttons and relays. And since you only have those connected to the D1, the logic is not being triggered since the D1 is itself not "misbehaving".

I'm not sure if you can restore the original Tuya module. That should be easy enough to flash... especially if it can be done before re-inserting into the device.

meingraham on 8 Aug 2019

I only take the three outputs of the touch keys out and the two drivers of the relays back in and I take the 3V3 power supply from the switch.

I instrumented the code to report the duration of the button signal.
I realized that the button debounce is different from the switch debounce, because the switch debounce ending with 1,2,3 includes the filtering, which is not there for the button. I hope that is not giving me new ghost switches...

MarcG046 on 8 Aug 2019

@MarcG046 I'm signing off at this point. I think your modification has rendered the ability to troubleshoot this effectively practically impossible... at least for me.

meingraham on 9 Aug 2019

@meingraham Of course. I understand.
Thanks for your support.

MarcG046 on 9 Aug 2019

Hi there!

To the original poster @Bloodyagent and @beppeserafini... thanks a lot!!

But... is there any minimum chance you could possibly post detailed pictures or instructions for the less digital-electronic-handy ones??? Schematics, videos, shopping lists, etc, etc.

On the other hand, has anyone tried the new wall switches??

https://www.itead.cc/sonoff-tx-series-wifi-smart-wall-switches.html

Does the problem persist?

Finally, is there any software way to disable RF receiving on the T1 switches?

Thanks again!

Cheers,

MO

mvorchard on 10 Aug 2019

small update from my side.
after upgrading to 6.6.0.3 (2.5.2 core) from 6.5 (2.3 core) I've experienced ghostswitching on switches which has been working before without any issues and without setoption40 10

So I kind of think it's not directly HW but SW smthing.

I'm not exactly familiar what is that core for, but it might be worth to reflash all ghosting switches with older code to see if it is still happening (i'll give it a shot)

kriznik on 16 Aug 2019

@kriznik

I think it would be worth giving the pre-2.6 Core a try. You will have to compile your own firmware but this is relatively straightforward. But to update, you will have to load sonoff-minimal and then the new binary in a two-step manual process using the file upload OTA method rather than the web OTA method.

meingraham on 16 Aug 2019

I have my images under 480Kb and for my switches even less, so I guess I dont need to go with sonoff-minimal

thanks for info, will try to compile actual, if there is any difference in behaviour

kriznik on 16 Aug 2019

👍1

I just bought three 3 gang US T1's and flashed them with Tasmota and am having the ghost switch issue on all three of them and found this thread as I was researching it. I downgraded to 6.3.0 and it is still doing it. I read through everything and it doesn't look like it but, just to be sure, there is no known solution yet, correct?

tjdavisbz on 16 Aug 2019

@meingraham silly question, where do i find pre-2.6? in current dev I can see 2.5.2 still

kriznik on 19 Aug 2019

@kriznik You will have to compile your own firmware. See here

meingraham on 19 Aug 2019

yes indeed.. i know how to make own bin ;), but current source doesnt seems to have 2.6 , all i can see is 2.5.2 in the dev branch?

kriznik on 19 Aug 2019

4. Open the platformio.ini file located in the /Sonoff-Tasmota root directory (scroll to the bottom of the file Explorer pane)
5. In the Editor (2) pane, a leading semicolon ; comments out (disables) a statement:
   In the [common] section, enable -DUSE_CONFIG_OVERRIDE if using your own user_config_overrid.h 
   In the [core_active] section, enable the platform and build_flags lines for the desired core

"enable the platform and build_flags lines for the desired core"

meingraham on 19 Aug 2019

yes... i know

[core_active]
; Select one core set for platform and build_flags
;platform = ${core_2_3_0.platform}
;build_flags = ${core_2_3_0.build_flags}
;platform = ${core_2_4_2.platform}
;build_flags = ${core_2_4_2.build_flags}
platform = ${core_2_5_2.platform}
build_flags = ${core_2_5_2.build_flags}
;platform = ${core_stage.platform}
;build_flags = ${core_stage.build_flags}
;platform = ${core_pre.platform}
;build_flags = ${core_pre.build_flags}

so shall I assume "core_pre" is 2.6?

kriznik on 19 Aug 2019

giving the pre-2.6 Core a try - the 'pre' in pre-2.6

STAGE and Pre are the development branches of the next version the ESP Core of. Pre is a known working code freeze version. Stage is the development version that is updated every day until the release date. 2.5.2 is the currently released _master_ Core. 2.6 is the next Core under _development_ and is due to be released on September 2, 2019

meingraham on 19 Aug 2019

cool, thanks for an explanation. Will give it a try

kriznik on 20 Aug 2019

running new version with pre_core, so I'll report if that behaviour is improved or not.

strangely it says I'm running STAGE even when I do have pre uncommented

Core/SDK Version | STAGE/2.2.2-dev(38a443e)

;platform                  = ${core_stage.platform}
;build_flags               = ${core_stage.build_flags}
platform                  = ${core_pre.platform}
build_flags               = ${core_pre.build_flags}

is that correct behaviour?

kriznik on 20 Aug 2019

@kriznik

I have the same experience, if I select pre, it still says STAGE. I have also tried STAGE directly (it still says STAGE). Both Core codebase result in good connectivity so I'm not sweating it :wink:

Mike

meingraham on 20 Aug 2019

update: it seems pre-2.6 to be even worse than on 2.5.2 it's ghosting every roughly every 10min.
will go with 2.3 core

kriznik on 20 Aug 2019

OK, there's that test checked off. Surprising.

What kind of RSSI levels does this device show?

meingraham on 20 Aug 2019

Just to make sure I understand - is there no ghosting at all with 2.3 core?

tjdavisbz on 20 Aug 2019

@meingraham
it's connected without issues, over 85% signal ... and it goes always like this

23:12:27 WIF: Checking connection...
23:12:27 WIF: Connected
23:12:47 WIF: Checking connection...
23:12:47 WIF: Connected
23:12:51 APP: Button1 multi-press 1
23:12:52 SRC: Button
23:12:52 MQT: home/dressroom/light/stat/RESULT = {"POWER":"ON"}

but sadly that's all what level 4 debug log can offer ... ;(
this is with pre-2.6 core, setoption40 10, buttondebounce 150, setoption13 0

on core 2.5.2 with same settings it's happening like once per day, on pre-2.6 it's way more often

it's as well fairly regular

00:37:28 WIF: Connected
00:37:48 WIF: Checking connection...
00:37:48 WIF: Connected
00:37:48 APP: Button1 multi-press 1
00:38:08 WIF: Checking connection...
00:38:08 WIF: Connected
00:38:28 WIF: Checking connection...
00:38:28 WIF: Connected
00:38:48 WIF: Checking connection...
00:38:48 WIF: Connected
00:39:08 WIF: Checking connection...
00:39:08 WIF: Connected
00:39:28 WIF: Checking connection...
00:39:28 WIF: Connected
00:39:44 APP: Button1 multi-press 1
00:39:48 WIF: Checking connection...
00:39:48 WIF: Connected

kriznik on 20 Aug 2019

Just to make sure I understand - is there no ghosting at all with 2.3 core or is that still being tested?

tjdavisbz on 22 Aug 2019

update:
on 2.5.2 it looks like these values:

backlog setoption1 1; setoption13 0; setoption40 8; buttondebounce 175

works.
with setoption40 10 it was ghostswitching once-twice per hour, with 8 it is not (second day) yet

kriznik on 22 Aug 2019

Awesome! I’m thinking I may be a ghost on this thread because I have asked a couple of simple questions with no response.

Anyways, I’ll try those settings and report back.

tjdavisbz on 22 Aug 2019

I do not want to be rude but pls stop adding to this issue. It's closed in March/April with the simple solution of concluding it's a hardware issue.

Pls read this issue from top down. All your questions and findings have long been answered.

Issue closed. Good luck.

arendst on 22 Aug 2019

So how are we supposed to communicate about possible findings if we don't comment here? kriznik has a possible fix, has worked for him for two days, is he just supposed to keep that to himself since the issue is closed? Not trying to be rude either but just seems ridiculous to not share if we have further findings or a question. Also - I don't see an answer to my question and I have read the entire thread several times. Would it be that hard to just answer my question?

tjdavisbz on 23 Aug 2019

@tjdavisbz the problem exists in the switches even on the stock firmware. The difference between the stock one and tasmota is the stock ones have some code to handle the ghost switching, which has been effectively replicated with the setoption40.

A stock switch will beep just as regularly as a tasmota switch and will also eventually trigger to a state change given a long enough time.

Unfortunately there is no software solution. What I have done is fill all the beepers with hot glue and run setoption40 10 on all the switches.

kyemacdonald on 23 Aug 2019

Got it, thank you very much. I had seen mixed feedback on whether or not it was a hardware issue. Thanks again for taking the time to clarify.

tjdavisbz on 23 Aug 2019

(Just my 2 cents)

My solution was to put the IC PCB as close to the front cover as possible. Barely holding on the pinheader connections and then press the front cover on.

This way the buttons are much more sensitive to finger presses and the ghost switching issue is almost gone since then.

TimelessNL on 23 Aug 2019

I don't want to be rude, but you are wrong.

It is simply not working in longterm, even with those settings I've reported it is ghostswitching at the end and with those up in this topic, it's not working either.

So TL;DR: there is NO reliable solution yet for Tasmota.

kriznik on 23 Aug 2019

@kriznik the ghost switching / beeps occur on stock unmodified switches.

Sonoff released a product with a hardware problem with some hacky software to mask it. But the problem still exists.

At best tasmota can have a similar hacky solution. But it will never be perfect.

kyemacdonald on 23 Aug 2019

We have problems with sonoff t1, it's a hardware problem. I bought sonoff t2 a new version of t1. When this switch arrives I will test and post the result about switch ghosting.

mpkoslovsky on 24 Aug 2019

My Sonoff T3 3CH US is Ghost Switching, did anyone else have problems with T3 models?

felipenanoclick on 20 Sep 2019

O have ghost switching with T1 us 3ch. After removing from welink, It stopped. So i Think Its a software problem. Dont have tasmota.

serralta on 20 Sep 2019

My Sonoff T3 3CH US is Ghost Switching, did anyone else have problems with T3 models?

it was a problem with the house power cables, I switched places and the ghost stopped

felipenanoclick on 25 Sep 2019

Hello Felipe, how do you fix or change these cables. Could give me an example?

Thanks
Michel

mpkoslovsky on 25 Sep 2019

Hello Felipe, how do you fix or change these cables. Could give me an example?

Thanks
Michel

I didn't change the wires, I changed the sonoff touch room, but the ghost came back this morning

felipenanoclick on 25 Sep 2019

Hello Felipe, how do you fix or change these cables. Could give me an example?
Thanks
Michel

I didn't change the wires, I changed the sonoff touch room, but the ghost came back this morning

I will buy the touch from zemismart

felipenanoclick on 25 Sep 2019

Hello All, I had the same problem with my sonoff T1 3CH. when started all was OK and few time after start to change state without needed... After putting a scope on PIO0 I found that ramdomly (apparently) some oscillations comes on GPIO0 and according to level, switch the 1st relay. After investigation I've found that to flash it, I have solder the GPIO0 cable at a place very very closed to another connection. No contact between the 2 but according to something: humidity presure or other, the residual grease become resistive !!!! After cleaning, no more problems.....

LaurentG91 on 19 Oct 2019

👍2

Thanks for this post!

This is great information for users. Could you edit the flashing guide in the wiki to add your findings? It would be greatly appreciated.

Mike

meingraham on 19 Oct 2019

Who has experience with replacing the C6 capacitor?
I replaced it by a 15nf capacitor, but had no luck. It is switching a few times a day.
Now I replaced it with a 18nf and will test it the next days....

Moskito99 on 21 Oct 2019

I'm with the same problem with 2 new Sonoff T1 US 3ch flashed with tasmota 6.6.0
The Sonoff T1 2ch don't have the problem.

lucianocf on 22 Oct 2019

Hi all,

Just one more example to the pool! :)

I've flashed two Sonoff T1 1CH last week or so, and both of them are experiencing ghost switching every now and then.
Before the flashing, they worked for 1 year without a single (detected) false switch.
I have no HA server, and only "talk" with the devices by http (post and websend), no mqtt whasoever.
The ghost switch is for on and off alike.
Tasmota version is 6.6.
Changed the locations of the switchs and nothing changed.
Changed buttondebounce to 60, 80, 100, 120 (actual) and no consistent change in ocurrences or interval.
I still haven't tried the setoption40 10.

rnogueiragh on 23 Oct 2019

Guys this is a hardware issue with the T1s. The switching will occur with an unflashed switch if you leave it long enough.

All that can be done is to try and mitigate it with the setoption40.

It sucks balls, but honestly there isn't anything that can be done short of modding the hardware in some way.

kyemacdonald on 24 Oct 2019

I agree with @kyemacdonald , I have tried everything other than hardware mod. I have tried leaving the glass cover off for about 2 days, on the switches that ghost the most, and I had no ghosting with the glass cover off. After adding the setoption40 command, it does not stop the ghosting, but it does prevent the light from switching on or off. The switch will just do its three beeps (which is annoying). Without removing the buzzer, I wish there was an option to disable it. Just to note setoption40 is not 100% fix. I still have a light here and there that switch on, during a ghost, perhaps once in two weeks roughly. I have 15 Sonoff T1's. One, Two and Three Channels, its always the same ones that ghost. not all of them, Also points to it being hardware issue.
I have one question, has anyone tried any of the new Sonoffs Tx range? Is this still an issue?

Collin7 on 24 Oct 2019

Sorry I cannot speak for TX range but I have had the T1 US 3 gang switch with ghosting and got feed up with them.
I went and bought the T2 US 3 gang switches. Put tasmota on them and no ghosting at all for 6 weeks. I now have 5 T2 US 3 gang running without issues.

justflashit on 24 Oct 2019

haven’t seen this idea yet:

touch switching is an analog thing and thus could be sensitive to power issues.

Anyone tried this on a bench with stable external power supply (without mains please), or with wifi completely switched off, or with aggressive sleep mode settings?

joba-1 on 24 Oct 2019

So does this means that the T2 are bullet proof?
Any other difference between T1 and T2 (as far as tasmota goes)?

I have now almost 2 full days with no ghost switching with both T1, but with a configuration than almost turns de manual switch unusable...

rnogueiragh on 24 Oct 2019

If it is a hardware problem, why i started have ghost swtiches after latest firmware and if i unlink from ewelink, it dont happen?

serralta on 24 Oct 2019

supports (but not proving) my power hypothesis

joba-1 on 24 Oct 2019

@joba-1 if it is dirty power then the switches are not equally sensitive. I have a dozen across my house, some never switch, others regularly do.

I just filled all the buzzers with hotglue, and then setoption40 5 pretty much sorts the rest.

kyemacdonald on 25 Oct 2019

that is the problem with analog: little differences (e.g. due to part tolerances) can result in different behaviour.

joba-1 on 25 Oct 2019

So, replaced the capacitor c6 with a 18nf, no luck!
Bought a T2EU1C and T2EU2C and both are ghost switching with tasmota and not with original firmware. I still think that tasmota is the problem with sonoff devices.
Iit is not usable for me with this issue! :-(

I'm looking now for an alternative for the sonoff switches.
If someone have infos, please share!

Moskito99 on 26 Oct 2019

@Moskito99 I found this thread while researching ghost switching of my own. I am seeing it with Sonoff devices, but I also have a pair of Etekcity ESWL01 devices that haven't shown this problem. Just fyi.

jp-bennett on 27 Oct 2019

@Moskito99 - I hooked up 2 unflashed stock switches to lamps and just let them sit there on the floor of my office. Across a 3 day period they both did the beep cycle without switching multiple times and then eventually both switched without the beeps.

kyemacdonald on 28 Oct 2019

i have anothers T1 US and work great., but I get a new t1 us 3 bang and fist button flicking all time, only first buttom ... I think this sensible is high! Try Setoption 40 but get christim light either! hahaahahah Need Help too!

djgel on 28 Oct 2019

@djgel - switching all the time? Ghost switching is usually 4hrs to 3 days or so. If it is happening a lot more than that I suspect you have another issue.

kyemacdonald on 29 Oct 2019

This is what i get from um T1 in a usual day:

30/10/2019 13:29:53 : Switch 0
30/10/2019 13:11:12 : Switch 1
30/10/2019 12:11:59 : Switch 0
30/10/2019 11:30:13 : Switch 1
30/10/2019 09:08:06 : Switch 0
30/10/2019 07:28:34 : Switch 1
30/10/2019 07:14:52 : Switch 0
30/10/2019 07:14:21 : Switch 1
30/10/2019 07:04:44 : Switch 0
30/10/2019 05:14:19 : Switch 1
30/10/2019 04:02:39 : Switch 0
30/10/2019 03:59:43 : Switch 1
30/10/2019 01:52:07 : Switch 0
30/10/2019 00:55:32 : Switch 1
30/10/2019 00:52:17 : Switch 0
29/10/2019 23:27:23 : Switch 1
29/10/2019 23:03:18 : Switch 0
29/10/2019 22:35:50 : Switch 1
29/10/2019 21:06:02 : Switch 0
29/10/2019 20:10:15 : Switch 1
29/10/2019 19:45:25 : Switch 0
29/10/2019 19:35:41 : Switch 1
29/10/2019 17:15:03 : Switch 0
29/10/2019 16:55:18 : Switch 1
29/10/2019 16:51:05 : Switch 0
29/10/2019 16:43:16 : Switch 1
29/10/2019 15:45:47 : Switch 0
29/10/2019 15:28:17 : Switch 1
29/10/2019 13:08:25 : Switch 0

I can see no pattern and i've tried every configuration i could think of...
The worst part is that now i have 2 worthless switches that i cannot use (and cant reflash to the original state). This cannot be an isolated thing!
And even if this is a defect in some equipments, the original Sonoff "software" seems to be able to mitigate this.
Is there anything else that i could do?

rnogueiragh on 30 Oct 2019

I've decided to put the sonoff T1 away, and bought a couple o TX 1 and 2 gangs, (T2EU1C and T2EU2C), and before flashing it i couldn't even register them with the ewelink app (something about my wifi network???), so i put the tasmota firmware (6.6.0.20) and everything is working very nicely, with no ghost switching whatsoever, in the exactly same spots the old ones were.

So i'll leave them for now, and hope that they behave accordingly, hoping that one day, another tasmota version could address the ghost switching problems of the old T1...

rnogueiragh on 12 Nov 2019

I have 3CH touch ITEAD switch (works the same as sonoff, but other brand) with original firmware 3.3.0, worked perfect for almost 2 years. Yesterday moved switch to different Wifi. First reset wifi: hold 2 times 7 seconds on the buttons to enter pair mode. After that (after 3 hours finding out how to pair) finally paired to the new WiFi. But from the moment I did reset WiFi, the switch started randomly on/off the lights (even without WiFi, so no broker or whatever). Tried to reset and pair again, nothing helps. Disconnected eWelink from Google Home as well from Home Assistant. But again, ghost switching even without WiFi connection. Any ideas? Hardware issue? Or switch keeps some states/events on his own hardware?

dimitriy-k on 26 Nov 2019

If you are running the firmware that came on the switch from the manufacturer (i.e., not Tasmota) then you need to inquire on their support forum. This forum is only for Tasmota.

meingraham on 26 Nov 2019

I experience the ghost switching on every Sonoff T1 R2 EU. Does not matter if it is covered or uncovered. They are new.

Did anybody solve the issue?

18:28:04 RSL: tele/touch-switch-1/STATE = {"Time":"2019-12-05T18:28:04","Uptime":"0T19:50:07","UptimeSec":71407,"Heap":27,"SleepMode":"Dynamic","Sleep":50,"LoadAvg":19,"MqttCount":0,"POWER":"OFF","Wifi":{"AP":1,"SSId":"UPCDD9D128","BSSId":"E8:94:F6:D4:44:C5","Channel":11,"RSSI":100,"LinkCount":1,"Downtime":"0T00:00:06"}}
18:28:24 WIF: Checking connection...
18:28:24 WIF: Connected
18:28:29 APP: Button1 multi-press 1
18:28:29 SRC: Button
18:28:29 RSL: stat/touch-switch-1/RESULT = {"POWER":"ON"}
18:28:29 RSL: stat/touch-switch-1/POWER = ON
18:28:29 CFG: Saved to flash at F7, Count 573, Bytes 4096
18:28:44 WIF: Checking connection...
18:28:44 WIF: Connected
18:28:59 APP: Button1 multi-press 1
18:29:04 WIF: Checking connection...
18:29:04 WIF: Connected
18:29:05 SRC: Button
18:29:05 RSL: stat/touch-switch-1/RESULT = {"POWER":"OFF"}
18:29:05 RSL: stat/touch-switch-1/POWER = OFF
18:29:05 CFG: Saved to flash at F6, Count 574, Bytes 4096
18:29:24 WIF: Checking connection...
18:29:24 WIF: Connected
18:29:29 APP: Button1 multi-press 1
18:29:29 SRC: Button
18:29:29 RSL: stat/touch-switch-1/RESULT = {"POWER":"ON"}
18:29:29 RSL: stat/touch-switch-1/POWER = ON
18:29:29 CFG: Saved to flash at F5, Count 575, Bytes 4096
18:29:44 WIF: Checking connection...
18:29:44 WIF: Connected
18:30:04 WIF: Checking connection...
18:30:04 WIF: Connected
18:30:24 WIF: Checking connection...
18:30:24 WIF: Connected
18:30:44 WIF: Checking connection...
18:30:44 WIF: Connected
18:31:04 WIF: Checking connection...
18:31:04 WIF: Connected
18:31:06 APP: Button1 multi-press 1
18:31:09 SRC: Button
18:31:09 RSL: stat/touch-switch-1/RESULT = {"POWER":"OFF"}
18:31:09 RSL: stat/touch-switch-1/POWER = OFF
18:31:09 CFG: Saved to flash at F4, Count 576, Bytes 4096
18:31:24 WIF: Checking connection...
18:31:24 WIF: Connected
18:31:44 WIF: Checking connection...
18:31:44 WIF: Connected
18:32:04 WIF: Checking connection...
18:32:04 WIF: Connected
18:32:24 WIF: Checking connection...
18:32:24 WIF: Connected
18:32:44 WIF: Checking connection...
18:32:44 WIF: Connected
18:33:04 WIF: Checking connection...
18:33:04 WIF: Connected
18:33:04 RSL: tele/touch-switch-1/STATE = {"Time":"2019-12-05T18:33:04","Uptime":"0T19:55:07","UptimeSec":71707,"Heap":27,"SleepMode":"Dynamic","Sleep":50,"LoadAvg":19,"MqttCount":0,"POWER":"OFF","Wifi":{"AP":1,"SSId":"UPCDD9D128","BSSId":"E8:94:F6:D4:44:C5","Channel":11,"RSSI":98,"LinkCount":1,"Downtime":"0T00:00:06"}}
18:33:24 WIF: Checking connection...
18:33:24 WIF: Connected
18:33:44 WIF: Checking connection...
18:33:44 WIF: Connected
18:34:04 WIF: Checking connection...
18:34:04 WIF: Connected
18:34:24 WIF: Checking connection...
18:34:24 WIF: Connected
18:34:29 APP: Button1 multi-press 1
18:34:41 SRC: Button
18:34:41 RSL: stat/touch-switch-1/RESULT = {"POWER":"ON"}
18:34:41 RSL: stat/touch-switch-1/POWER = ON
18:34:41 CFG: Saved to flash at FB, Count 577, Bytes 4096
18:34:44 WIF: Checking connection...
18:34:44 WIF: Connected
18:35:04 WIF: Checking connection...
18:35:04 WIF: Connected
18:35:10 APP: Button1 multi-press 1
18:35:11 SRC: Button
18:35:11 RSL: stat/touch-switch-1/RESULT = {"POWER":"OFF"}
18:35:11 RSL: stat/touch-switch-1/POWER = OFF
18:35:11 CFG: Saved to flash at FA, Count 578, Bytes 4096
18:35:24 WIF: Checking connection...
18:35:24 WIF: Connected
18:35:32 APP: Button1 multi-press 1
18:35:33 SRC: Button
18:35:33 RSL: stat/touch-switch-1/RESULT = {"POWER":"ON"}
18:35:33 RSL: stat/touch-switch-1/POWER = ON
18:35:33 CFG: Saved to flash at F9, Count 579, Bytes 4096
18:35:44 WIF: Checking connection...
18:35:44 WIF: Connected
18:36:04 WIF: Checking connection...
18:36:04 WIF: Connected
18:36:24 WIF: Checking connection...
18:36:24 WIF: Connected
18:36:44 WIF: Checking connection...
18:36:44 WIF: Connected
18:37:01 APP: Button1 multi-press 1
18:37:04 WIF: Checking connection...
18:37:04 WIF: Connected
18:37:12 SRC: Button
18:37:12 RSL: stat/touch-switch-1/RESULT = {"POWER":"OFF"}
18:37:12 RSL: stat/touch-switch-1/POWER = OFF
18:37:12 CFG: Saved to flash at F8, Count 580, Bytes 4096
18:37:24 WIF: Checking connection...
18:37:24 WIF: Connected
18:37:44 WIF: Checking connection...
18:37:44 WIF: Connected

jardous on 5 Dec 2019

After updating all devices to firmware 8.1 I forgot what is ghost switching.

mpkoslovsky on 3 Mar 2020

After updating all devices to firmware 8.1 I forgot what is ghost switching.

Really? Need any aditional configuration? I'll try to upgrade one of my devices to check.

lucianocf on 3 Mar 2020

For now (2 days straight), i have the same results with the newest firmware (8.1). No ghost switching.
But I'll wait a little more to be sure… :)

rnogueiragh on 3 Mar 2020

No, no need any changes. I'm with more than 10 days without ghost switching.

mpkoslovsky on 3 Mar 2020

tasmota 8.11, bye bye ghost!

felipenanoclick on 17 Mar 2020

I can also confirm, 8.1 seems to do the trick. After setOption40 came out, I had the rare ghost that the light would actually come one, but mostly my switches would do the 3 beeps, multiple times a day. Its been about 2 weeks now with tasmota 8.1, and not a single ghosting incident or beeping incident. Thank you @arendst

Collin7 on 17 Mar 2020

My Sonoff T2 3x Gang switches which were ghost switching several times a day. V8.1 has solved it, not a single ghost in weeks!

REALogics on 17 Mar 2020

Ok. Upgraded one of my switches to 8.1 that usually ghosts switches. Will now see what happens over the next few weeks. Can i ask though, does anybody know what change in 8.1 has fixed the issue specifically.

wookie666 on 17 Mar 2020

Pushing update to all my switches right now! Oh please! I have a pile of switches not installed because of this issue.

kyemacdonald on 18 Mar 2020

@arendst

5 days and 14 switches - no ghost switches on 8.1

kyemacdonald on 22 Mar 2020

Confirmed. 20 days without ghost switching. Version 8.1 solved completely. Thanks everyone.

lucianocf on 23 Mar 2020

Ok. Upgraded one of my switches to 8.1 that usually ghosts switches. Will now see what happens over the next few weeks. Can i ask though, does anybody know what change in 8.1 has fixed the issue specifically.

I also like to know that

Fusseldieb on 15 Jun 2020

Tasmota: Sonoff T1 3CH US Ghost Switching

Most helpful comment

All 323 comments

Support Information

Related issues