Vanishing 1-wire sensors



  • @jasperjackson said in Vanishing 1-wire sensors:

    @tomas_hora looks like disabeling the rtc kernel module fixed the problem. Things are still running perfectly :)

    Too early....

    unfortunatly i got another problem now after 3 days of operation. The problem is not the one wire sensors. I resetted the 1-Wirebus by disconnecting all the sensors. I then did a softreset of the raspberry/unipi, no sensors where found. It turns out there is a problem with the DS2482.

    
    [  273.108220] Driver for 1-wire Dallas network protocol.
    [  279.979503] ds2482 1-0018: DS2482 reset status 0x00 - not a DS2482
    
    

    Only a hardreset of the unipi resolves the problem. I'm running out of options here.. any help would be appreciated. I'll mount the rtc battery for now and see if the problems will go away. The main problem is that the unipi is controlling my heating in Austria and I live in holland. I'll probably need a second unipi to reset the main one :(.

    Regards,

    Jasper


  • administrators

    @jasperjackson If you have a free relay, you can try to wire the Data and 5V (or only GND) through a relay, and when it the sensors vanishes, restart the whole bus by the relay



  • @tomas_hora

    Hi Thomas!
    Thanks for your reply!

    When i disconnect the 1-Wire bus and reconnect it, (even when i give the unipi a softreset with no sensors connected before reconnecting the sensors again), the problem will not go away.

    It seems different than the problems other users reported. In their cases a hard reset of only the 1-Wire bus solves the problem, thus, putting it on a Relay makes it possible to remotely reset.
    In this case a hardreset of the 1-Wire bus does not resolve the problem.
    Could it be a problem with the I2C? The kernel driver reports that the DS2482 is not a DS2482. So it seems to me that the problem is I2C related, or the DS2484 hangs for some reason.
    Can I Analyse this somehow? I don't understand the i2cdetect command very well, maybe you can point me in the right direction here.

    I Have 2 Unipies by the way, one with shitty unshielded telephone cables through my whole house for the 1-Wire (Works flawlessly) and the one with CAT-5 FTP, (causing troubles).

    The Unipi is remotely accessible through SSH, so i can try almost everything remotely.

    Regards,

    Jasper Jackson



  • @jasperjackson said in Vanishing 1-wire sensors:

    @tomas_hora

    Hi Thomas!
    Thanks for your reply!

    When i disconnect the 1-Wire bus and reconnect it, (even when i give the unipi a softreset with no sensors connected before reconnecting the sensors again), the problem will not go away.

    It seems different than the problems other users reported. In their cases a hard reset of only the 1-Wire bus solves the problem, thus, putting it on a Relay makes it possible to remotely reset.
    In this case a hardreset of the 1-Wire bus does not resolve the problem.
    Could it be a problem with the I2C? The kernel driver reports that the DS2482 is not a DS2482. So it seems to me that the problem is I2C related, or the DS2484 hangs for some reason.
    Can I Analyse this somehow? I don't understand the i2cdetect command very well, maybe you can point me in the right direction here.

    I Have 2 Unipies by the way, one with shitty unshielded telephone cables through my whole house for the 1-Wire (Works flawlessly) and the one with CAT-5 FTP, (causing troubles).

    The Unipi is remotely accessible through SSH, so i can try almost everything remotely.

    Regards,

    Jasper Jackson

    Did some more research, this is the output of the failing unipi

    root@unipi:~# i2cdetect -y 1
         0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f
    00:          03 04 05 06 07 08 09 0a 0b 0c 0d 0e 0f
    10: 10 11 12 13 14 15 16 17 18 19 1a 1b 1c 1d 1e 1f
    20: 20 21 22 23 24 25 26 27 28 29 2a 2b 2c 2d 2e 2f
    30: 30 31 32 33 34 35 36 37 38 39 3a 3b 3c 3d 3e 3f
    40: 40 41 42 43 44 45 46 47 48 49 4a 4b 4c 4d 4e 4f
    50: UU 51 52 53 54 55 56 UU 58 59 5a 5b 5c 5d 5e 5f
    60: 60 61 62 63 64 65 66 67 68 69 6a 6b 6c 6d 6e 6f
    70: 70 71 72 73 74 75 76 77
    

    And this is the output of my functioning unipi

    
    root@unipi:~# i2cdetect -y 1
         0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f
    00:          -- -- -- -- -- -- -- -- -- -- -- -- --
    10: -- -- -- -- -- -- -- -- UU -- -- -- -- -- -- --
    20: 20 -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
    30: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
    40: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
    50: UU -- -- -- -- -- -- UU -- -- -- -- -- -- -- --
    60: -- -- -- -- -- -- -- -- 68 -- -- -- -- -- -- UU
    70: -- -- -- -- -- -- -- --
    
    

    I have a hard time imagining that this has something to do with my 1-Wire setup.


  • administrators

    @jasperjackson Nice catch. Anyway, I am afraid we cannot help you much. It is probably caused by some EMC on the 1W or I2C. You can try to reset reload the I2C driver if it helps:

    sudo rmmod i2c_dev
    sudo rmmod i2c_bcm2708

    sudo modprobe i2c_bcm2708
    sudo modprobe i2c_dev



  • @tomas_hora
    Ok! Thanks for the reply!, i'll look for emc sources, and i'll change the unipi and the raspberry with my other one to rule out hardware failure.
    I'll keep you posted if i find anything.

    Resetting the driver didn't work.
    I read in some other forum that i2cdetect returning all addresses usually means that the SDA of the I2C is bound to ground somehow.
    Maybe there lays a problem.

    Regards,

    Jasper Jackson



  • @jasperjackson said in Vanishing 1-wire sensors:

    @tomas_hora

    Hi Thomas!

    I didn't find anything yet...
    Looking at the cabeling again...

    What surprises me by the way is that you sell 1-wire sensors with 10 meters of unshielded cable??

    https://www.unipi.technology/1-wire-temperature-sensor-p63

    How can that be right?



  • @tomas_hora

    So.. I finally got to the source of the problem. After a lot of measuring en swearing we discovered that under some conditions (rapid temperature change) some sensors kept the data line low to approx. 1.6V rendering the other sensors unable to communicate. It turned out that only a powerreset of the unipi board resolved the problem for some time what pointed me in the direction of the UNIPI board. Wrong! After some googleing i found the following blog:

    https://blog.peku33.net/aliexpress-ebay-china-market-fake-ds18b20-sensors/

    So, there you go... Lots work to find out that the DS18B20 sensors I got from AliExpress where fake ones and did exactly what was described in the Blog.

    I'll order new (genuine) ones soon.

    So for everyone experiencing these problems. Check where your sensors came from.

    Tanks for the help anyway ;)

    Regards,

    Jasper Jackson



  • Hello Jasper

    Thank's a lot for this hint. I'm searching since a while.
    I have exactly the same problem with my L503. I have 12pc. DS18B20 Sensors on the 1 Wire bus and always after a huge temp change of app. 60 deg. all Data of the Temp. Sensors are frozen. Rebooting the the Neuron dosen't help. Only disconnecting from the power source helps. I brought al my sensors direct from unipi. I will test them asap.

    Regards
    Stefan


  • administrators

    Hello @stoeffel,

    If this happens, you do not need to reboot the whole neuron, just reset the 1Wire bus using Modbus or OW_off variable in case of Mervis.

    Let us know what you find out.