How to disable and enable the 1Wire bus (M503)
-
Hi,
How can i reset (disable and enable) 1-wire bus in Neuron M503?
Vesa
-
At the moment this isn't possible via the evok API, as the 1Wire bus device is common to both Neuron and UniPi 1.1 (and resetting the 1Wire bus is not possible on the latter). We are considering adding this option to evok but it will require some changes to the way 1Wire is handled.
It is possible to reset the bus by using the internal Modbus server, but this is somewhat insecure, non-trivial and requires editing systemd scripts.
-
@tomas_knot said in How to disable and enable the 1Wire bus (M503):
It is possible to reset the bus by using the internal Modbus server, but this is somewhat insecure, non-trivial and requires editing systemd scripts.
Insecure is my middle name :D
So can you tell me how that can be done with Modbus server?
Vesa
-
You will need to edit the following line in /etc/systemd/system/neurontcp.service:
ExecStart=/opt/neuron-bin/neuron_tcp_server -p 502Change it to say:
ExecStart=/opt/neuron-bin/neuron_tcp_server -p 502 -l 0.0.0.0instead.
Afterwards you will need to send a Modbus write of 1 to coil 1001 and then another write of 0 to coil 1001. This can be done via a short python script using PyModbus, or through another TCP Modbus library, of which there is a wide variety. The Modbus server resides on port 502 with the ip of your Neuron (or 127.0.0.1 if you are doing it locally on the device). The Python script can then be run for some given condition.
-
It worked, thank you!
In my case:
ExecStart=/opt/neurontcp/neuron_tcp_server -p 502 -l 0.0.0.0So...this maneuver should fix any problems with 1-wire sensors if the fault is not on the connector(s) or the 8 port hub?
Vesa
-
Hopefully yes, but it really shouldn't be necessary, it's unfortunate you have to use it.
We'll try to add it to EVOK soon.
-
Hi,
I just ran into the same problem as Vesa (one wire bus suddenly lost) and I can confirm that this procedure worked, but I had to reboot the rpi afterwards.
I wonder if the change to /etc/systemd/system/neurontcp.service is really necessary, I initially tried polling the server without modifying the file and it seemed to work anyway.
If someone else needs this, the only cli tool to poll the modbus server I found functioning on the rpi is this one:
https://datapacket.dl.sourceforge.net/project/mblogic/MBLogic-All/MBLogic_All_Release-32/mblogic_all_2011-04-16.zipYou will have to set the line endings in the file to linux style, before running:
./pollmb.py -p 502 -u 1 -f 5 -a 1001 -q 1 -d 1
./pollmb.py -p 502 -u 1 -f 5 -a 1001 -q 1 -d 0 -
@rob
It listens on 127.0.0.1 by default. The tool is written by us (you can find the source code at http://git.unipi.technology:3000/UniPi/neuron-modbus-tools, though note that not all of our software is opensourced) so I am quite confident it is necessary. It is of course not necessary if the client software is run locally via 127.0.0.1.That said, the Linux libmodbus library did contain a bug where this setting was not applied correctly and ignored instead. Somewhat humorously I believe that the bug is still there in the version used on the main Debian repository (we statically link against a newer version of the modbus library instead, which does not have this issue), and as a result quite a few IoT devices likely unintentionally listen on all IPs.
(Incidentally this issue was uncovered by DavidCozens, to give proper credit)
-
Hi @tomasknot
I'll come back to this topic again. Is it possible to do the same for Modbus in M503? I mean reset the bus (disable/enable)
Br,
Vesa -
Hello @vesa-kauppinen,
yes, check our Modbus Registers Maps: https://kb.unipi.technology/files:products:02-neuronFor M503 it is coil 1001 - Enable/Disable 1Wire Bus
Best regards,
Martin -
Hi,
That's what we've done with 1-Wire and it works.
I mean, is it possible to do this same 'reset' for Modbus (RTU) itself? We sometimes lose a modbus device.
Br,
Vesa -
Oh, sorry for the misunderstanding @vesa-kauppinen.
We have a functionality called Master Watch Dog (MWD). When enabled, the underlying HW will watch when the last communication happend. If it will be longer than the preset time, it will reboot itself. Now, this can be set per group in the PLC, and it will watch if the ModbusTCP server communicates with the given group. If not, the board will reboot. Similarly, this can be set on the extensions as well. If nothing communicates with the extension over ModbusRTU, the extension reboots itself.
This can be set via modbus. Look for the "Enable MWD" coil and "MWD timeout" register. The timeout is in miliseconds.
Best regards,
Martin -
-
Sorry, I've never used the pollmb.py. Would be the mbpoll example sufficient? You can install the mbpoll from our repository:
apt update && apt install mbpollMartin
-
-
For Group 1 on M503
Verify the status of the MWD (if 0 - MWD disable, if 1 - MWD enabled but never rebooted, if 2 - MWD enabled and already rebooted):
mbpoll -m tcp -a 0 127.0.0.1 -0 -1 -r 6Set the MWD timeout to 5000ms:
mbpoll -m tcp -a 0 127.0.0.1 -0 -r 1008 5000Enable the MWD:
mbpoll -m tcp -a 0 127.0.0.1 -0 -r 6 1Write the settings to the NVRAM:
mbpoll -m tcp -a 0 127.0.0.1 -0 -t 0 -r 1003 1Verify the MWD by stopping the UniPi Modbus TCP server:
sudo systemctl stop unipitcpThe both lines of LEDs should blink for 1sec after the MWD timeout
Disable the MWD:
mbpoll -m tcp -a 0 127.0.0.1 -0 -r 6 0Write the settings to the NVRAM:
mbpoll -m tcp -a 0 127.0.0.1 -0 -t 0 -r 1003 1For Group 2 on M503
Use the MWD control register 114 instead of 6
Use the MWD timeout register 1108 instead of 1008
Use the Save configuration coil 1103 instead of 1003 -
Thanks for the good instructions! After enabling MWD, setting timeout and writing the settings:
mbpoll -m tcp -a 0 127.0.0.1 -0 -1 -r 6 mbpoll 1.2-6 - FieldTalk(tm) Modbus(R) Master Simulator Copyright (c) 2015 epsilonRT, All rights reserved. This software is governed by the CeCILL license <http://www.cecill.info> Protocol configuration: Modbus TCP Slave configuration...: address = [0] start reference = 6, count = 1 Communication.........: 127.0.0.1, port 502, t/o 1.00 s, poll rate 1000 ms Data type.............: 16-bit register, output (holding) register table -- Polling slave 0... [6]: 3...or...
mbpoll -m tcp -a 0 127.0.0.1 -0 -1 -r 114 mbpoll 1.2-6 - FieldTalk(tm) Modbus(R) Master Simulator Copyright (c) 2015 epsilonRT, All rights reserved. This software is governed by the CeCILL license <http://www.cecill.info> Protocol configuration: Modbus TCP Slave configuration...: address = [0] start reference = 114, count = 1 Communication.........: 127.0.0.1, port 502, t/o 1.00 s, poll rate 1000 ms Data type.............: 16-bit register, output (holding) register table -- Polling slave 0... [114]: 3What does that result (3) mean? Its not 0, 1 or 2.
Btw
sudo systemctl stop unipitcpshould be
sudo systemctl stop neurontcpin Neuron M503...right??
Br,
Vesa -
Hi @vesa-kauppinen,
sorry the value 3 is 1 + 2, which means board has MWD enabled and the MWD already rebooted the board.The service is called unipitcp, but in older packages it was called neurontcp. The service is the same for Neurons and Axons, so we renamed it to unipitcp to make it more clear.
Best regards,
Martin -
Aaaaa...of course :D
But what about this: either 11 or 3
root@residevice:~# mbpoll -m tcp -a 0 127.0.0.1 -0 -1 -r 6 mbpoll 1.2-6 - FieldTalk(tm) Modbus(R) Master Simulator Copyright (c) 2015 epsilonRT, All rights reserved. This software is governed by the CeCILL license <http://www.cecill.info> Protocol configuration: Modbus TCP Slave configuration...: address = [0] start reference = 6, count = 1 Communication.........: 127.0.0.1, port 502, t/o 1.00 s, poll rate 1000 ms Data type.............: 16-bit register, output (holding) register table -- Polling slave 0... [6]: 11 root@residevice:~# mbpoll -m tcp -a 0 127.0.0.1 -0 -1 -r 6 mbpoll 1.2-6 - FieldTalk(tm) Modbus(R) Master Simulator Copyright (c) 2015 epsilonRT, All rights reserved. This software is governed by the CeCILL license <http://www.cecill.info> Protocol configuration: Modbus TCP Slave configuration...: address = [0] start reference = 6, count = 1 Communication.........: 127.0.0.1, port 502, t/o 1.00 s, poll rate 1000 ms Data type.............: 16-bit register, output (holding) register table -- Polling slave 0... [6]: 3 root@residevice:~# mbpoll -m tcp -a 0 127.0.0.1 -0 -1 -r 6 mbpoll 1.2-6 - FieldTalk(tm) Modbus(R) Master Simulator Copyright (c) 2015 epsilonRT, All rights reserved. This software is governed by the CeCILL license <http://www.cecill.info> Protocol configuration: Modbus TCP Slave configuration...: address = [0] start reference = 6, count = 1 Communication.........: 127.0.0.1, port 502, t/o 1.00 s, poll rate 1000 ms Data type.............: 16-bit register, output (holding) register table -- Polling slave 0... [6]: 11Br,
Vesa -
Ok, this is a little bit tricky. The register has 16bits, and some of the bits represent some state or function.
Bit 0: Set or read the status of the MWD
Bit 1: The MWD rebooted the boarde
other bits contain some debug information, which we haven't documented yetSo value 11 in decimal is 0000 0000 0000 1011 in binary. The bit no. 0 (first from the right) is 1, which means MWD is enabled. The bit no. 1 (second from the right) is 1, which means MWD already rebooted the board. Other bits don't apply to you.
You can "mask" the lowest 2 bits with boolean operation AND, typically with bit mask in hexadecimal representation:
register_6 && 0x0003
Or with binary mask:
register_6 && 0b11
I hope this helps you.
Martin
