how to write coils (function 0x05) #1271

Neukam · 2025-04-08T01:41:05Z

Neukam
Apr 8, 2025

Checklist

I made sure that there are no existing issues or discussions - open issues,closed issues,open discussion, closed discussion - which may answer my question or that I could contribute my information to.
I have read the Known Issues Wiki Page and my question / issue is not listed.
I have provided the version of the software and the operating system in my post.
If I have an issue with genmon not communicating with the controller, I have read and followed the serial troubleshooting guide.
I have included the generator make and model in my submission.

Version

Genmon Version V1.19.05
Ubuntu 24.04.2 LTS Proxmox LXC

Have you submitted your log files from the About page? Note that outbound email must be working for logs to be submitted.

I have not submitted logs

What is/are your question(s)?

I am creating a custom config for a SmartGen HGM4010NC controller that I have installed on my 21kW generator (Isuzu 4LE1 powered Allmand light tower). I have the RS485 connected to an EW11 wifi bridge. I have communication between Genmon and the controller and can see status and alarms. I would like to add buttons to set the generator's mode (Auto, Manual, Start, Stop, etc). If I understand correctly, buttons command_sequence uses function code 0x10 to write to holding registers. My controller needs function code 0x05 to write to coil registers to set the mode. Is this possible with genmon?

Attached is a modbus dump, a PDF of the HGM4000 series protocol documentation, and my config JSON (work in progress).

Additional information

modbusdump.txt
SmartGen_HGM4000.json
HGM4000N_Protocol_en.pdf

jgyates · 2025-04-08T11:24:45Z

jgyates
Apr 8, 2025
Maintainer

Hi @Neukam

I can add the ability to write coil registers. The reason this is not a functionaliy today is that I have yet to run into a controller that uses coils that can be written. The ones I have tested with only used coils as read only and reserved write functionality for registers or modbus file access.

I will test this later today, but so far you JSON file looks great. Give me a couple of days to add writing coil registers.

Let me know if you have any feedback on the JSON file documentation in the wiki as I would like it to improve over time to make it easier to add support for new controllers.

7 replies

Neukam Apr 8, 2025
Author

I initially set register 002a to a length of 4 and used that for hours, but it didn't give the result that I expected. Register 002a returns 0001 and register 002b returns 0901, so reading it as 4 bytes, you get 00010901 which is 67841 hours. I wish it was that simple...

I will work on the "disableoutagecheck" and let you know if I need further help.

Would it be better to add a default value to the config functionality, or put phase = 1 into genmon.conf as an initial placeholder? Is the phase parameter read automatically from certain controllers?

jgyates Apr 8, 2025
Maintainer

Adding a default value to the config functionality is pretty trivial. That is the current plan

I looked at the PDF you sent and it looks like they are not treating these two registers like a single 32 bit value like most controllers do, but the LSB is 0-9999 and the MSB is 0-9000. Due to this we will have to treat this as two values and use the "list" type to combine multiple registers into one value with a little bit of shifting ("shiftright" or "shiftleft" JSON directives) on the MSB. I can add this as well.

jgyates Apr 8, 2025
Maintainer

as a point of reference, how may hours do you have on your controller? I am trying to figure out how they are formating their MSB. I assume it is like this:

add 10,000 to the MSB, then add it to the LSB

1 in MSB would give you 10000, 901 hex in LSB would give 2305 so that would be 12305 hours if I am guessing right.

jgyates Apr 8, 2025
Maintainer

also, it looks like the registrers support minutes and seconds so I guess we can display hours min and seconds

Neukam Apr 8, 2025
Author

1 in MSB would give you 10000, 901 hex in LSB would give 2305 so that would be 12305 hours if I am guessing right.

That is correct, I have 12305 hrs on the controller.

also, it looks like the registrers support minutes and seconds so I guess we can display hours min and seconds

That is correct. There's also a pair of registers for the number of starts, in the same format as the runtime hours.

Neukam · 2025-04-10T00:43:32Z

Neukam
Apr 10, 2025
Author

I've been working on on the config. I've made some good progress and added some functionality. I have the hours working and disableoutagecheck. I have a placeholder in for the 1/3 phase check. I have datetime working, but settime is not working. Setting the time through Modbus might not be supported.

I feel like my config is at a good point to share my progress.

SmartGen_HGM4000.json

9 replies

jgyates Apr 11, 2025
Maintainer

Submit your logs from the About page. This will send me all of your log files and a snapshot of your registers.

if that does not yield fruit then there are more verbose logging options we can enable that will give more data regarding what is an is not happening via modbus.

I just now updated the ./genmon/OtherApps/modbudump.py program to allow for writing to coil registers. Before it could read coils, input and holding registers, but could only write holding registers. Now it will attempt to write holding registers so we can try it out without genmon in the way.

Docs for modbusdump: https://github.com/jgyates/genmon/wiki/1----Software-Overview

since you are in a managed python environment you will need to activate before running modbusdump like it shows in this page:
https://github.com/jgyates/genmon/wiki/Appendix-S---Working-in-a-Managed-Python-Environment

The following two examples use IP 192.168.11.196 port 9989:

To read coils 1 - 4 with modbus address 64 hex (100 decimal) type (from the OtherApps folder):

  python3 modbusdump.py -i 192.168.11.196 -t 9989 -a 64 -s 1 -e 5 -c

To write a value of 1 to mobus coil reg 6 with modbus address 64 hex type:

  python3 modbusdump.py -i 192.168.11.196 -t 9989 -a 64 -s 6 -w 1 -c

Log files for modbusdump.py are stored in the same folder and modbusdump.py (myserial.log, myserialtcp.log and mymodbus.log)

jgyates Apr 11, 2025
Maintainer

I will test out the power graph tomorrow. I did not look a that one before.

jgyates Apr 11, 2025
Maintainer

I checked in another update. It should fix the power graph.

jgyates Apr 11, 2025
Maintainer

also, forgot to mention, you have to stop genmon to run modbusump.py

after you send your logs, the next time you run the generator, send another set of logs so I can get a copy of the registers while running.

So are all of the .1 multipliers 1? or just the generator voltage, current and power output (i.e. battery and Frequency are still .1)?

Neukam Apr 11, 2025
Author

Submit your logs from the About page.

Sent.

I just now updated the ./genmon/OtherApps/modbudump.py program to allow for writing to coil registers

Does this use 0x0f or 0x05?

I just now updated the ./genmon/OtherApps/modbudump.py program to allow for writing to coil registers. Before it could read coils, input and holding registers, but could only write holding registers. Now it will attempt to write holding registers so we can try it out without genmon in the way.

I'll update and give this a try.

I checked in another update. It should fix the power graph.

I'll test this in the morning and send logs. I'm at work right now.

So are all of the .1 multipliers 1? or just the generator voltage, current and power output (i.e. battery and Frequency are still .1)?

Battery voltage and frequency are .1 Power factor is .01 and I think all the rest are 1. Attached is the config I was working with this afternoon. I think all the multipliers there should be correct.
SmartGen_HGM4000.bak.json

Neukam · 2025-04-11T02:48:57Z

Neukam
Apr 11, 2025
Author

I think the powergraph is fixed. It now shows a blue line all the way across at 0. It doesn't show the data from when I ran the generator this afternoon. I'm assuming the data doesn't survive a restart.

Here is the output from modbusdump.py when attempting to write to coil 3 (auto)

(genenv) root@genmon:~# python3 genmon/OtherApps/modbusdump.py -i 192.168.77.54 -t 8899 -a 10 -s 3 -w 1 -c
IP Address : 192.168.77.54
TCP Port : 8899
Address is : 10
Start Register : 3
Reading Coil Values
Using serial over TCP.

Modbus Timeout Error(s) detected: 1
Program Complete

Same output from modbusdump.py when attempting to write to holding register 74 (controller time: day)

(genenv) root@genmon:~# python3 genmon/OtherApps/modbusdump.py -i 192.168.77.54 -t 8899 -a 10 -s 74 -w c
IP Address : 192.168.77.54
TCP Port : 8899
Address is : 10
Start Register : 74
Using serial over TCP.

Modbus Timeout Error(s) detected: 1
Program Complete

9 replies

jgyates Apr 11, 2025
Maintainer

What is odd is that when you initial started working on this, you enabled the "Automated Feedback" option which sends emails to me with logs if genmon run into unexpected data (this can happen when starting to develop a new controller JSON file from scratch) and I got emails from you, but that was only up until 4/7/2025. After that I go no emails. The mymail.log file does not have errors when sending, so my guess is some other filter or firewall is prohibiting it from arriving. It is odd since this is just one gmail address to another.

Regarding the modbus writes, it is interesting that both holding and coils writes fail. To me this is a clue that there may be a controller setting that needs to change. I have seen other controllers that allow modbus reads by default, but modbus writes must be explicitly allowed in the controller menu settings. I tried looking that the SmartGen user manual but it does not have a lot of info on the menus, only the external connections. It may be worth browsing the menus looking for modbus related security features that allow / disallow modbus writes.

Neukam Apr 11, 2025
Author

I have been keeping Automated feedback off most of the time until the 9th to keep from blasting you with emails while I was building the config. I would occasionally turn it back on momentarily to clear the "unknown error occurred" warning flag, and then turn it back off. I just checked my sent mail and genmon has sent a minimum of 3 submissions per day since the 7th. Is it getting filtered on your end?

I haven't noticed any settings related to modbus writes on the controller or in the manual, but I will take a closer look.

jgyates Apr 11, 2025
Maintainer

It actually was in the spam folder for the account email.

It is odd that it times out on writing holding registers too. So far, every generator model I have tried writing holding registers works with the multiple write function. genmon uses the multiple register read and write calls for all holding, input and coil registers, both reading and writing. If you dont find anything I can possibly create a test to see if it responds to single register / coil writes.

Neukam Apr 11, 2025
Author

Hold off on creating a test until I can test with a direct connection.. It's possible that the EW11 wifi adapter or something else is mucking things up.

I'm testing now with the new config. I see only 2 issues.

The gauge for Total Apparent Power still had a multiplier of 0.1
Reactive Power needs "signed16": true to show negative values.

I'm waiting for the tower lights to cool down so I can test the changes. I will send a log when the lights are back on. Powergraph is working.

jgyates Apr 11, 2025
Maintainer

I made the changes to the JSON file. Let me know how it goes.

Neukam · 2025-04-11T16:25:29Z

Neukam
Apr 11, 2025
Author

Reactive Power needs to be "type": "int" to show a negative value. I changed that on my end. Logs sent

1 reply

jgyates Apr 11, 2025
Maintainer

good catch. Thanks!

I updated the JSON again. I also made a small tweak to the code related to the JSON interaction with the menu color.

Neukam · 2025-04-11T17:38:51Z

Neukam
Apr 11, 2025
Author

Everything's looking good with the config. Oil temperature can probably be removed from status. I put it there in case others wanted it and had it hidden based on returning zero. I'm not using it and I doubt many others are going to use it either. I think it depends on CAN communication from an engine ECU.

I was able to hook my laptop up directly to the controller and was able to control it using 0x05 but not 0x0f. I forgot to test writing the date and time. I might take the controller with me to work and test it later.

1 reply

jgyates Apr 11, 2025
Maintainer

what were you using to control it when it was hooked to your laptop?

I have some time tomorrow. I will start added the support for singled writes.

Let me know how the time test goes. If writing holding registers does not work with 0x10 (multiple registers), I can also add holding register write with 0x06 (single register)

Neukam · 2025-04-11T19:26:15Z

Neukam
Apr 11, 2025
Author

what were you using to control it when it was hooked to your laptop?

I'm using PowerHud Modbus Tester with a cheap CH340 based rs485 adapter.

Let me know how the time test goes.

Same situation here, works with 0x06 but doesn't work with 0x10

2 replies

jgyates Apr 11, 2025
Maintainer

I added the function 0x5 and 0x6 stuff in the code, updated the JSON and tested on my end.

To test the writing of the time and date, go to the maintenance page, then click the "Set Generator Time" button. Then go back to the status page and watch the generator time and monitor time at the bottom. The software will wait for the monitor time minute to change then it will update the generator time. It does it this was since most controllers do not allow changing the seconds so this is helps put the controller and the monitor system within one minute of each other.

Let me know how your testing goes.

Neukam Apr 11, 2025
Author

Thanks. I brought the controller with me to work. I'll reinstall it and test when i get home in the morning.

Neukam · 2025-04-12T15:02:37Z

Neukam
Apr 12, 2025
Author

Let me know how your testing goes.

Writing single coils and holding registers is working, however the server seems to hang for up to a minute before sending the command, most of the time (but not always). When this happens status bar goes red and the notification says "Disconnected from server". Photo attached. This behavior was observed before you implemented the changes as well. I'm not sure what's causing it, it could be the wifi adapter. The command does eventually get sent and the system does resume normal operation.

Setting time and date is working. Does/can genmon periodically set the time automatically or after a DST event?

The test button doesn't do anything on this controller and should probably be removed. Attached is a copy of the HGM4000LT protocol manual. It doesn't show any function for input coil 2 and shows status coil 0x28 as Reserved. I can't think of what function "Test" would perform on this series of controller, unless it has something to do with ECU communication with electronically controlled engines.

It would be useful to have Switch State, Engine State, and Generator Output on the Maintenance page. Feedback from these is important when controlling the generator remotely, for example to make sure the generator is "Normal Running" before transferring the load.

HGM4100LT_Protocol_en.pdf

0 replies

Neukam · 2025-04-12T15:08:17Z

Neukam
Apr 12, 2025
Author

Logs sent. Ignore anything before 8:53. That's when I reconnected the controller

0 replies

jgyates · 2025-04-12T17:35:38Z

jgyates
Apr 12, 2025
Maintainer

The time and data behaves like this:

You can press the Set Generator Time button to initiat a sync of the generator time to the monitor system time.
Alternatively, you can set "Sync Time" on the setting page and genmon will sync the time of the controller to the generator every hour
If you enabled "Sync Daylight Savings Time" on the settings page genmon will only sync the controller time to the monitori system when there is a change in daylight savings time.

I will remove the test button.

I can duplicate the switch state, etc on the maintenance page, but I will leave it on the status page as this is in line with all the other controllers supported by genmon as this is probably the most basic status information.

It is not clear about the timing of the "Disconnect" messages, are they happening before you issue a command or after (include the time commands in this as well as the coil commands). Does this occur if no commands have been issued?

The "Disconnected from Server" messages occur when the web app failed to communicate with genmon. The web app polls genmon every in two threads. One polls every 3 seconds and the other polls every 5 seconds. If more than 5 messages are missed the message is displayed in the web interface. Normal modbus reads occur in a separate thread and have no bearing on the web app, however commands (like buttons) that issue a modbus write are a different. If the controller delays responding to the modbus command it could possibly cause this message. I looked at your logs and nothing is reported as abnormal.

3 replies

Neukam Apr 12, 2025
Author

Is "Sync Daylight Savings Time" dependent on "Sync Time"?

I can duplicate the switch state, etc on the maintenance page, but I will leave it on the status page as this is in line with all the other controllers supported by genmon as this is probably the most basic status information.

That's perfect. It should be in both locations.

It is not clear about the timing of the "Disconnect" messages, are they happening before you issue a command or after (include the time commands in this as well as the coil commands). Does this occur if no commands have been issued?

I get the disconnect message only after pressing a command button. After a period of time, the command is sent to the controller and the disconnect message clears and the server becomes responsive again. The server starts responding again around the time that the command reaches the controller. I can't tell whether the command reaches the controller first or the server starts responding first. 2 times I've restarted the server after several minutes of the server being unresponsive, but it's usually unresponsive for less than 1 minute.

Setting the time doesn't seem to cause the server to disconnect and become unresponsive like issuing a command does, but it sometimes seems to take several minutes to update the time.

F438

jgyates Apr 12, 2025
Maintainer

No, Sync DST is not dependent on Sync time.

Setting the time by clicking the Set Generator Time button can take up to one minute for the time to change. The software will wait for the monitor time minute to change (e.g. wait for 4:05 to become 4:06) then it will issue the command to update the generator time. It does it this was since most controllers do not allow changing the seconds so this is helps put the controller and the monitor system within one minute of each other. If you click the button close to the time of the minute changes (for example 4:05 and 45 seconds) then the delay in updating will be less. This is by design.

We can look at this in more detail, but if the issue is the controller is taking a while to respond (TCP modbus timeout is around 10 seconds) I can't change the delay as this is the controller. I can increase the number of missed messages on the web UI before a notice is displayed. The button commands are synchronous in nature. I can look at the timeout on the AJAX calls the web interfaces uses. It could be that they need to be increased for button command calls into genmon as these can take longer than other calls into genmon if the controller delays a response.

We can pull up the Developer Console in a browser and it will show how long each calll the web UI makes into genmon and how long it takes. My guess is that the button call is taking much longer. If Chrome this is under Settings -> more tools -> Developer Tools, then select the "Netowrk" section in the developer view. Once this is selected then pull up genmon and try to click one of the coil buttons.

Neukam Apr 12, 2025
Author

Is "Sync Daylight Savings Time" dependent on "Sync Time"?

I think I answered my own question here. Syncing during a daylight savings event wouldn't really be necessary if you sync every hour, right?

Neukam · 2025-04-12T18:57:37Z

Neukam
Apr 12, 2025
Author

2 replies

Neukam Apr 12, 2025
Author

I'm working on setting up genmon on a pi zero 2w to test with a direct connection to the controller. That should tell me if the EW11 bridge is causing the problem. I wish there was a pi zero hat that had RS485 and buck converter.

jgyates Apr 12, 2025
Maintainer

You can get a HAT for a pi that will do that.

https://www.amazon.com/Rapsberry-Transceiver-Instrument-Interface-Raspberry/dp/B0C3GH6TTP/ref=asc_df_B0C3GH6TTP

probably better to use it with a pi4 instead of a pi zero as it has a more compatible footprint regarding the mounting

Neukam · 2025-04-12T20:43:53Z

Neukam
Apr 12, 2025
Author

I have a pi zero sized hat for rs485, but it doesn't solve the power problem. Pis are great, but they're not always the easiest or cleanest to deploy.

0 replies

jgyates · 2025-04-12T21:09:52Z

jgyates
Apr 12, 2025
Maintainer

I checked in an update to the JSON file. Let me know what you think. I put some sections on the maintenance page to break things up a bit.

Here is the output I have on my test system. My actual generator (different controller obviously) is simliar.

I am still thinking about what I can do to make this not as effected by slow controller response.

0 replies

Neukam · 2025-04-12T21:12:49Z

Neukam
Apr 12, 2025
Author

I've started testing with the Pi and everything seems to be working fine. Commands get sent almost instantly. The server isn't disconnecting or hanging. The clock now updates on the next minute change. Next I need to test the Pi with the EW11 to confirm that the EW11 is the problem.

0 replies

Neukam · 2025-04-12T21:18:00Z

Neukam
Apr 12, 2025
10000 Author

Here is the output I have on my test system.

That looks good.
One other thing that might be worth adding under maintenance

  {
    "comment": "Time remaining until maintenance is due",
    "type" : "list",
    "format": "%sh %sm",
    "title" : "Maintenance Timer",
    "value" : [
        {
          "reg": "0020",
          "type": "int",
          "mask" : "ffff",
          "title": "Maintenance Timer Hours"
        },
        {
          "reg": "0021",
          "type": "int",
          "mask" : "ffff",
          "title": "Maintenance Timer Mins"
        }
      ]
  },

1 reply

jgyates Apr 12, 2025
Maintainer

Done. I checked in the updated JSON

Neukam · 2025-04-12T21:19:48Z

Neukam
Apr 12, 2025
Author

I haven't said this enough, Thank you so much for all the effort you've put into helping some random guy on the internet get his one off generator system working

0 replies

Neukam · 2025-04-12T21:38:44Z

Neukam
Apr 12, 2025
Author

More weirdness, I connected the pi to the controller with the EW11 and everything is still working well. It seems the EW11 is not the problem. Not sure what's going on with the proxmox/ubuntu setup...

5 replies

jgyates Apr 12, 2025
Maintainer

It could be related to the python Flask libraries that power the web server. they work great most of the time but I have seen a case or two where a reboot solved strange problems with the web server.

Neukam Apr 12, 2025
Author

I'll try a reboot and make sure everything is updated. Is there a linux distro that you know work well that I can try if I continue to have problems?

jgyates Apr 12, 2025
Maintainer

I use the raspberry pi os which is Debian based, formally called raspbian

Neukam Apr 12, 2025
Author

I meant something I can run on the Proxmox server. If I can do without the pi, that would be preferable.

jgyates Apr 12, 2025
Maintainer

I don't have a strong recommendation, Ubuntu and Debian have a lot of use for non pi users on the project.

Neukam · 2025-04-13T02:09:34Z

Neukam
Apr 13, 2025
Author

I reinstalled Ubuntu as a VM and had the same problems.
I installed Debian 12 as a VM and everything works properly.
I reinstalled Debian 12 as an LXC and had problems.
I'm going to stick with Debian 12 VM.

I don't know what the issue is. I'm wonder if it's a dependency issue. Every install was a minimal install without a GUI

0 replies

Uh oh!

how to write coils (function 0x05) #1271

Uh oh!

Neukam Apr 8, 2025

Checklist

Version

Have you submitted your log files from the About page? Note that outbound email must be working for logs to be submitted.

What is/are your question(s)?

Additional information

Replies: 17 comments · 40 replies

Uh oh!

jgyates Apr 8, 2025 Maintainer

Uh oh!

Neukam Apr 8, 2025 Author

Uh oh!

jgyates Apr 8, 2025 Maintainer

Uh oh!

jgyates Apr 8, 2025 Maintainer

Uh oh!

jgyates Apr 8, 2025 Maintainer

Uh oh!

Neukam Apr 8, 2025 Author

Uh oh!

Neukam Apr 10, 2025 Author

Uh oh!

jgyates Apr 11, 2025 Maintainer

Uh oh!

jgyates Apr 11, 2025 Maintainer

Uh oh!

jgyates Apr 11, 2025 Maintainer

Uh oh!

jgyates Apr 11, 2025 Maintainer

Uh oh!

Neukam Apr 11, 2025 Author

Uh oh!

Neukam Apr 11, 2025 Author

Uh oh!

jgyates Apr 11, 2025 Maintainer

Uh oh!

Neukam Apr 11, 2025 Author

Uh oh!

jgyates Apr 11, 2025 Maintainer

Uh oh!

Neukam Apr 11, 2025 Author

Uh oh!

jgyates Apr 11, 2025 Maintainer

Uh oh!

Neukam Apr 11, 2025 Author

Uh oh!

jgyates Apr 11, 2025 Maintainer

Uh oh!

Neukam Apr 11, 2025 Author

Uh oh!

jgyates Apr 11, 2025 Maintainer

Uh oh!

Neukam Apr 11, 2025 Author

Uh oh!

jgyates Apr 11, 2025 Maintainer

Uh oh!

Neukam Apr 11, 2025 Author

Uh oh!

Uh oh!

Neukam Apr 12, 2025 Author

Uh oh!

Neukam Apr 12, 2025 Author

Uh oh!

jgyates Apr 12, 2025 Maintainer

Uh oh!

Neukam Apr 12, 2025 Author

Uh oh!

jgyates Apr 12, 2025 Maintainer

Uh oh!

Neukam Apr 12, 2025 Author

Uh oh!

Neukam Apr 12, 2025 Author

Neukam
Apr 8, 2025

Replies: 17 comments 40 replies

jgyates
Apr 8, 2025
Maintainer

Neukam Apr 8, 2025
Author

jgyates Apr 8, 2025
Maintainer

jgyates Apr 8, 2025
Maintainer

jgyates Apr 8, 2025
Maintainer

Neukam Apr 8, 2025
Author

Neukam
Apr 10, 2025
Author

jgyates Apr 11, 2025
Maintainer

jgyates Apr 11, 2025
Maintainer

jgyates Apr 11, 2025
Maintainer

jgyates Apr 11, 2025
Maintainer

Neukam Apr 11, 2025
Author

Neukam
Apr 11, 2025
Author

jgyates Apr 11, 2025
Maintainer

Neukam Apr 11, 2025
Author

jgyates Apr 11, 2025
Maintainer

Neukam Apr 11, 2025
Author

jgyates Apr 11, 2025
Maintainer

Neukam
Apr 11, 2025
Author

jgyates Apr 11, 2025
Maintainer

Neukam
Apr 11, 2025
Author

jgyates Apr 11, 2025
Maintainer

Neukam
Apr 11, 2025
Author

jgyates Apr 11, 2025
Maintainer

Neukam Apr 11, 2025
Author

Neukam
Apr 12, 2025
Author

Neukam
Apr 12, 2025
Author

jgyates
Apr 12, 2025
Maintainer

Neukam Apr 12, 2025
Author

jgyates Apr 12, 2025
Maintainer

Neukam Apr 12, 2025
Author

Neukam
Apr 12, 2025
Author