Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Moving from Alfawerks to Comfort2MQTT
#11
(06-06-2026, 01:39 PM)Ingo Wrote:
(06-06-2026, 01:23 PM)DavidJ Wrote: Ok Ive successfully updated both UCM and the firmware but unfortunately the  Error sending command 'cc00', closing socket problem persists . Interestingly before I fully transitioned from Alphawerks I tested this a few months ago and aside from clashing with Alphawerks I'm (nearly) sure I didnt get these issues so its possibly due to some recent changes.
Ive also opened an issue on Github, hopefully it can be resolved as this would be superb.
Thanks

Did you upgrade the Eth03 firmware as well? Make sure you don't have a duplicate IP, or even another device trying to access the same IP and Port.

Yes I did update the  UCM firmware (7.141 → 7.225) but not the ETH03 which I believe is done through CSM and wasnt as confident doing that. I also updated the console firmware.
I had the previous Alphaweks module module which I literally just pulled the network cable out - not sure if that can be still polling perhaps through the board, i'll try stopping that just in case?
Whatever port I use it reverts to 1001. Something maybe Alphawerks module could be holding that?
The Comfort ETH03 is on 192.168.0.50 and obviously I disconnect the Configurator.
Appreciate any advice.

Ok I unplugged Alphawerks which made no difference.


Ive had some help from Claude - That may or may not be helpful... Heres findings

Also critically important — this is a Tibbo DS device (V3.34), not actually a Cytech ETH03. It's a Tibbo Device Server that Cytech used as the basis for their ETH03. The firmware version shown is V3.34 of the Tibbo firmware, which dates from around 2017 and corresponds to the ETH03 General 2.x firmware we were looking at earlier.
This also confirms it's single port only — there is no port 2. Port 1001 is all you have.
So to summarise what we now definitively know:
  • Connection timeout = 0 ✅ — not the cause
  • Single port device — port 1002 will never work
  • No duplicate connections — device shows idle
  • ETH03 firmware is Tibbo V3.34 (equivalent to ETH03 2.x)
The
cc00
drop is definitively a comfort2mqtt software bug — nothing on the hardware side is causing it. The panel, UCM, ETH03 and network are all configured correctly.
Reply to Ingo with these findings — specifically that the Tibbo DS connection timeout is already 0-Disabled and it's a single-port device. This gives him the definitive information he needs to diagnose the bug in the code.
Reply

#12
I think you might have missed this critical section of the documentation. The software works correctly as designed.

The following Cytech Universal Communications Modules (UCM) Ethernet modules are supported:
  • [UCM/Eth01] - Obsolete/Untested
  • [UCM/Eth02] - Obsolete/Not recommended. Not all functions work with this module.
  • [UCM/Wifi01] - Not Recommended (WiFi) - Firmware 7.176 or later.
  • [UCM/Eth03 or Eth03 Mainboard Plug-in] - Recommended (LAN) - Firmware 7.176 or later.
Reply

#13
I wasn't being critical.  As I said this was an AI response. 
Im merely looking for a resolution.
Reply

#14
Hi @Ingo,
Is it possible or practical to add a timeout/retry limit — if for example cc00 fails 3 times, stop sending it for the session rather than disconnecting?
Or will that break something? 
This is a wonderful addition to Home Assistant, fantastic work by you, also I suspect there's numerous others running the older hardware and who still use the Alphawerks solution, would like to swap over.
Reply

#15
That is not possible. Your suggestion will break auto-reconnect.
Reply

#16
(06-07-2026, 08:03 AM)Ingo Wrote: That is not possible. Your suggestion will break auto-reconnect.
Ok, appreciate response.
Reply

#17
Hi Ingo,
Further to my earlier messages about the cc00 keepalive drop on non-ARM CM9001 hardware, I've spent considerable time today investigating and I with the help of AI believe a possible fix for those with older hardware. I've also verified through extensive testing that all other hardware variables have been eliminated:

Comfort panel firmware upgraded to 8.030
UCM firmware upgraded to 7.225
ETH03 Tibbo firmware V3.34, connection timeout = 0 (disabled)
No duplicate connections or competing clients
Single port device (port 1001 only — no port 2 available on this hardware)

The cc00 drop occurs every 22 seconds consistently regardless of firmware version. The fix is a 3-line change to SendCommand() in comfort2.py:
python        except:
            if command == "cc00":
                logger.warning("Keepalive command 'cc00' failed to send - ignoring to maintain connection.")
                return
            logger.error("Error sending command '%s', closing socket.", command)
            self.comfortsock.close()
            raise
I believe this catches cc00 failures silently without closing the socket, while preserving existing behaviour for every other command — including the full reconnect logic for genuine connection failures. Auto-reconnect is completely unaffected.
I attempted to test this fix by installing from a forked repo but was unable to get HAOS to rebuild from source rather than using the cached image.
I appreciate this hardware isn't your primary test platform, but hopefully this is a straightforward fix that would make comfort2mqtt usable for anyone on older non-ARM CM9001 hardware. Would you be willing to consider including it in a future release?
Many thanks for your hard work on this addon.
David
Reply



Forum Jump:


Users browsing this thread:
1 Guest(s)

Powered By MyBB, © 2002-2026 Melroy van den Berg.