Changed behavior of HM-CC-RT-DN (radiator valve) after firmware update to 1.5

[blueznl]

Yesterday evening I've updated the firmware in my HM-CC-RT-DN radiator valves to 1.5. Am I correct in assuming its behavior has changed?

It seems it now has a much more 'granular' control over the valve opening (more smaller steps over shorter time spans). I noticed this when looking at the valve opening graphs that I use to verify my floor heating solution. It could be that it is re-learning my setup (due to fresh firmware) and will slow down over time. It seems it is making adjustments up to every minute, or so the valve position diagram that I maintain suggests, that would mean up to 500..1000 adjustments per day. Won't that eat up the battery fast?

The changed behavior may be great for regular heating radiators, but my solution is slightly different. Unfortunately it may be a bit too 'adaptive' now for my application, as I trigger my pumps on the valve position. (If valve open more than 5 % then start pump etc.)

I will let it run with the current setup for a day or two, then will increase the pump treshold to see if that reduces the run time of the pumps. (It looks like I'm using the pumps as a kind of electric heater, instead of the city block heating ('stadsverwarming') and that can never be a good thing

Is there a hysteresis option or something similar for the HM-CC-RT-DN radiator valve, or for the HM-TC-IT-WM-W-EU thermostat to make it behave a bit 'sloppier'? (I thought there was a hysteresis option in the past, I may be overlooking it or I may be a complete idiot, probably the latter.)

[blueznl]

The more I look at it, the more I think something changed with 1.5.

Here's a dump of the last few days, to the left the old firmware, to the right the new firmware.

[tomi_cc16]

Your assumption is correct - it changed and not everybody is happy with that:
viewtopic.php?f=27&t=47569

[blueznl]

Darn. I need to downgrade the firmware then, this isn't an improvement.

I did some testing, and it looks like they have opted for a more 'soft approach' where the delta T between actual and target temperature defines a limitation to the valve opening.

It seems to operate independent of the adaptive learning mechanism too, probably it still has the same adaptive control on-board, just limits the max valve opening on the output.

How do I downgrade until this has been solved? Is there a read-me / release note for this firmware?

[Sergo]

I have noticed same shit two weeks ago. In addition, one of regulators “stacking” periodically in returning valve at 21%, while physically closed. Over two weeks behavior improved a bit, but overall behavior stays the similar level. As you said, it’s good for central heating, but not acceptable for local heating system.

IMO, need to think about falling back to previous firmware version. As suggestion for eQ3: it’s better to introduce switch in options, for different type of heating system (local & central)

[tomi_cc16]

If you want to downgrade manually to 1.4 better do it quick.
Old firmware is avaible but after some time could be it will be taken down from the servers.

https://www.eq-3.de/Downloads/Software/ ... 141020.tgz

Here a video how to manually up- or downgrade:
https://youtu.be/jCNqTpXwdA4

[blueznl]

Downloading and downgrading as we speak. I'll change the ground floor valve to the old software and keep the first floor on the new one for comparison. But looking through the German parts of this board it seems I am not the only person who thinks the new software is NOT an improvement...

I'll report the results after a few days.

[Sergo]

Thanks!
Will wait for your feedback, after downgrade.
I have two rooms on HM-TC-IT-WM-W-EU & HM-CC-RT-DN, both upgraded.
First week after upgrade, each night, my gas heater was halting operation, due to overheat of system. Had to play with schedule/temp settings, to temporary workaround this.

Just now did an quick check and found inconsistent behavior:
Was:
in first room - temp set for 18, actual 20.4
in second room - temp set for 18, actual 20.1
Made a change of set temp to 20.5 for both. In room 1 radiator moved to 15%, in second to 21%.
After 20 min did a check:
In first room - current temp in room 20.6, radiator at 15%, really warm for hand
In second room - current temp in room 20.1, radiator at 21%, radiator almost cold, just lightly warm for hand

Few weeks ago, before upgrade, both worked well, quickly heating after command is set. Overall behavior were better compare to even HMIP radiator controllers in other rooms

After 30 min since new temp set:
Room 1 - actual temp is 21.2, valve at 0%. This room is bigger compare to second one.
Room 2 - actual temp at 20.2, valve at 21% seems like again, inconsistency between reported & actual valve position. This will result in overheat trigger by gas heater if continued for >3 hours...

[tomi_cc16]

If you also want to downgrade HM-TC-IT-WM-W-EU
https://www.eq-3.de/Downloads/Software/ ... 150827.tgz

[blueznl]

I did some more testing, and there is definitely something wrong with the new firmware. I have a very slow system (I'm actually controlling floor heating with it, with a delta T of 1 degree in an hour, so yeah )

What I noticed is the following:

- 1.5 keeps 'changing' the valve position, even when there's nothing to change (ie. room temperature is so slowly changing it is constant, but the valve just keeps running up and down)

- 1.5 limits valve opening depending on distance between actual and target temperature (probably to avoid overshoot, but when your system is slow it means it never reaches its target temperature), I tested this by increasing the difference between target temp and actual temp, and then you'll see it use a larger initial valve opening. (1.4: 90% / 1.5 small delta T: 20% / 1.5 large delta T: 40%)

- 1.5 doesn't seem to 'learn', at least after a few days in my setup it did not change its behavior in any noticeable way

Note that I set up a direct connection thermostat to valve in the CCU. I have a feeling (unproven though) that the 1.5 was mostly focused on use of the valve as a radiator valve only. Perhaps EQ3 missed something in testing?