Page 1 of 4 1234 LastLast
Results 1 to 10 of 36

Thread: New Install - Banging Head Against The Wall - Hot Water Sensor!

  1. #1
    Automated Home Jr Member
    Join Date
    Jan 2020
    Posts
    36

    Default New Install - Banging Head Against The Wall - Hot Water Sensor!

    Hi

    I installed a few weeks ago our new Evohome installation. The CH is working as hoped but the hot water regularly (every few days) appears to overshoot the hot water temp by a considerable margin (set at 55, gets to 67). I've hooked our install up to Home Assistant so I can clearly see the hot water temps being reported overtime.

    System overview:

    • Wifi controller
    • 14 HR91s
    • 2 BDR91s (one for HW, one for CH)
    • Hot Water Kit, unvented cylinder, wired in series with existing high-limit stat.
    • Water temp 55c, differential 5c (also tried 10c).
    • Boiler via Opentherm, temp limited to 70c and being followed

    I've read the forum, so I'm aware of others having/or have had similar issues. I've tried adjusting the battery contacts, moving the sensor, bringing the controller closer etc.. etc.. but yet still the problem persists. The only thing to be aware of is the hot water insertion sensor currently has no thermal paste applied, but it's sitting snuggly within its well.

    A few observations for the hot water:

    • Sensor signal strength is reported as excellent (5 flashes)
    • Removing and replacing sensor batteries results in an immediate temp update being reported.
    • Temps appear to be reported in consistently largely every hour, but sometimes every couple of hours if no water is being used.
    • Temps appear to be reported more frequently when temp is changing significantly - as expected, every 5-10 mins
    • However, the overshoots ONLY appear to happen when the water temp doesn't reach the set temp within 5-10 mins of the water heating, our cylinder reheats quickly. So for example, the first temp report after having a shower, occurs within 5-10 mins of starting the shower and the temp reads 54.9, no further temp is reported for 1 HOUR! By which point the water temp reaches the 66-67c, which I guess is the limit possible with the boiler set to 70c.
    • If the water is reheat is reached within 5-10 mins, say 55.1, the system works perfectly, within another temp reading reported within 5 mins, the temp creeps to say 57 but otherwise is perfect.

    Any help, or guidance would be really appreciated.

    Given the temps are being reported, I've not bothered with the thermal paste so far.

    The obvious issue is the temp not being reported for an hour whilst re-heating, the moment it's reported the system shuts off etc... perfectly.

    This morning, following a couple of showers we've had two consecutive 67c overshoots, yet for the previous 2-3 days things appeared to be working as planned.

    This graph shows the overshoots, the green dots show when temp readings are being received. Notice the lack of readings when the temp is climbing above the set point.

    Water Graph.jpg

    Thanks
    Dan
    Last edited by [email protected]; 10th September 2020 at 03:01 PM.

  2. #2
    Automated Home Legend
    Join Date
    Sep 2014
    Location
    Scotland
    Posts
    2,361

    Default

    Hi Dan,

    Sorry to hear you've having these problems. You're not alone. It looks like you have researched the problem well. I'm sure you've read some of the threads I've posted about this issue in too.

    Unfortunately after having the CS92A hot water kit for 3 something years and having spent countless dozens of hours troubleshooting this issue and talking to other people with the same problem I believe it is a systemic problem which cannot be solved. In other words there is one or more design flaws in the system.

    I think there are actually two interwoven issues that interact with each other.

    One is unreliable comms from CS92A to Evotouch - in short, sometimes a temperature update sent by the CS92A is not received and processed properly by the Evotouch as if it didn't "hear" the message. I've actually gone to the lengths of building a DIY receiver for the Evohome's protocol (NanoCUL with CC1101 receiver and software called evohome_rf) to independenty monitor and log all transmissions in the house for later analysis.

    This is situated about a metre from the Evotouch (and operates receive only, so no chance of interference) so should be able to receive anything the Evotouch can and vica versa.

    And after comparing these logs with the times when I see hot water issues the conclusion is unequivical - the Evotouch is missing temperature transmissions from the CS92A which my independent receiver is correctly receiving, confirming that the "missed" messages are in fact being sent. This was quite a surprise to me because before building the monitoring system I had always assumed the fault lay with the CS92A not sending a temperature update.

    Whether there is an issue with the receiver built into the Evotouch or there is something a bit "funny" about the timing or quality of the transmission from the CS92A is hard to determine for sure, but at the end of the day the independent monitor receives the "missing" transmissions just fine.

    In theory a collision with the Evotouch and CS92A both sending at once could cause the Evotouch not to hear the CS92A - however if that was the case neither would my independent receiver, which is much closer to the Evotouch than the CS92A, so I don't think that's the reason.

    The second issue is the transmission schedule of the CS92A is "sub-optimal" shall we say. As you've noted, if the temperature isn't changing it will send an update once an hour. If the temperature is changing within the differential band, it sends frequent updates, however below or above the differential band it only sends infrequent updates with changing temperature.

    This is very unfortunate. What seems to happen is that as the hot water reheats through the differential band it sends frequent updates and everything seems fine, however once it goes above the differential band (above 55c in your case) it seems the CS92A only sends one further update to say that the hot water temperature has been satisfied, then goes silent for at least 20 minutes despite the temperature still rising. That means if the Evotouch doesn't receive that one message (due to the reception problem described earlier) the reading will sit at 54c or whatever the previous reading was for something like 20 minutes. If you have a fast hot water cylinder the result is a major oveshoot.

    This is a fundamental flaw in the power saving algorithm in the CS92A - when the temperature is above the differential band and still increasing it should not go into low power mode sending infrequent updates.

    While the firmware in the CS92A is not field upgradable and hence can't be fixed without hardware replacement, it's the Evotouch which tells the CS92A what the differential range is, so in theory the Evotouch firmware could be modified to "lie" about the differential range to make the CS92A send fast updates over a wider temperature range. (for example it could lie and say your differential range was 45-65C instead of 45-55C) Whether that would cause compatibility issues with other devices that might monitor the hot water parameters I don't know.

    Honeywell don't officially acknowledge any of these problems or that there might be a design flaw and I've basically given up trying to solve this problem now. It doesn't happen very often for me (maybe 2 or 3 times a month) but when it does it's certainly annoying. I've considered ripping out the hot water part of my Evohome system and replacing it with a traditional programmable timer.

    Sorry to sound like a wet blanket but I don't think this problem can be solved, at least not by us end users.
    Last edited by DBMandrake; 10th September 2020 at 08:16 PM.

  3. #3
    Automated Home Jr Member
    Join Date
    Jan 2020
    Posts
    36

    Default

    Hi DBMandrake. Thanks for the detailed reply, having read your other posts I'd feared you'd say something like this.

    The option I'm thinking of is wiring back in the adjustable tank stat back in series with the existing high limit stat and the CS92A. When I installed the CS92A I opted to remove the adjustable stat, and only leave in series the high-limit stat. This would mean that adjustable stat if say set to 60c would kill the power to the BDR91 controlling the HW valve, avoiding the water getting to hot. My question is, as the boiler is controlled by OpenTherm would killing the power to the BDR91 then trigger a 0% heat demand to the boiler, I'm guessing not as the CS92A would still be reporting a lower temperature to the EvoTouch controller?

  4. #4
    Automated Home Legend
    Join Date
    Sep 2014
    Location
    Scotland
    Posts
    2,361

    Default

    Putting a regular thermostat in series with the relay set a bit higher than Evohome would certainly limit the overshoot, but there is also the risk that if it triggers before 55C it will stop the cylinder heating but the Evohome will keep trying to heat the cylinder. As a result the boiler will keep running, unable to heat the water.

    Mechanical thermostats also aren't very precise and have a differential of about 8C so that has to be taken into account when choosing what to set it to, as well as the fact that it will be sensing the cylinder in a slightly different location than the sensor going to the Evotouch.

    It could take a little trial and error to find what to set it to.

  5. #5
    Automated Home Jr Member
    Join Date
    Jan 2020
    Posts
    36

    Default

    I’m guessing also that if the mechanical stat correctly kicked in at crica 60c ish, but above 55c, if the EcoTouch hadn’t received the temp update from the CS92A it would also still continue to try and heat the water. Only difference would be that the boiler water wouldn’t be circulated through the cylinder as the valve would be shut.

  6. #6
    Automated Home Legend
    Join Date
    Mar 2017
    Posts
    507

    Default

    Out of interest what is your differential set at?

    I've installed a few CS92a and there's only one in particular that i've had issues with. That got replaced, whilst the newer was problematic it was much better than the old one. I had to replace a failed Unvented cylinder on that job and in the process I took DBMandrakes advice and rotated the CS92A 90 degrees and moved it. I haven't been back to the job yet but i'll be interested to see if it's made a difference....

  7. #7
    Automated Home Jr Member
    Join Date
    Jan 2020
    Posts
    36

    Default

    Differential is currently set to 5c

  8. #8
    Automated Home Legend
    Join Date
    Mar 2017
    Posts
    507

    Default

    I'd maybe up it to 7oc, although I doubt it'll make the slightest bit of difference.

    Any photos of the installed location? Is it close to anything etc.

    My CS92A works flawlessly at home, so I think a lot of it comes down to install location.

  9. #9
    Automated Home Legend
    Join Date
    Sep 2014
    Location
    Scotland
    Posts
    2,361

    Default

    Quote Originally Posted by mtmcgavock View Post
    Out of interest what is your differential set at?

    I've installed a few CS92a and there's only one in particular that i've had issues with. That got replaced, whilst the newer was problematic it was much better than the old one. I had to replace a failed Unvented cylinder on that job and in the process I took DBMandrakes advice and rotated the CS92A 90 degrees and moved it. I haven't been back to the job yet but i'll be interested to see if it's made a difference....
    I actually rotated mine back again...

    While it made the signal stronger for my independent receiver (which has a vertically polarised antenna) it didn't really solve the problem, and I think the antenna in the Evotouch might be mostly horizontal. Like a lot of "fixes" I've tried it seemed to help for a while but ultimately it still plays up after a while.

    I still get "lost comms" from the hot water sensor logged several times a month, more often in the middle of the night, despite my independent receiver receiving the hot water sensor just fine in the times that the Evotouch is reporting problems.

    Those people who don't seem to have problems, count your blessings.

  10. #10
    Automated Home Sr Member
    Join Date
    Feb 2019
    Posts
    78

    Default

    Happened to me after 4 months following installation. Shot up to 66 degrees C. Showed the temp graphs to my installer who replaced the probe that goes to the cylinder and it's been fine since. So may be change the probe?

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •