Strange HR92 behaviour

[Tempted]

I've had Evohome installed for just over a month now and in generally it is great. As with anything gadgety, there's a tendency to play with it too much and to over analyse every little anomaly. In the most part, I think I'm just letting it do it's thing. Admittedly, I have been playing with schedules a lot since I got it, but I have been making less adjustments the more we live with it.

That being said, I have noticed a strange anomaly in one of my zones which seems a little too 'perfect' to be just a series of one-offs. So the porch in our house is scheduled to be 10degrees from 10pm at night until 9am the next day. Nothing exciting here. The issue I'm having is that my graphs show the following 'targets' which do not correlate with my schedule (in any of its iterations):

5deg 17:00 - 5th October
5deg 18:00 - 7th November
8deg 06:00 - 8th November
5deg 12:00 - 17th November

The reason this seems odd to me is that 1) the temperatures seem very specific and 2) the times are exactly on the hour. I can tell you with a good degree of certainty that I have not made typo's with my schedule and would be hard pressed to accept that as an explanation to this. I could maybe put one or at a push two of these down to my stupidity, but not four . Am I seeing some kind of 'fail safe' type behaviour? I will admit to clicking the button on one of the HR's in another location and noticing it seemingly 'boot up' in front of my very eyes, although haven't noticed this with the porch. I don't know if that's a different issue though.

Any suggestions?

[paulockenden]

At these times are you actually seeing those targets on the HR92 itself?

The reason I ask is that (assuming you're using the API rather than an HGI80) what you are graphing is several steps away from the HR92 (Rad valve -> controller -> Honeywell cloud -> API) so the problem could be anywhere in that chain.

P.

[DBMandrake]

Sounds strange. Firstly how are you graphing, using the Web API or an HGI80 ? Using what software ?

When you list the "targets" above, I assume you mean that you were graphing the set points and the set point jumped to those values at those times ? Or do you mean the measured temperature was reported as being those temperatures ?

It would be useful to see the entire schedule for a day for that room - you say it is supposed to be 10 degrees from 10pm until 9am, and yet only one of the "targets" you listed actually occurs between those times, so I'm not sure what you're getting at there.

When you think the set point is wrong, what does it show on the controller - does the controller agree with what your graphs show ? Does it show a manual override icon ? Overrides made by an HR92 or at the controller show a pocket watch icon, overrides made via the phone apps or web API shows a rectangular phone icon.

Another possibility if you see the set point change to 5 degrees for a while when it is not scheduled to is that the zone has gone into "window open" mode. This is triggered by a sudden drop in measured temperature which will temporarily set the zone set point to 5 degrees (by default for half an hour, or until a sudden increase in temperature is detected) then back to what it was. This is to minimise wasted heat if a door or window is left wide open. At this time of the year both our front and rear doors will set off window open mode on the nearest radiator if they are left wide open for even a couple of minutes.

In your graphs window open mode will just look like an unexpected set point change to 5 degrees and then back to the previous set point. On the controller there will be a window icon however, and an HR92 display will show the word window at the bottom. The phone apps will just show a 5 degree current set point with no explanation as to why it is 5 degrees when the schedule says otherwise, and no override icon.

[Tempted]

Thank you both for the input.

Sounds strange. Firstly how are you graphing, using the Web API or an HGI80 ? Using what software ?

I'm using the API which is plotting into emoncms.

When you list the "targets" above, I assume you mean that you were graphing the set points and the set point jumped to those values at those times ? Or do you mean the measured temperature was reported as being those temperatures ?

I'm referring to the set points.

It would be useful to see the entire schedule for a day for that room - you say it is supposed to be 10 degrees from 10pm until 9am, and yet only one of the "targets" you listed actually occurs between those times, so I'm not sure what you're getting at there.

The entire schedule is as follows:
18 09:00:00 16 11:00:00 18 17:00:00 16 20:00:00 12 22:00:00

When you think the set point is wrong, what does it show on the controller - does the controller agree with what your graphs show ? Does it show a manual override icon ?...

At these times are you actually seeing those targets on the HR92 itself?

I must admit, I have only observed these sporadic set points after analysing my graphs. I don't generally check what the HR92's are reporting unless I happen to be in a room and want to know what the target vs actual is.

Another possibility if you see the set point change to 5 degrees for a while when it is not scheduled to is that the zone has gone into "window open" mode. This is triggered by a sudden drop in measured temperature which will temporarily set the zone set point to 5 degrees (by default for half an hour, or until a sudden increase in temperature is detected) then back to what it was. This is to minimise wasted heat if a door or window is left wide open. At this time of the year both our front and rear doors will set off window open mode on the nearest radiator if they are left wide open for even a couple of minutes.

An interesting theory. My graphs do indeed show how the porch is susceptible to quicker than average temperature drops given the lack of ground insulation, 2 large windows, 2 velux windows and an exterior door, however, I feel the timings are too precise for this to be an explanation, surely?

[DBMandrake]

I will admit to clicking the button on one of the HR's in another location and noticing it seemingly 'boot up' in front of my very eyes, although haven't noticed this with the porch. I don't know if that's a different issue though.

That could be a problem with lack of tension on the battery contacts, see the following thread regarding this:

http://www.wordpress-1219309-4387497.cloudwaysapps.com/vbulletin/showthread.php?4957-HR92-battery-contact-problem

The entire schedule is as follows:
18 09:00:00 16 11:00:00 18 17:00:00 16 20:00:00 12 22:00:00

I don't see 10 degrees there, or did you mean to say 12 in your original post ?

I must admit, I have only observed these sporadic set points after analysing my graphs. I don't generally check what the HR92's are reporting unless I happen to be in a room and want to know what the target vs actual is.

An interesting theory. My graphs do indeed show how the porch is susceptible to quicker than average temperature drops given the lack of ground insulation, 2 large windows, 2 velux windows and an exterior door, however, I feel the timings are too precise for this to be an explanation, surely?

If the problem is happening on a regular schedule try to catch it in the act - check the set point both as reported on the controller (along with any icon next to the set point) and on the HR92 itself.

You might also want to have a browse through this thread:

http://www.wordpress-1219309-4387497.cloudwaysapps.com/vbulletin/showthread.php?4728-Phantom-override

I would suggest adjusting the battery contact tension as described in the first link (which will also cause a reboot of the HR92 when the batteries are removed) and see if that fixes it. But try to catch it in the act first before disturbing the batteries so we can get a better idea of what is going on for future reference, even if it does fix it.

[Tempted]

Sorry - the schedule has been all over the place the last few weeks. I keep changing my mind. Version 38 was 10 degrees, version 39 is 12 degrees. It was never 8 or 5 degrees though, that's for sure.

I thought I'd read something about the contacts in the HR92's. I'll have a read and check mine over.

Thanks