evoHome / boiler cycle settings

[mart1711]

Dear all,

Many thanks again for helping me getting started with evoHome a few days ago. evoHome has had a few days to learn and so far I'm duly impressed. I've set up temperature logging with evoLogger - many thanks to all of those having put in hard work into getting the evohome module for Python done (and then also sharing them out to everybody)

There is one question I could do with a little bit of help for, please. Recap - I have a Worcester Bosch Greenstar 24RI boiler with an open-vented system for HW and CH. evoHome is talking to the BDR91 as a boiler relay, which will operate the CH valve, which in turn will fire the boiler. I don't have the HW kit (yet). I've set up optimum start / stop. I'm pretty sure I haven't double bound the BDR91 by accident. I deleted the previous binding, before binding it once as a boiler really in evoHome.

I have four HR92s already fitted, with another four arriving next week.

Like many others I have noticed that the boiler is on much more often since evoHome is "in charge", but for much shorter intervals. I'm aware of the explanation in other threads - it's probably TPI (which I've read the white paper on). I like the analogy that was given in another thread - once you've reached 70 mph in your car you then don't floor it to maintain 70, but instead apply "small doses" to keep the speed up.

My question is around the cycles / hour. The evoHome controller help menu says "refer to the boiler manufacturer". I've been up and down the WB installation manual for my boiler, but it doesn't talk about cycles per hour.

It's currently set to the default 6, with a default on time of 1 minutes. I've definitely seen the boiler come on for 1-ish minutes multiple times once rooms were up to temperature, probably just to satisfy a low heat demand through TPI.

However, I also have a three-minute default pump overrun (controller by the boiler). So every 10-ish minutes, the boiler might come on for 1 minute, then the pump runs for another 3 minutes.

I have also found a passage in the WB manual that says the boiler needs 45 seconds to get the flame stabilised.

So now I'm wondering what I should set the cycles / hour and the minimum time to.

As the pump overruns by 3 minutes, I wonder whether 6 / hour is too much. I'm tempted to reduce down to 3 / hour. As the flame needs 45 seconds to stabilise, I'm also wondering whether I should set the minimum on time to 2 minutes.

I'm wondering what the long term effects are of the valves operating, plus the boiler firing potentially every 10 minutes for 1 minute at the default settings.

So I guess my questions really are:

1) does anybody have any recommendations with regards to cycles / hour and on times for this particular boiler?
2) based on what I have described above, does it sound right to change to 3 cycles / hour with a minimum of 2 minutes on time?
3) am I worrying too much about long term effects? Should I just leave the system to do what it is designed to do and stop worrying about it?

Many thanks in advance, please let me know if any clarification needed

Martin

[mart1711]

Hmmmm....also there might another part that I have misunderstood. I thought the cycles / hour were separate to those when the HR92s demand heat, but maybe not?

I had changed the cycles to 3 / hour with a 2 min on time, just to get a feel for what would happen.

Very soon after that I raised the set point in two rooms by ~ 1 degree. Both HR92s opened within the 4 minute period, but the boiler only fired about 10 minutes later.

Do the 3 cycles / hour now mean that even if the HR92s are calling for heat, the boiler will only ever come on every 20 minutes if required??

[DBMandrake]

So I guess my questions really are:

1) does anybody have any recommendations with regards to cycles / hour and on times for this particular boiler?
2) based on what I have described above, does it sound right to change to 3 cycles / hour with a minimum of 2 minutes on time?
3) am I worrying too much about long term effects? Should I just leave the system to do what it is designed to do and stop worrying about it?

Many thanks in advance, please let me know if any clarification needed

Martin

You might be worrying too much - as far as I've heard just about any gas boiler should be OK with the default 6 cycles per hour and less cycles per hour is mainly for other types of boilers like oil fired.

One minute minimum on time is probably OK, but I don't see any harm setting it to two minutes if it worries you. If you watch the flow temperature indicator when it comes on during each 10 minute cycle how long does it take for the temperature to start rising ? Does it take a full 45 seconds ?

One drawback of setting minimum on time too high is that if there is a very small heat load - such as only one radiator running during moderate conditions in an already up to temperature room, the demand to keep that room at a constant temperature might be below the minimum on time - in which case the boiler won't fire at all.

The room will gradually fall below the set point by half a degree or more before enough demand builds up to satisfy the minimum on time then the boiler will fire and the room will start to heat up again. This can lead to the temperature cycling below the set point and back above it again.

I saw this "problem" last winter because we had only our living room scheduled to be on between 8pm and 11pm - and the room would have to drop below the set point by at least half a degree before the boiler would come back on and start heating again, but when other rooms were also demanding heat (thus causing the boiler to fire more often) this didn't happen. And that was with one minute minimum on time - I tried two minutes and it made it worse.

Your system may be different though, you won't know unless you try it.

Hmmmm....also there might another part that I have misunderstood. I thought the cycles / hour were separate to those when the HR92s demand heat, but maybe not?

Not sure what you mean by that. Here is a simplified explanation of what happens:

Every HR92 sends a heat demand from 0 to 100% wirelessly to the Evotouch. This is proportional to how much "heat" it thinks it needs from the boiler to reach or maintain the set point. The controller combines the demands of all the different HR92's in some way and forms an aggregate demand also from 0 to 100% which it sends to the boiler relay.

The boiler relay then switches on for this percentage of time in each cycle. So with the default 6 cycles per hour and 1 minute minimum on time a 10% demand is the minimum required to get the boiler to fire at all. A 30% demand would mean 3 minutes on 7 minutes off.

I had changed the cycles to 3 / hour with a 2 min on time, just to get a feel for what would happen.

Very soon after that I raised the set point in two rooms by ~ 1 degree. Both HR92s opened within the 4 minute period, but the boiler only fired about 10 minutes later.

Do the 3 cycles / hour now mean that even if the HR92s are calling for heat, the boiler will only ever come on every 20 minutes if required??

Sort of, yes. Setting less cycles per hour will inherently make the boiler respond more slowly to small changes in demand.

Think of it this way - say you are on the default 6 cycles per hour. Lets say that the 10 minute cycles are aligned on the hour (which they won't be) and start and end at every 10 minute interval.

Say your demand is 30% so it goes on for 3 minute off for 7 minutes. Now you turn the set point up by a degree and lets say the demand increases to 50%. This may not result in an immediate change in the boiler relay status.

Lets say that you were 6 minutes into the 10 minute cycle when you increased the set point - it had already been on for 3 minutes and off for 3 minutes by that point, you have already passed the 5 minute period where it would have switched off had you already been at 50%, so you have to now wait another 4 minutes with the relay off to reach the next on period, and this next on period will be 2 minutes longer than the last one was, so it is actually a full 7 minute before anything different to the previous cycle happens.

On the other hand if you had been say 2 minute into the on part of the cycle when you made the change then the 2 minute longer on time would take effect in the current cycle. So depending on timing and what the before and after demand percentage is there might be a fairly long delay before any real change occurs.

If you use only 3 cycles per hour you double the cycle length to 20 minutes and doing so on average you would make the boiler relay twice as slow to respond to small changes in demand, because the TPI cycle period is so long.

Large changes in demand like going to full demand or zero demand aren't delayed - say you were at 30% demand, no matter whether you were in an on or off portion of the 10 minute cycle going to 100% means on all the time at any point in the cycle, so the relay will immediately come on if it was in an off part of the cycle. It's only the response time of small changes in demand that takes a while.

I would definitely stick to the default 6 cycles per hour for a normal gas boiler. Hope that makes sense.

PS what is Evologger and where did you get it ?

I also use the evohome python client libraries, but I use Domoticz and the munin-evohome plugin to do my graphing.

[mart1711]

This is brilliant info. Thank you very much!

The demand I had generated was an increase in temperature in one room from 17.5 to 20.5, which I thought would have been a big enough change to make the boiler relay come on. But maybe the overall demand was still too small...

Either way, I have now set the cycles / hour back to 6. I will also adjust the minimum on time back to 1 minute and just see how it goes.

I think I'll have to learn (like others have said before) to "trust the system" and "let it do its thing". I can fully see the merit of TPI, but it goes a little against "what feels right". As someone said in another thread - let's just hope we won't be hit with a £2000 boiler replacement bill in a few years when it turns out it didn't take too well to firing that often for such short bursts after all

I came across Evologger here: https://github.com/freeranger/evologger, also here.

It uses the evohome2 client and has modules for easy upload to Plotly and other services included. I'm running this on a Raspberry Pi. I was looking at Domoticz but haven't had the time to install yet.

I am also tempted to follow paul777's suggestion here and use optical sensors connected to another Raspberry Pi to monitor the boiler and BRD91 LEDs. But maybe - as I said above - I just need to learn to "trust the system" - so maybe it's better if I don't

[DBMandrake]

This is brilliant info. Thank you very much!

The demand I had generated was an increase in temperature in one room from 17.5 to 20.5, which I thought would have been a big enough change to make the boiler relay come on. But maybe the overall demand was still too small...

Depends what the current room temperature is though, if the room was already 20 degrees, going up to 20.5 degrees is only a small increase in demand. The HR92's use a "proportional band" of +/- 1.5 degrees, which means:

* Set point is at least 1.5 degrees below current measured temperature - always 0% heat demand.

* Set point is between 1.5 degrees below current temperature and 1.5 degrees above current temperature - proportional band - heat demand can be anywhere between 0% and 100% as required. This depends on the current error in temperature, (proportional term) the averaged long term error (integral term) and how quickly the temperature is changing. (differential term) This is a so called PID controller at work.

* Set point is at least 1.5 degrees above current measured temperature - always 100% heat demand.

Either way, I have now set the cycles / hour back to 6. I will also adjust the minimum on time back to 1 minute and just see how it goes.

I think I'll have to learn (like others have said before) to "trust the system" and "let it do its thing". I can fully see the merit of TPI, but it goes a little against "what feels right". As someone said in another thread - let's just hope we won't be hit with a £2000 boiler replacement bill in a few years when it turns out it didn't take too well to firing that often for such short bursts after all

I've been using it with an ancient 20+ year old boiler which is so old that it doesn't even have any electronics, just a pilot light and a mechanical stat and burner control, with it cycling on and off with TPI for the last year without any problems, so I doubt you'll have any problems.

TPI has been in commercial use for nearly 10 years so any recently designed boiler should be OK with it. The evohome is not the only system that uses TPI - I think all the smart thermostats these days use it.

It would be interesting to see what cycles per hour and minimum on time other people use though.

One thing to bear in mind is that if you have an S-Plan or Y-Plan config with hot water control later and only have two BDR91's you can't adjust cycle time and minimum on time like you can when you configure a boiler relay - you are stuck with the default settings.

With the hot water kit you can only get access to these settings if you use a 3x BDR91 config as I do.

[mart1711]

Again - thanks for your answer!

Depends what the current room temperature is though, if the room was already 20 degrees, going up to 20.5 degrees is only a small increase in demand. The HR92's use a "proportional band" of +/- 1.5 degrees, which means:

* Set point is at least 1.5 degrees above current measured temperature - always 100% heat demand.

Target temperature was 20.5, actual measured temperature 17.5 (as recorded in evologger data received from the API), which should have equated to a 100% demand. Curious that the relay took so long to respond.

One thing to bear in mind is that if you have an S-Plan or Y-Plan config with hot water control later and only have two BDR91's you can't adjust cycle time and minimum on time like you can when you configure a boiler relay - you are stuck with the default settings.

I do have an S-Plan config, but do not have a hot water kit. As per my other post from a few days ago, I ended up binding the (only) BDR91 as a boiler really, even though it really operates the CH valve.

[DBMandrake]

Target temperature was 20.5, actual measured temperature 17.5 (as recorded in evologger data received from the API), which should have equated to a 100% demand. Curious that the relay took so long to respond.

You're right, a setting of 20.5 with a current temperature of 17.5 should give a 100% demand. (boiler relay on constant) It does take the valve about 30 seconds to wind from fully closed to fully open, and it won't send the new boiler demand until after it finishes that - is that how long it took for the relay to react ? If so that's normal.

One thing you'll also learn when you get used to the system, is that when there is an overall low system demand (boiler on time is low) and zones are near their set points, if you make small increases in set point of say half or one degree, it can take quite a long time for the system to react to that.

The HR92 will make a small adjustment to the valve as soon as it receives the set point change (immediately if you turn the knob on the top, or up to 4 minutes if you adjust it remotely) but it can take some time for the boiler on time to ramp up enough to actually cause that room to start heating. For half a degree it can take as much as half an hour. For a full degree 5-10 minutes would be a typical reaction time.

This is partly due to how TPI works as I was explaining earlier, but also partly due to the system not wanting to "over react" to a small requested increase in temperature. It's actually pretty difficult to take a room that is nicely balanced at a certain temperature and then increase it by only half a degree without overshooting the target.

If it was to open the radiator valve a lot or ramp the boiler on time way up then by the time it detected an increase in temperature (it only samples temperatures every 4 minutes) the radiator would be really hot and would then cause the room to overshoot the target. So for small changes like that it inches its way there so as not to overshoot, but it can make it seem like the system isn't doing anything for a while. But be patient and it will get there, and with a minimum of overshoot.

I do have an S-Plan config, but do not have a hot water kit. As per my other post from a few days ago, I ended up binding the (only) BDR91 as a boiler really, even though it really operates the CH valve.

Yep I know - I was just pointing out that if you go to a 2x BDR91 hot water control arrangement later you'll lose the cycles per hour and minimum ontime setting - they're only accessible if you have a boiler control relay and you don't have one on a two relay S-Plan configuration.

Not that it probably matters because the defaults are probably fine.

[bruce_miranda]

If anything, i would say a 3 min pump overrun is too short. I have mine set to 15 mins and my pipes are still hot after that.

[DBMandrake]

If anything, i would say a 3 min pump overrun is too short. I have mine set to 15 mins and my pipes are still hot after that.

Is it really three minutes though or is it a temperature based overrun ?

Some boilers have pump overrun timers, some have pump overrun stats on the flow output - I think the latter is more common in combi boilers, at least older ones, where the pump will keep running until the flow temperature is down to a low enough temperature that stopping the pump won't damage the boiler heat exchanger. Therefore the overrun time will depend on the flow temperature and to a lesser extent the load the radiators are providing.

I have my pump overrun set really long too - about 12 minutes, but that's mainly because my old boiler has a tendency to kettle (the heat exchanger is probably a bit limescale contaminated but I've done what I can for it to little avail) so if the pump shuts off in every TPI cycle and the flow temperature is at 70+ it tends to start generating little steam bubbles when the pump stops and starts which eventually starts making things noisy. So I set the overrun long enough that it doesn't quite stop in each 10 minute TPI cycle - it will only stop when there really is no demand. As long as the pump doesn't keep stopping and starting in each TPI cycle it doesn't kettle.

But this should not be necessary on a modern boiler that doesn't have limescale problems - 3 minutes is probably fine, but again every system is different and should be treated on its own merits.

[paulockenden]

My (quite old) boiler's over-run is certainly based on temperature rather than time.