Control of domestic night-storage heaters

PhilHornby

PhilHornby

Senior Member
I'm intrigued as to the improvement this yields - surely it's altering the 'charge time', based purely on the room temperature at the time it's charging (i.e the middle of the night)? This is what my XLS24N does (albeit using a mechanical thermostat, positioned in a stupid place).

But what you really need, is to be able to set the 'charge time', based on a temperature at some other point in time. Ideally, this 'point' would be sometime tomorrow(!), but maybe the previous minimum, or average of the previous cycle would be better than nothing. Also, if the heater is only going to be on for (say) 3 hours, you'd want those to be the last 3 hours of the economy 7 period, rather than the first 3.

I've rather given up on improving the XLS24N - it has two shortcomings. 1) is the inability to keep the heat in, until I want to release it (the insulation is not good enough) and 2) the inability to get the heat out fast enough, when I demand it (no fan). (I do have a piece of bent metal, that can be inserted to lift the internal flap, when the heater doesn't otherwise want to cooperate o_O )

Wrt to wireless thermostats in general though, I agree they're like hen's teeth. I couldn't find any (at reasonable prices) to control my wall-mounted convectors - so I used Picaxe 14M2's, talking to commercial (Energenie) RF sockets. Every house should have a heating system controlled by 10 Picaxes :)
 
B

Baker Steve

Member
Long answer - sorry. I agree with all that you say: my 'dream' system would take into account the predicted temperatures for the next few days drawn from the Met Office's website. However, after The Beast From The East last February I was looking for a simple quick solution. My observation only is that there is improvement over trying to second-guess the weather and using the storage heaters' own controls (which, I omitted to say, I set to max charge and min output, as control is now coming from the wireless stat).

When thinking this out, I'd assumed that a wireless stat would emulate a conventional bimetallic mechanical one. Obviously such a stat could only influence the storage heater(s) during the off-peak period, which in my case is 11-12 pm and 1-7 pm. My (faulty) reasoning was this. There would be four possible states during that period: 1. Zone temperature remains below command temperature during whole period; 2. Zone temperature starts below command temperature but rises above it during charge period; 3. Zone temperature starts above command temperature but falls below it during charge period; 4. Zone temperature remains above command temperature during whole charge period. The results would be: 1. Full charge; 2 and 3. Partial charge; 4. No charge. Nothing could happen outside the charge period v-a-v storage heater charging, obviously. All I expected was an improvement, not accurate control, as clearly the stat cannot do anything useful outside the charge period.

However, I am seeing better control than I expected, although I have yet to track operation through a whole winter. Firstly, the Honeywell wireless stats don't work in the same way as a mechanical stat, they work as PID (proportional–integral–derivative) controllers (see Wikipedia entry although I expect you understand these things: I didn't before). Certainly the instructions use the term 'intelligent control', and there's way more electronics in them than you'd expect, including a crystal on the room-stat circuit board. My guess is that, as the room-stat part is battery operated, it tracks zone temperature outside the charge period, which will influence its error feedback signal internally. If the zone temperature is below the command temperature during the day (which, so far, it isn't, as the central heating, which is downstairs only, will be on), the room-stat will be sending 'on' commands to the relay module. But of course nothing will happen, as the relay modules is on the off-peak circuit, which is unpowered. Thus, when the heaters are powered on, they will take a larger charge. As the opposite is true so far, the room-stat part is sensing that for most of the day the zone temperature is above the command temperature: as expected, this results in the slightly counterintuitive behaviour of the heaters not being switched on immediately when power is restored even if the zone temperature has fallen below the command temperature.

Of course the above explanation could be wrong, but all I claim is that by observation the heater charge is inversely tracking the outside temperature and that I get much better zone temperature control than before. I've attached a graph, which shows outside temperature, bedroom temperature and the surface temperature of the two storage heaters (one in the bedroom, one just outside on the landing) at approximately 8 am each morning. Each storage heater has its own stat although they are physically close. This is a detached house on an exposed site close to the sea. In the pre-wireless-stat days, I could see a bedroom temperature change of even as much as ten degrees C if a cold southerly gale set in unexpectedly.

It is quite possible that when the weather gets a lot colder the two 1700 watt storage heaters won't be able to cope, and then of course control will be lost. I'll make a note to post an updated graph at the end of the winter. Incidentally, my electrician tells me that 'optimizing' storage heaters will become legally mandatory soon for new installations, if they haven't already.
 

Attachments

PhilHornby

PhilHornby

Senior Member
Well, hats off for attempting to measure this - one measurement being worth a thousand opinions, and all that :)

I suppose we shouldn't be too surprised, that this yields an improvement, given that it's a more sophisticated version of what Dimplex did, to arrive at the XLSN series. It's not a strategy that works for me, because there is additional heating in the room with the Storage Heater (so it's always still warm at 11pm).

Incidentally, the claim of "PID" control, is something that I've seen in other On/Off Room stats. (such as the Siemens REV24 that I use back at base for my gas Central Heating). Given that the "P" in PID stands for proportional, it's baffling how they can use the term in relation to something that only contains a simple 'switch' :unsure:. (I know the Siemens unit 'learns' the room characteristics, to try and avoid overshoots (and undershoots)). Maybe the Honeywell unit is similar.

Dimplex's latest "LOT20" compliant Storage Heater is the XLE Range; apparently, "XLE storage heaters offer dynamic storage calculation". There seems to be a considerable overlap with their Quantum range...
 
Last edited:
B

Baker Steve

Member
I only believe in things I can measure - I like graphs :)

When considering replacing the heaters a few years back, I looked at the Quantum range, but the problem for me was that they required a permanently live supply in addition to the off-peak, which would have involved me in a *lot* of very costly rewiring, ruling them out. It looks as if this applies to the XLE range too (and they're ugly).

I think the 'P' is somewhat justified in the case of the Honeywell stats, as the way in which they seem to work is to divide each hour into six ten-minute periods, then switch on the load for a variable proportion of each period depending on the amount by which the ambient temperature is below the command temperature.

It will be interesting to see the performance of this setup as it passes through the colder part of the winter (although I sincerely hope that it won't be as bad as the last one!).
 
B

Baker Steve

Member
I'd advise against those.I actually use the UK version as control for backup boost heating (with an oil-filled electric radiator), but the temperature control is pretty crude. Plus the remote eats batteries.
 
You must log in or register to reply here.
Top