wall thermostat for electric storage heater

buzz11

I want to connect a room thermostat to my storage heater so that it turns off when the room reaches the desired temperature. (I know that the storage heater will only be live at night)

The storage heater is a Creda TSR24MW 3.4KW Manual, so what current rating does the room stat need to be?
I'm looking at a basic dial thermostat that can handle 15a....would that be suitable?

2011

Normally thermostats only switch small loads.
In this case I would reccomend that you use the thermostat to switch a contactor that in turn switches the storage heater that you intend to control.

The load in this case is about 15A.
I always use a larger contactor than required as this extends the life of the unit and increases overall reliability.
So I would suggest using a 20 or 25A contactor.

Connections should always be tight!

buzz11

Excellent, thanks for that.

Its just one heater in a room so I'm hoping it will prevent the massive overheating that comes with predicting Irish weather

2011

No bother.

Remember: The stat should be protected by a suitabely rated MCB. So if the stat is rated to switch 6A then the MCB should be rated for no more than 6A.

Working on a distribution board and making a modification such as this should only be done by a qualified electrician IMHO.

M cebee

that was the old method of adjusting the input to heaters
.stat and contactor

-im not certain that a room stat will regulate the input correctly on a storage heater as you envisage

leeomurchu

M cebee wrote: »

that was the old method of adjusting the input to heaters
.stat and contactor

-im not certain that a room stat will regulate the input correctly on a storage heater as you envisage

Agree storage heaters aren't a very precise form of heating a room perhaps a time clock on the heater might be a better idea.

Have storage heaters not got a built in stat anyway?

M cebee

the devireg will adjust the input to the weather outside
im not sure that a room stat will work well

2011

As stated above, on off control such as the OP is suggestion will not work well with storage heaters.

For example:
Say the room stat has a set point is 20 DegC.
Although the stat will switch off the heater when it registers this 20 DegC the heater will continue to emit heat after it is switched off, which could raise the room temperature to a peak of say 25 DegC. Eventually the room will cool and the stat will "call" for heat again once it registers a temperature of less than 20 DegC. When this happens the heater will be switched on once more. However it can take quite some time for the heater to emit heat once more to the point that the room temperature starts to increase (due to the large time constant of storage heaters). During this lag time the room may cool to 15 DegC. So despite the fact that the desired room temperature is 20 DegC the actual temperature may fluctuate either side of this from 10 to 25 DegC.
This is illustrated on the graph below.

A time clock is not going to help this situation at all as the temperature will still overshoot and undershoot the set point.

If this were an industrial application that had to use storage heaters and these fluctuations had to be eliminated a PI or PID type controller may well be used.


Despite the fact that this on off control is very crude, some control is better than none.

My experience with storage heaters has always been very negative. They are always expensive to run and the end user is always too hot or too cold. If they were mine I would bin them.

Bruthal

M cebee wrote: »

the devireg will adjust the input to the weather outside
im not sure that a room stat will work well

Room stat wont work properly at all. All that can be done is regulate the input to the storage heater, through a stat in the heater, or outside temp stat probably.

It has to be sort of predicted what the weather/temperature will be like the next day, to even decide if the heater will be required, so controlling it with a room stat is a waste of time and effort.

Bruthal

2011 wrote: »

As stated above, on off control such as the OP is suggestion will not work well with storage heaters.

For example:
Say the room stat has a set point is 20 DegC.
Although the stat will switch off the heater when it registers this 20 DegC the heater will continue to emit heat after it is switched off, which could raise the room temperature to a peak of say 25 DegC. Eventually the room will cool and the stat will "call" for heat again once it registers a temperature of less than 20 DegC. When this happens the heater will be switched on once more. However it can take quite some time for the heater to emit heat once more to the point that the room temperature starts to increase (due to the large time constant of storage heaters). During this lag time the room may cool to 15 DegC. So despite the fact that the desired room temperature is 20 DegC the actual temperature may fluctuate either side of this from 10 to 25 DegC.
This is illustrated on the graph below.

A time clock is not going to help this situation at all as the temperature will still overshoot and undershoot the set point.

If this were an industrial application that had to use storage heaters and these fluctuations had to be eliminated a PI or PID type controller may well be used.


Despite the fact that this on off control is very crude, some control is better than none.

None of that works with storage heaters. They are designed to heat up but retain their heat at night, and release during the day, so the room stat will not work properly at all.

Temperatures overshooting set points applies to real time heating. A storage heater is supposed to be able to regulate the heat being released, well after the electrical input has ceased, and this cant be regulated by a room stat, only by mechanical means.

A time clock will actually work better, as it can be timed to come on at 4am and off at 9am as an example, to enable the storage heater to heat to around half capacity etc.

But that also is crude. Storage heaters themselves are crude, their main intention being to even out the base load generated over 24 hours.

2011

robbie7730 wrote: »

None of that works with storage heaters.

+1
As I said: My experience with storage heaters has always been very negative. They are always expensive to run and the end user is always too hot or too cold. If they were mine I would bin them.

They are designed to heat up but retain their heat at night, and release during the day

In theory they are yes, in reality things work quite differently.

Storage heaters themselves are crude, their main intention being to even out the base load generated over 24 hours.

+1
But they never deliver this.

Bruthal

2011 wrote: »

+1
As I said: My experience with storage heaters has always been very negative. They are always expensive to run and the end user is always too hot or too cold. If they were mine I would bin them.


In theory they are yes, in reality things work quite differently.


+1
But they never deliver this.

Indeed

buzz11

I agree with above observations about storage heaters, they appear to be designed for 'consistent' climates where it might stay below/near freezing all winter.

Anyhow my purpose with the room stat is to crudely eliminate room-overheating which is a regular event, but with heat emmission from the storage heater being so delayed and limits by night clock, I don't know how it will work in reality.

The application is an office so people rarely check the settings, except to turn them to max when a few days are cold, then when mild weather comes nobody turns down the controls and the place overheats. This overheating is what I hope the room stat may reduce.

2011

Perhaps the simplest solution is to replace the storage heaters with panel heaters.

The stat will not help you maintain a set point during the day as the supply to the heater will be off.

However panel heaters can be powered on during the day and if controlled by a stat can provide reasonable temperature control. Panel heaters are also available with built in timers, so you can ensure that they are off during the hours that the office is empty.

M cebee

2011 wrote: »

Perhaps the simplest solution is to replace the storage heaters with panel heaters.

The stat will not help you maintain a set point during the day as the supply to the heater will be off.

However panel heaters can be powered on during the day and if controlled by a stat can provide reasonable temperature control. Panel heaters are also available with built in timers, so you can ensure that they are off during the hours that the office is empty.

prob not for office hours

storage would be cheaper and better than panel heating

a timer will be ok as robbie said -set for the end of the off peak period

if i recall the devireg also outputs similarly at the end of the off-peak period
http://www.devi.co.nz/index.cfm/3,124,350/710.pdf

leeomurchu

Eh hmmm did I not say time clock :)

2011

M cebee wrote: »

storage would be cheaper and better than panel heating

It would be cheaper in terms of each unit of electricity costs less.
But what I have found is that the storage heaters give out quite a bit of heat when powered up i.e at night when nobody is present. Then when the power is off they give out all of the stored heat too quickly. After lunch time they are cold. Messing around with the dampers helps, but not much.

I have found that they dump so much of their heat in the morning that people open windows to cool down. After lunch they either freeze or plug in/switch on other heaters at peak rate. This can negate any saving made by using the off peak rate to heat the storage heater.

Perhaps other people have had better experiences than me.

This is why I would prefer to use a panel heaters. Although they may be using peak rate electricity the control is so much better that savings can be made.

a timer will be ok as robbie said -set for the end of the off peak period

Yes, I see what he is trying to achieve but this is about using a timer to save money rather than to keep the temperature at a set point.

Bruthal

2011 wrote: »

Yes, I see what he is trying to achieve but this is about using a timer to save money rather than to keep the temperature at a set point.

The less heat put into the storage heater blocks, the less hot the room being heated will likely be.

It will do the same thing as the input stat really, except it could control more than one heater.

Bruthal

2011 wrote: »

The on/off square wave needs to move a quarter wavelength to the right in that diagram imo, or does it?

2011

robbie7730 wrote: »

The on/off square wave needs to move a quarter wavelength to the right in that diagram imo, or does it?

No, the diagram is accurate.

Remember the room temperature continue to increase for a time after the electrical supply is switched off.

2011

robbie7730 wrote: »

The less heat put into the storage heater blocks, the less hot the room being heated will likely be.

True.

Less heat into the storage blocks could be an issue on very cold days as it could mean that the room never gets to the desired temperature.

It will do the same thing as the input stat really, except it could control more than one heater.

Not really.
A stat takes ambient temperature into account which ensures that the supply to the heater will only be switched on when the ambient temperature is below the set point.

On the other hand the timer will switch the power on to the heater at the appropriate time regardless of the temperature.

Bruthal

2011 wrote: »

No, the diagram is accurate.

Remember the room temperature continue to increase for a time after the electrical supply is switched off.

The diagram shows the stat going on and off at exactly the same temp, which is a problem for stats. They need a hysteresis setup, something inherent in mechanical stats, or they could rapidly switch on and off at the set point,.

If that diagram is correct, then none exists for the stat or stats the diagram is based on.

2011

robbie7730 wrote: »

The diagram shows the stat going on and off at exactly the same temp, which is a problem for stats. They need a hysteresis setup, something inherent in mechanical stats, or they could rapidly switch on and off at the set point,.

You are correct, stats are designed (and do) have an amount of hysteresis for the reason you stated.

However I think that the diagram should be taken for what it is: A basic sketch of a graph that is designed to illustrate the limitations of on/off control.
A more detailed graph should show exactly what you are saying.
You could also argue that it is highly unlikely that the temperature would rise at exactly the same rate that it would fall, so the graph is inaccurate from that aspect as well

When I said that the diagram was correct it was in response to your post stating that:

The on/off square wave needs to move a quarter wavelength to the right in that diagram imo, or does it?

Bruthal

2011 wrote: »

When I said that the diagram was correct it was in response to your post stating that:

he on/off square wave needs to move a quarter wavelength to the right in that diagram imo, or does it?

If the on/off part was moved to the right, that would then show a hysteresis in the stat. Off would be just above set point, and on would be just below it.

Bruthal

2011 wrote: »

Not really.
A stat takes ambient temperature into account which ensures that the supply to the heater will only be switched on when the ambient temperature is below the set point.

On the other hand the timer will switch the power on to the heater at the appropriate time regardless of the temperature.

Do the same as the input stat in terms of varying the on to off time during charging. I think everyone on boards knows how a stat works compared to a timeclock.

2011

robbie7730 wrote: »

If the on/off part was moved to the right, that would then show a hysteresis in the stat. Off would be just above set point, and on would be just below it.

If you moved the "on/off square wave needs a quarter wavelength to the right" as you suggested the graph would be incorrect.

It would incorrectly show that once the stats switches off the temperature immediately stops increasing. It would also show that once the stat switches the supply back on the temperature immediately stops dropping, again this is incorrect.

I would think that the purpose of the graph is to demonstrate the limitations of on/off control.

leeomurchu

Goin around in a nice big circle there lads :eek:

2011

leeomurchu wrote: »

Goin around in a nice big circle there lads :eek:

Yeah, you are right. I'm off to my bed

Summary:
1) Don't use storage heaters or expect good control from them with a stat and/or time clock.

2) Don't use crude graphs to illustrate a point :D

Bruthal

leeomurchu wrote: »

Goin around in a nice big circle there lads :eek:

This is war. Get out now or become collateral damage:)

leeomurchu

robbie7730 wrote: »

This is war. Get out now or become collateral damage:)

Feck that I'm takin me ball and goin home, good luck

Bruthal

leeomurchu wrote: »

Feck that I'm takin me ball and goin home, good luck