HONEYWELL DT90E ROOM THERMOSTAT

[MJN]

Of course... That makes more sense!

 

[slj33]

thank you all,

so leave the proportional band width on 1.5 ?

 

[vicvapour]

thank you all,

so leave the proportional band width on 1.5 ?

Yes, or....

Change your boiler to a fully modulating burner boiler which means it adapts the output according to the heating demand without constantly switching itself on and off. This results in less cycling and increases reliability and efficiency in gas and electricity consumption - saving you money both in the short and long term. For e.g. Baxi Platinum Combi HE A

 

[D_Hailsham]

so leave the proportional band width on 1.5?

Does the actual temperature rise above the set temperature (overshoot) and is your house very well insulated? If so try increasing the PB, a bit at a time, until the overshoot stops.

 

[slj33]

i have it set for 18.5 and it always reaches 19 but no more. (its sited on the landing - not ideal i know!)

the house is well insulated (max in roof, cavity wall, double glazing)

goes from 19 at night to 18.5/18, in morning so not losing much overnight.

so maybe ill set the p band width to 2 and see how that goes.

the oil settings of cr3 and on/off 4 min seem to be better suited to this gas boiler though.

 

[MJN]

To illustrate, if your target temperature is 21C and the proportional bandwidth setting set to 1.5C then the proportional window is therefore 20.25C to 21.75C (or is it 19.5C to 22.5C? D_Hailsham needs to mark this homework - he seems to know these stats inside out!).

According to Honeywell's explanation of TPI, if the PB is 1.5C and the required temperature is 21C, then the band is from 19.5C to 21C. There's no point having the band going above the set point!

I was thinking about this today and something struck me... If we are saying that the proportional bandwidth is the size of the window below the set point then that means if the room temperature was above the set point, but falling, then the stat would only fire once the room temperature fell to, or below, the set point. Such an operation methodology would likely suffer from undershoot as a result.

To illustrate, let's say the proportional bandwidth is 1.5C and the set point 21C. If the room temperatute is 23C (let's just say from solar gain for argument's sake, or even overshoot) yet is now falling... If the stat behaves likes a normal on/off stat outside of the proportional bandwidth then, by implication, it will remain off at 23C. At 22C is will stay off. At 21.5C, 21.4C, etc.... it'll stay off until it hits 21C. By this time we are potentially too late to keep maintain our target because of the inherent heat delay.

I therefore would expect that the bandwidth is the size of the window surrounding the set point (whether that is 1.5C above and below, or 0.75C either side I don't know; probably the former) so that even if/when the room temperature is above the set point there may still be demand for heat (at a level in accordance with the proportional part of TPI control). In order to ensure minimum overshoot this is where the integral aspect of the function would come into play (i.e. self-adjusting in response to the measured effect of the heating system).

Do you recall where you saw Honeywell's definition of the bandwidth being the window beneath the set point, and not surrounding it?

Mathew

 

[MJN]

Do you recall where you saw Honeywell's definition of the bandwidth being the window beneath the set point, and not surrounding it?

Answering my own question, this presentation says it is beneath the set point on pages 48 and 50. Hmm.... not sure what to make of that!

Mathew

 

[D_Hailsham]

Let's say the proportional bandwidth is 1.5C and the set point 21C. If the room temperatute is 23C (let's just say from solar gain for argument's sake, or even overshoot) yet is now falling... If the stat behaves likes a normal on/off stat outside of the proportional bandwidth then, by implication, it will remain off at 23C. At 22C is will stay off. At 21.5C, 21.4C, etc.... it'll stay off until it hits 21C. By this time we are potentially too late to keep maintain our target because of the inherent heat delay.

If the temperature has risen above the set point due to overshoot, then the PB is too narrow.

Page 12 of the presentation says that the temperature control accuracy is ±0.5K at 20°C. I'm unclear what they mean by this. Do they mean that it will maintain the temperature within ±0.5K of the set point or do they mean that, when the stat reads 20°C, the actual temperature could be 0.5°K higher or lower. There's a subtle, but important difference. The display only moves in steps of 0.5° and there's no way of knowing at what actual temperature it changes. My guess is that 20° means it's somewhere between 19.75° and 20.25°.

Although the display only moves in 0.5° steps, the sensor is obviously more sensitive than that, otherwise it would not be able to provide a fine enough control in the PB. If 1.5° below set point means on 100%, 0.5° below is 33% on. So 10% on is equivalent to the temperature being 0.15° below the set point.

 

[slj33]

this system seems to hold the temp well but ...... it does seem hard for the ordinary householder to understand????

when i bought this stat the heating engineer fitted it and left it on the recommended settings for the 15 yr old gas boiler - but these seemed totally unsuitable! two engineers returned to try and work it out - no luck! they didnt understand it much more than me. now set to oil boiler settings.

my temp 'overshoots' by 0.5 (if that IS an overshoot??), i increased the pb to 2 but it didnt seem to make much difference - temp set to 18.5, still firing at 3 cycles 4mins min - four hours later @ 19 degrees!

surely if the pb was - set point and BELOW, - mine would not fire at all when the temp was 19????? but it does! so maybe it is above and below the set point?

does this help ? http://www.fitathermostat.com/pdf/FAQ-TPI-Control.pdf

would really like to set the thing and forget about it !

thanks to all for the replies so far.

 

[mikely]

I therefore would expect that the bandwidth is the size of the window surrounding the set point (whether that is 1.5C above and below, or 0.75C either side I don't know; probably the former) so that even if/when the room temperature is above the set point there may still be demand for heat (at a level in accordance with the proportional part of TPI control). In order to ensure minimum overshoot this is where the integral aspect of the function would come into play (i.e. self-adjusting in response to the measured effect of the heating system).

Do you recall where you saw Honeywell's definition of the bandwidth being the window beneath the set point, and not surrounding it?

Mathew

I think you are quite right Matthew. The integral part of the controller positions the proportional band about the set point. D_H and I discussed this in a thread a while back. See //www.diynot.com/forums/plumbing/honeywell-cm927-thermostat.297354/

 

[MJN]

Thanks Mikely; an interesting read.

Mathew

 

[vicvapour]

(off topic a tad)

MJN, (my good sparing partner), what is this I rad about your internet cat feed, ive got two cats, tell me more.

 

[MJN]

Check out http://www.newtonnet.co.uk/catfeeder where all will be revealed!

Before you say it, yes, I probably do have too much time on my hands...

Mathew

 

[vicvapour]

Check out http://www.newtonnet.co.uk/catfeeder where all will be revealed!

Before you say it, yes, I probably do have too much time on my hands...

Mathew

Nice one im impressed, are these for sale of is this just a prototype for yourself.

I thought i was good with my ipcam that i can log into from any pc or smartphone any, this is so much more practical.

 

[MJN]

Nice one im impressed, are these for sale of is this just a prototype for yourself.

People have expressed an interest but to be honest it would end up being quite expensive in its current design. To make it commercially viable would likely require a different approach. More to the point, I really just built it for my own interest/benefit as I've always been into working out how things work and using that knowledge to solve real-life problems. Much to my parent's annoyance at the time I was always taking apart anything a screwdriver would allow me to from a very early age and so this is just a natural evolution of that!

I have made some money on the back of it by designing and building a not too dissimilar device (but fully automated) for the BBC to try and train crows to pick up litter at a service station but again it wasn't really for the money - more for the challenge and opportunity to solve another problem.

I thought i was good with my ipcam that i can log into from any pc or smartphone any, this is so much more practical.

You will know the satisfaction of first viewing a live camera stream remotely as even though you might know exactly how it all works, actually seeing the practical results of the theory still holds a certain something. The cat feeder is just a natural extension of that by also being able to interact with what you are seeing!

Mathew