Annoying thermostat - Siemens RDJ100 - help?
Discussion
Our thermostat is proving to be somewhat irritating...
Set temperature is 19.5°C - therefore I would expect the thermostat to call for heat when the temperature drops to 19°C (I would also expect a small delay - perhaps 30 minutes or so to prevent 'hunting').
However, the indicated temperature drops to 19°C and will sit there for 3 - 4 hours before deciding to call for heat, by which time we have usually manually increased the set temperature in order to call for heat.
Similarly, when the room warms up, it ends up getting too warm - as if the setpoint is increased to 20°C then at times the indicated temperature is up as far as 21.5°C before the thermostat switches off.
Is this just a lousy thermostat, or am I doing something wrong?
I have tried fiddling with the PID settings and found "1 = 2P, 0.5K" to be optimal (we have a wet system with an electric boiler) the other settings call for heat too frequently and at the same time turn the boiler on and off at specific intervals, instead of leaving it to run and achieve it's maximum efficiency.
Explanation of PID settings:
The factory setting for control is “PID slow”, ideal for most heating systems. Optimum control can be selected if control does not provide the desired result:
2-position, 1 K
2-position controller with 1 [K] switching hysteresis
?For systems with small capacity that appear slow
?For applications requiring extended runtimes or where frequent switching causes problems
?For difficult control loops where hunting may result
Typical applications:
?Dry floor heating systems
?Heat pumps
?Electric heating with contactors
2-position, 0.5 K - - - CURRENT SETTING
?2-position controller with 0.5 [K] switching hysteresis.
?For general control situations. Provides better comfort than 1 [K] switching hysteresis.
?Can also be used for difficult control situations.
PID slow
PID control behaviour for slow heating systems that require longer minimum On times and a limited number of switching cycles per hour.
Typical applications:
?Wet floor heating systems, oil fired boilers
?Can also be used for all other types of heating applications. (Alternative setting)
PID fast
PID control behaviour for fast heating systems that tolerate a high number of switching cycles.
Typical applications:
? Electric heaters with current valve
? Gas boilers
? Fast thermal actuators
P.S. no thanks to the antiquated formatting on PH it took me ages to compose this post!!!!
Set temperature is 19.5°C - therefore I would expect the thermostat to call for heat when the temperature drops to 19°C (I would also expect a small delay - perhaps 30 minutes or so to prevent 'hunting').
However, the indicated temperature drops to 19°C and will sit there for 3 - 4 hours before deciding to call for heat, by which time we have usually manually increased the set temperature in order to call for heat.
Similarly, when the room warms up, it ends up getting too warm - as if the setpoint is increased to 20°C then at times the indicated temperature is up as far as 21.5°C before the thermostat switches off.
Is this just a lousy thermostat, or am I doing something wrong?
I have tried fiddling with the PID settings and found "1 = 2P, 0.5K" to be optimal (we have a wet system with an electric boiler) the other settings call for heat too frequently and at the same time turn the boiler on and off at specific intervals, instead of leaving it to run and achieve it's maximum efficiency.
| Parameter Description | Factory setting | Setting range | Remark |
|---|---|---|---|
| P01 | Control behavior | PID slow (4) | 0 = 2P, 1.0 K / 1 = 2P, 0.5 K / 2 = PID fast / 4 = PID slow |
| Resolution of settings and displays | |
|---|---|
| Temperature setpoint | 0.5 °C |
| Display of actual temperature value | 0.5 °C |
Explanation of PID settings:
The factory setting for control is “PID slow”, ideal for most heating systems. Optimum control can be selected if control does not provide the desired result:
2-position, 1 K
2-position controller with 1 [K] switching hysteresis
?For systems with small capacity that appear slow
?For applications requiring extended runtimes or where frequent switching causes problems
?For difficult control loops where hunting may result
Typical applications:
?Dry floor heating systems
?Heat pumps
?Electric heating with contactors
2-position, 0.5 K - - - CURRENT SETTING
?2-position controller with 0.5 [K] switching hysteresis.
?For general control situations. Provides better comfort than 1 [K] switching hysteresis.
?Can also be used for difficult control situations.
PID slow
PID control behaviour for slow heating systems that require longer minimum On times and a limited number of switching cycles per hour.
Typical applications:
?Wet floor heating systems, oil fired boilers
?Can also be used for all other types of heating applications. (Alternative setting)
| Minimum switch on/off time | > 4 minutes |
| Minimum period | Approximately 12 minutes |
PID fast
PID control behaviour for fast heating systems that tolerate a high number of switching cycles.
Typical applications:
? Electric heaters with current valve
? Gas boilers
? Fast thermal actuators
| Minimum switch on/off time | > 1 minute |
| Minimum period | Approximately 6 minutes |
P.S. no thanks to the antiquated formatting on PH it took me ages to compose this post!!!!
I have no knowledge/experience of that particular thermostat (so apologies for the inevitable anti climax following a new post notification!) but with it being a PID controller I would expect it to maintain an extremely stable output with little variation around the setpoint so something does indeed seem off.
Has it always been like this? Have you tried a factory reset? Just wondering if there's another setting hiding away that's having an impact, or if there is any algorithm values stored from a self-learning feedback loop that might be causing issues if corrupt? Even out of the box I would expect it to perform 'reasonably well' which, from the hysteresis described it doesn't look it is.
Has it always been like this? Have you tried a factory reset? Just wondering if there's another setting hiding away that's having an impact, or if there is any algorithm values stored from a self-learning feedback loop that might be causing issues if corrupt? Even out of the box I would expect it to perform 'reasonably well' which, from the hysteresis described it doesn't look it is.
Edited by tux850 on Monday 20th November 18:28
tux850 said:
Stuff.
I too expected it to maintain quite a stable output.From memory it wasn't this bad last winter, although nothing has changed in the interim...
I don't remember resetting it, but I do wonder if your theory about algorithm values could be the issue, as it does seem to behave as it it's 'learnt' something somehow.
I think I'll try a reset then!
My Mum has a Honeywell wireless stat that has similar, but not quite as bad, behaviour. It only talks to the controller every few minutes, so it does overshoot/undershoot a bit. If it's too far away from the controller it won't start or stop the heating reliably, so it sometimes the heating stays on or off.
Simpo Two said:
I have no idea what a PID setting is, but if that thermostat has a bimetal strip in it I guess it's just not sensitive enough. Does it buzz? Have you tried cleaning the contacts?
It fires the boiler for shorter and shorter periods as it gets closer to the target temperature, it may also fire the boiler when it's at the target temperature to stop it dropping below the target. In theory it should not over shoot or under shoot...
Does the heating actually turn on and off with the stat? Pondering if a valve is sticking causing the heating to stay on or not come on when it should.
Simpo Two said:
I have no idea what a PID setting is, but if that thermostat has a bimetal strip in it I guess it's just not sensitive enough. Does it buzz? Have you tried cleaning the contacts?
It's an electronic thermostat, so no bimetallic strip.PID:
Proportional - the variance between the setpoint and the current process temperature
Integral - the previous variance from the setpoint
Derivative - the predicted future variance based on previous and current variance
In basic terms the PID feature tries to anticipate when heating (or in some cases cooling) is required, based on the above parameters.
.:ian:. said:
It fires the boiler for shorter and shorter periods as it gets closer to the target temperature, it may also fire the boiler when it's at the target temperature to stop it dropping below the target.
In theory it should not over shoot or under shoot...
Does the heating actually turn on and off with the stat? Pondering if a valve is sticking causing the heating to stay on or not come on when it should.
The heating does actually turn on and off (cycle) with the stat. There are no valves involved, it's a straightforward call for heat to the (electric) boiler.In theory it should not over shoot or under shoot...
Does the heating actually turn on and off with the stat? Pondering if a valve is sticking causing the heating to stay on or not come on when it should.
There is an indicator on the boiler that shows when there is a call for heat. Likewise, there is an indicator in the thermostat.
Nether indicates a call for heat - the thermostat is apparently just "biding it's time" based on this (apparently useless) PID program.
As suggested in a previous post, I'll reset the stat today - however I suspect it's 'learning' algorithms are just poor.
Gassing Station | Homes, Gardens and DIY | Top of Page | What's New | My Stuff