Thermostat Latency?

Tips, tricks and frequently asked questions
ReeX
Posts: 16
Joined: 20 Nov 2015, 23:35

Thermostat Latency?

Post by ReeX » 20 Nov 2015, 23:45

Good evening, I have freshly installed a Netatmo Thermostat in wifi mode, so these are my first thoughts.

The product seems to be excellent! I love the design, the functionality, user interface, options etc.

I have noticed that the thermostat latency is very high, that is temperature changes are tracked very slowly. This will lead to delayed response, and delayed actions on the boiler.

For test purposes I have placed the thermostat on a bookshelf with a digital thermometer by it's side. Fluctuations are tracked quickly by the digital thermometer, slowly by Netatmo. This will lead to delays turning off/n (particularly off)

Is this normal/by design?

TypeC
Posts: 289
Joined: 30 Dec 2014, 21:35

Re: Thermostat Latency?

Post by TypeC » 21 Nov 2015, 09:02

It depends what you mean by latency. If you're expecting three readings a second, then no, you won't get that. If you're waiting 5 minutes for a change to be detected then that might be considered a problem.

You don't want your heating going on and off very frequently because it will increase wear, but (assuming you're working in PID mode) changes in the room temperature are disconnected from the operation of the boiler by a layer of abstraction. The aim is to maintain the correct temperature by operating the boiler at the rate necessary to avoid the thermostat detecting changes (or big ones, at least). If you look at it like this, a small amount of latency (which smoothes the set of readings) is not important.

ReeX
Posts: 16
Joined: 20 Nov 2015, 23:35

Re: Thermostat Latency?

Post by ReeX » 21 Nov 2015, 10:53

I understand that frequent measurements may lead to tearing on/off cycles. Some questions:
1) How can I know if the "inertia" is correct or not?
2) How do I know if netatmo thermometer is calibrated properly?

BTW I don't know if netatmo is currrently set in PID mode.....

ReeX
Posts: 16
Joined: 20 Nov 2015, 23:35

Re: Thermostat Latency?

Post by ReeX » 21 Nov 2015, 11:01

One factor that may influence this beahviour is .5 centigrades increments... Who knows if this is due to thermometer reading increments (not capable of correct measurments below .5 centigrades) or a design constraint from netatmo?

TypeC
Posts: 289
Joined: 30 Dec 2014, 21:35

Re: Thermostat Latency?

Post by TypeC » 23 Nov 2015, 15:31

The Netatmo is calibrated well if the temperature in your room doesn't fluctuate much during a normal day (with a constant set point). It is constantly adjusting it's calibration too, so you don't need to worry about it (to help, you can avoid giving it surprises like moving the thermostat around frequently). This applies to PID mode only, of course,

If you haven't yet been asked whether you want to enable PID mode, you will still be in hysteresis mode. You can check in the web app by going to Settings -> Advanced Settings -> Heating Algorithm which will be either "Hysteresis" or "Advanced" (PID) or a suitable translation for your locale.

The 0.5c increments are by design, I believe. My thermostat gives repeatable and smoothly altering readings to 0.1c (although I haven't compared this with any other devices, so it could be out by a constant) and I think this is a common precision for modern electronic thermal sensing, even at the budget end of the market.

Kraicheck
Posts: 2
Joined: 06 Jan 2016, 13:34

Re: Thermostat Latency?

Post by Kraicheck » 06 Jan 2016, 14:05

I've also noticed that the thermometer in the thermostat is really slow to register temperature increases.
I've done the following simple test:
Temperature at 20.8C on both the thermostat and a different electronic thermometer.
Request 22.0C.
An hour later the thermostat thinks it's 22.1C and stops the boiler, but in fact it's already 23.2C as read on the thermometer.
With the residual heat, the temperature continues to climb to about 24.1C 45 minutes later.
The thermostat still only thinks it's 23.4C.
The room then starts to cool off again and thermostat and thermometer converge on 23.6C, 1.6C above the requested temperature.

I've already returned a Honeywell Round for this exact problem.
This was not a learning thermostat so there was no hope of better results in the future.
I'm hoping PID mode will resolve the problem.
Even though it was probably only designed to handle the case of residual heat further heating the room, I hope it will also compensate for the slow reaction speed to temperature increases.

Why are all these thermostat thermometers so slow to sense temperature changes?
I can understand reacting slower to changes/fluctuations, but sensing should be as fast as possible.

TypeC
Posts: 289
Joined: 30 Dec 2014, 21:35

Re: Thermostat Latency?

Post by TypeC » 06 Jan 2016, 16:51

That's an interesting experiment.

In answer to your question, "Why?", I wonder if it's lack of air flow inside the case. I don't know where the sensor is in the Netatmo device, but if it's in the middle of what is essentially a large bubble (with no obvious air vents), I imagine this itself could be a factor. You'd think they'd have thought of that though!

ReeX
Posts: 16
Joined: 20 Nov 2015, 23:35

Re: Thermostat Latency?

Post by ReeX » 06 Jan 2016, 22:57

Nice debate. I'd like to add one more point: It looks like the thermostat senso is quite fast to adjust to increasing temperatures while it is slow to match lowering temps. Is it just me, or homebody else noticed this behavior?

BTW, it could make sense to remove the back cover from the thermostat? I once had the same thought...how can the thermostat react quickly if the sensor is packed in a plastic cage?

Kraicheck
Posts: 2
Joined: 06 Jan 2016, 13:34

Re: Thermostat Latency?

Post by Kraicheck » 06 Jan 2016, 23:04

ReeX wrote:It looks like the thermostat senso is quite fast to adjust to increasing temperatures while it is slow to match lowering temps. Is it just me, or homebody else noticed this behavior?
For me it's actually the opposite of this.
It's just slow in sensing changes. Because heating up goes a lot faster than cooling down, it can't follow the normal thermometer but follows it without problem when the temperature drops naturally when the heating is turned off.
ReeX wrote:BTW, it could make sense to remove the back cover from the thermostat? I once had the same thought...how can the thermostat react quickly if the sensor is packed in a plastic cage?
The thermostat doesn't work without the back cover :-(

ReeX
Posts: 16
Joined: 20 Nov 2015, 23:35

Re: Thermostat Latency?

Post by ReeX » 07 Jan 2016, 00:10

Kraicheck wrote: The thermostat doesn't work without the back cover :-(
Did you try this?

Post Reply

Return to “General questions”