The ‘MyVaillant’ (Android) phone app is reporting an outside temp. of 7.5C when two other references say it’s about 2C, and sticking my head out the door confirms the latter. Some time back I put an offset in for the OAT of -1 which seemed to work. Now it’s colder and the reading’s way-out.
The important thing is the heat pump is performing as-per its weather compensation curve of 0.4 (currently), giving a flow temp of 32C at 3C outside, so I’m not too ‘bovvered’ but wonder if the app reports some kind of 24-hour average ? If the sensor is faulty then surely the unit wouldn’t be giving the right flow temp?
Any ideas?
Anyway, as this is its first cold spell it looks like ‘Baby’s First Defrost’ - aaaahhhhh. I’ll get me coat.
Yes, I’ll have to look deeper. Right now it’s not important as I’m running on ‘pure’ weather comp and all seems well, so as you say it must be measuring outside temp accurately somewhere!
It’s been a few months since I had this issue and (from memory) I checked the terminations of the cable in the outside sensor as well as the ‘other end’ in the controller/interface, and it’s working as expected now, so I suspect a faulty connection somewhere. The sensor itself looks like a simple thermistor working on very small current so any bad resistance will have a large effect. Hope this helps.
How long? I think it’s a 10k NTC like all of Vaillant’s temperature probes. Even a long wire would add so little resistance to not really skew the results (we’re talking less than an Ohm in series with around 10kOhm). I don’t think there’s any shielding.
How much does your sensor overread? Mine is on the north side of the house but does get sunlight from the side in the morning, which does inflate temperature significantly in the summer.
The probe is mostly accurate until the temperature drops below ~10C; only then does it begin to over-read. For example, the other night we hit 3C overnight, but the sensor didn’t read below 8.
The sensor is mounted on a wall in the side passage gap between two houses, cable run ~25m. I placed a Thermopro bluetooth sensor next to the Vaillant sensor, and checked the readings periodically overnight.
Even when comparing to the ‘Inlet Air Temp’ reading on the VWZ AI controller, whenever the inlet temp is below 10C there is a discrepancy.
This definitely sounds like a defect. For me the temperatures all match relatively closely and I don’t have any significant deviations from my other sensors. Maybe a contact issue?
I’m now of the belief that the temperature reading discrepancy is simply being caused by the Vaillant outdoor sensor being mounted to a solid, uninsulated brick wall, which is at above-ambient temps due to heat loss, with the delta increasing as outside temperatures decrease.
This explains why the discrepancy becomes more apparent when actual outdoor temp hits sub-10C.
I will be moving the sensor away from the wall, and fitting it inside a small Stevenson shield. Will report back with results.
I totally agree with your diagnosis. As outside temps fall below around 2 degrees C, the discrepancy between actual and measured temp increases, due to the mass temp of the wall the sensor’s attached-to. I re-mounted the sensor on an insulating spacer and the delta is now within 0.5C. Before at sub-zero temps the ‘offset’ range of 3C wasn’t enough to compensate. I should say the wall in question is standard 1930s brick/cavity/block with 1970s foam cavity fill, so reasonably well insulated.
The Vaillant sensor is not well protected against radiative influcences from the environment. See this comparison of one of my own outdoor sensors, Vaillant’s sensor (ebus) and the nearest weather station (ignore the oscillations, those are weird interpolation errors). The strong dips in Vaillant’s sensor appear with very clear night skies. Current sky temperature measured by IR thermometer is -20 to -30 °C, whereas the air temperature is still above 0.
My sensor is on the north wall and essentially faces into the sky, hence the low measurements. If it’s in a more closely populated urban environment and close to warmer surfaces even if not in direct contact, this will skew the temperature upwards. It appears I’ll need a Stevenson screen fitted over the sensor.
The other day, I mounted my outdoor sensor inside a Stevenson shield, and decoupled it from the wall by mounting the assembly against a waste pipe instead, wrapped in a strip of Chinese ‘aerogel’ insulation.
The sensor now reflects the actual OAT much better, but it’s still not perfect as it’s located in a narrow alleyway between two houses, both with solid brick walls built c.1890, which I’m assuming is having a ‘canyoning’ effect and warming the air slightly. I’ll try repositioning onto a more exposed side of the house.
