Hey all, I have an Air/Sky combo, but I’m curious how the Tempest, when placed in full sun, is supposed to get accurate temperature readings? Since there are no moving parts, it doesn’t have a fan to circulate the air, so how does it not read a higher temperature than is actually the case? My Air is cited in full shade and aside from being a little off on very cold nights (the trees above keep the temp a degree or two warmer than it should be), it is extremely accurate. Seems like keeping the sensors separate might’ve made more sense.
it will measure the temperature at its location. But as the unit is under influence of the sun and the wind, the unit itself will likely change temperature. Luckily the unit measures the sun and the wind, and by knowing all the physics of the unit, it compensates for the sun and the wind, so the temperature stays accurate. If you placed the unit at the required height in an open field, it will be very accurate. No need for a box to shield it from the sun, or a fan to circulate the air inside.
So an algorithm is compensating for the inaccurate temperature reading? That makes it seem like the unit is only as good as the algorithm…
but the algorithm is based on the physics of the unit. it should be fine.
I mean, I guess…I’m a meteorologist and weather models are based on physics, but they’re still wrong a lot…
yeah, but that is because weather is a kind of chaotic problem. That’s totally different then calculating the heat absorbed by the unit due to sunshine.
anyway, if it wasn’t working as expected, we would see a flood of messages in this forum of people complaining about it. But having a separate Air/Sky unit has the advantage that you can measure the temperature at a place you actually want to know the temperature.
Keep in mind, the Tempest is a medium range consumer product.
Totally understand Gary! I’m just wondering out loud, in theory, does the Air sensor itself measure temperature during the day more “accurately” (i.e. with no external AI modifying the values) since it is in the shade, than the Tempest would in full sun? I think the answer is yes?
I think so too. I have two Air units in shade. One 12 inches above ground level and one 12 feet above that.
Still awaiting info from WF on how we can tell how much their magic math is adjusting what the Tempest reads. For those of us comparing multiple sensors that would be really helpful to know.
I added a DFARS to my VP2 yesterday to try to ensure its accuracy in full sun, and definitely believe it far more than the adjusted Tempest temperature. Tempest is reporting 7+ degF lower than the VP2 in full sun today. Tempest at 75, VP2 reports 82+. It’s every bit 82 here today for sure.
I’ve seen this this mentioned in another post. This was really noticed by owners of the Tempest after the latest firmware update.
I guess my point here is that if math it all it takes to get an accurate temperature reading in full sun, why haven’t other companies done the same and saved money by selling cheaper weather instrumentation that could be marketed to more people?
I disagree with this because I don’t know many people that can tell the difference between 82 degrees and say, 85 degrees.
Do others have all the sensing equipment necessary in the same unit and profiled the thermal properties of that unit? It may not be as cheap as it sounds, just like software isn’t cheap even though it is cheap to copy.
I had both a sky and tempest for a time. The air measured within .2 of my tempest even on hot 95-100 degree days. That’s good enough for me.
why don’t other companies have ultrasonic wind sensors in their consumer grade weather stations? why don’t other companies have haptic rain sensors? I guess they just aren’t that innovative. Besides many weather stations don’t have the necessary sensor for brightness and I guess many others go for an even cheaper solution… just having an external temperature sensor. (you can buy a complete Bluetooth thermometer for slightly more than 2 euro https://www.aliexpress.com/item/4001221135164.html)
I feel like you all are taking this personally. That’s not my intention! I guess it just bothers me that the sensor relies on algorithms rather than a true estimate of the air temperature as read by the sensor itself. Math can be wrong sometimes when trying to model thermal properties. As a meteorologist, I prefer true measurements (and yes, I realize even a $100,000 ASOS temperature sensor’s error is +/- 1.8 deg F) to algorithm-modified ones.
but you appear to think that it is inaccurate or don’t trust the used physics to get to the temperature. As vtyler98 just showed… in his case it is within 0.2 F of the measurement of the air unit.
I don’t think anyone would argue that the Tempest will be able to consistently out perform a properly shielded and vented stand-alone temperature sensor using its temperature compensation algorithm.
That being said, the vast majority of weather stations at this price point are generally all-in-one devices. I would assume that there is sufficient market research to show that this is the preferred option over having separate devices. While these all in one devices often have radiation shields to help mitigate against solar heating (including the Tempest), they are always going to struggle when the temperature rises and the wind speed falls. WeatherFlow could have decided that this was simply a design limitation and accepted that the temperature recorded by the Tempest would be artificially high compared to the “true temperature”. Instead, they have decided to use their knowledge of the thermal properties of the device and the observed solar radiation/wind speed to adjust the Tempest temperature closer to the “true temperature” when the in-build solar radiation shield is overwhelmed. Overall I think this is probably the best approach available for all-in-one devices.
The downside of this approach, however, is that one poorly calibrated sensor (such as the solar radiation) will begin to impact the quality of the observations from another sensor (in this case the temperature). It could well be that variability in the quality of the solar calibration could account for the accuracy differences seen between different stations.
Of course the best of both worlds would be an option to turn this correction off if you wish, and/or a clear indication of the magnitude of the adjustment it is making. This way each user can choose the approach that they like. And perhaps in the future WeatherFlow will provide a new stand-alone temperature sensor that can be properly shielded for those that like/require the associated accuracy.
Doesn’t the Tempest also uses other surrounding weather sensors to calibrate the temperature? If you have an Air sensor near it, the system knows that and uses one to modify the other.