Display Temperature to the hundreths of a degree

Got it. I also use ‘wfpiconsole’ as well as the App. . .which formats temps to nearest tenth of a degree.

Same thing. It’s an external source formatting the raw data in a way that’s most useful for an end user.

A weather station such as the tempest with its 3 second wind data and temperature to an extra decimal point of precision does have a use. Most weather enthusiasts do not analyse or have much reason to study thermals. But us paraglider pilots do have a use. I currently use the one minute wind readings to know where a thermal is lifting off from near my Sky at our paragliding launch. If I was in the tow paddock then the thermals are building and breaking off regularly and knowing their pattern is really helpful for knowing when to tow up. It doesnt matter if the actual temperature is a few degrees in error but knowing the change in temperature to fine precision can have its use, even if it is not ‘accurate’ and has a delay due to thermal mass.
cheers Ian :slight_smile:

1 Like

you must be a super duper professional paraglider if you ever are going to make a decision based on an 0.01 degree temperature increase. :wink:

Thank you this is exactly why I suggested it, not so much for the actual, but for the trans data in the acceleration or deceleration of that trend

It can show a trend that is otherwise invisible. And I neglected to mention the pressure trend is also valuable. In the future some one may use that information to map the thermals. But I already do by using the 1min wind. Usually as glider pilots we watch many other signs because this data has not been as easy to gather and access as it is now. For example I know Penrith rowing have several along their course. They may find benefit in understanding how these variables can forecast short term wind changes based on thermal behaviou and prediction using the current days behaviour. Ther temp gradually increases while a thermal bubble builds then at a certain temperature it breaks off, causes wind, cools and repeats at regular intervals. And often makes a cloud. So yes you could even forecast when a cloud will appear above you, which stuns the people who dont understand how cumulus clouds are created. Fun stuff.
Cheers Ian

1 Like

Not if those values are outside of the margin of error of the device. If Tempest is saying that the device’s precision is .1 despite it reporting .01, I think it’s a fair assessment that any precision in the hundredths place should be considered inaccurate. This would make even perceived trends as being invalid as well.

I don’t question if there is value to you to having that level of precision, but I think what is being expressed to you is that the device simply isn’t capable of providing accuracy at that level of precision.
If the device shows you temperatures of 30.14, 30.13, and 30.12, but due to the margin of error the values were actually, 30.13, 30.14, and 30.11, does that do you any good?

Yes, if the relative precision is accurate. The operative word here is relative. Just like the pressure sensor is accurate to the hundredth (in inHg) it is precise to the thousandths. This is how a trend can be detected before a hundredth of an inHg changes value.

amazing that one can predict clouds or changes in wind based on a relative increase in temperature of way less than 0.1 degree. I didn’t know that.

Another interesting reason for temperature to atleast 0.1 degrees is when back casting electricity consumption in Western Sydney a difference of 0.1 degrees C is significant. While some people will not believe it the forecaster in the company I worked for explaned to me that using 0.1 degrees was very important to his calculations. But that is not relevent to this topic but just another example where greater precision can be valuable information.
What is relevent to this topic that no one has mentioned is that the Tempest temperature value is not only derived from the temperature sensor. The solar radiation and wind speed warm and cool the temperature sensor due to minimum shielding which is compensated by an invisible calculation which throws a spanner into the works. The Tempest shielding is not designed for precise temperatre readings. It is likely more accurate during the night. Which may still provide something useful to some one. Perhaps a farmer wanting to know if the night may frost and they want to see the trend to determine if to activate their water sprays?
Cheers Ian :slight_smile:

well…from an earlier comment of mine:

“My suspicion is that it goes in steps of 0.1, but as the tempest has some algorithm to compensate for the sun heating the instrument or wind cooling it, they report more digits, so the conversion to fahrenheit can be rounded correctly (as it doesn’t make sense to round the intermediate results). So you can get results like 4.24 but that would be due to change of sun/wind conditions,”

but then it was a quite night with no wind and the value still varied by 0.01 degree.

1 Like

Former control system and systems engineer here.

Anybody making business decisions based on these consumer grade devices being +/- hundredths of a degree is simply a bit delusional.

You cannot make imprecise hardware better than it natively is by tweaking the values with kludged compensation algorithms based on ‘multiple’ possibly even more imprecise sensors.

To me, the best accuracy of the Tempest temperature sensor would be in a completely 100% always shaded location with no wind flow basically…in other words a perfectly sited Stevenson Screen at night.

And even then, you’re limited by the chip they picked which is ‘not’ precise to 0.01 deg kind of numbers.

It is what it is in terms of precision (how close two readings are to each other) and accuracy (how close a reading is to reality).

3 Likes