I’d like to see if people would like to see the temperature displayed as such: 43.27 F
vs. 43.2 F Since the pressure provides more detail than most with an extra digit for measuring the smallest changes in pressure - I’d like to see the temperature displayed in hundreths or at least be given the choice to carry out that far.
Gary
How many decimal places does the actual underlying UDP API display ?
what use is it do display those numbers if the device isn’t that accurate in the first place. If memory serves me correctly it is 0.1 degree celcius.
yes, presently to a tenth of a degree but I’d like to get greater resolution to .00
What @sunny is saying is that maybe the sensor is not capable of providing that level of precision. The only reason the pressure reading makes sense at its current level of precision is to make relative pressure comparisons even though you need to only use the hundredth place (when using inHg) when reading a single isolated pressure value.
The precision of the measured temperature is tenths, not hundredths, of a degree so only one decimal place is warranted.
you could ask for 0.00001 degree, which is just as useless. you’re asking a device to tell you information it isn’t rated for. It doesn’t have any way to tell you reliably the difference between 15.00 and 15.03 degree celsius.
You don’t have to be mean, why does the current pressure reading go beyond traditional retail pressure instruments?
I see value in one more level of accuracy, if you don’t like it, move on please.
So why does Tempest offer three figures after the decimal point when it too is not needed?
I completely agree and understand if it can’t provide that level
of detail. Commercial grade thermometers frequently offer this feature (museums for example)
Appreciate your courtesy response without being belligerent.
Gary
FWIW, the UDP API (WeatherFlow Tempest UDP Reference - v171) seems to indicate 0.01 degC from the example data there, but that’s just me reading between the lines from their sample data.
I can’t find an official ‘specifications’ page online that answers the underlying question of what you can reasonably expect.
A @WFsupport post pointing to the actual documentation of the Tempest specs would be needed for you to know what you can reasonably expect in terms of precision…
I can’t answer that question but .001 inHg seems too precise to be useful. To put that precision in perspective, the decrease in atmospheric pressure is about 0.001 inHg per foot increase in elevation in a standard atmosphere near sea level. So for a Tempest, at say, 100 feet above sea level if the user estimated the elevation as 95 or105 feet there would be a 0.005 inHg error in barometric pressure. And that assumes differences of .001 inHg could be accurately measured by the Tempest.
I’m sorry if you feel like that. I was just trying to explain with a more absurd example why it doesn’t make sense. But apparently that didn’t work for you.
With a better sensor you might measure with the precision you like. But you are right I see no value in measuring atmospheric temperatures that precise. What do you think is the use case of such precise measurements if it were possible?
If anyone has ever breadboarded the Bosch BME-280 pressure sensor and looked at the output, they would see the measurements bouncing wildly due to drafts in the room making the extra precision essentially useless. Hence, mathematical rounding to a more useful precision for a home/prosumer weather station…
These type of precise temperature features are in place
already for applications such as wine cellars, museums, curation storage units
and security areas. So, I’m not trying to invent something that already doesn’t exist.
Obviously there are needs for this precision as industries already use them. I was
simply wondering if Tempest had the capability, why not offer the option to go out that
precise. Over and out on this.
Gary
I use the hundredth of change w pressure to gauge the minor pressure
changes and rate of decline in half hour increments for approaching storms and warm fronts.
Gary
You guys seem to want to argue. Sigh. How about you just figure it out experimentally and lets move on here.
Run my python listener (link) and see what number of decimal points they actually report.
You can’t do better than what the hardware emits and you two guys arguing back and forth is not going to change the answer the hardware tells you.
Disagreement isn’t always arguing. I’m done
here, was just a thought. Over and out.
Gary
It reports 2 decimal places
Here is what the Tempest reports for obs_st:
{'serial_number': 'ST-00061881', 'type': 'obs_st', 'hub_sn': 'HB-00068027', 'obs': [[1675372154, 1.95, 2.86, 4.15, 246, 3, 990.62, **4.23**, 45.25, 24871, 0.62, 207, 0.0, 0, 0, 0, 2.805, 1]], 'firmware_revision': 165}
Epoch: 1675372154
Wind Lull: 1.95
Wind Avg: 2.86
Wind Gust: 4.15
Wind Dir: 246
Wind Int: 3
Pressure: 990.62
Air Temp: **4.23**
Rel Humid: 45.25
Lux: 24871
UV: 0.62
Solar Rad: 207
Rain Acc: 0.0
Precip Typ: 0
Avg Stk Dist: 0
Strike Cnt: 0
Battery: 2.805
Report Int: 1