F7 pressure laboratory compare just for fun

EDIT: Added TEMP test too.

Today I tried to compare F7 standard barometer with calibrated proffesional pressure monitor. I also compared the watch of colleagues - Amazfit GTR4 and older Honor watch GS Pro.

Measured "right" value was 960.8hPa

1st. place for Amazfit GTR4 with 961hPa (doesn´t show decimals)

2nd. place Honor watch GS Pro with 960hPa (doesn´t show decimals)

3rd. place F7 with 958.9hPa

When I owned F7Pro solar for two weeks, it shows 1hPa higher than my F7. So it was closer to the reference. In the end it doesn´t matter, because the trend is more important than real value, but if there was a possibility to add offset in some menu, it could be able to show perfectly accurate value :-).

I´m going to do some next temperature measurent tests in temperature chamber. Will post later :-).

  

  • Hm, I just tried to recompile the app after adding the plotting of the diff rawAmbient - ambient pressure, but there were too many changes in the SDK in the time I did not touch the sources, so I am getting too many code incompatibilities, that will require some time to fix. Unfortunately, right now I cannot allocate the time to it.

    And looking at my old code, it is possible I gave you false information. Probably the experimental version I used did not really show the rawAmbientPressure. Although originally I had it there, and also kept the label "raw pressure" for the graph, I might have replaced the function with the plain pressure for some testing, and did not put it back to the original state. The difference of the ~2 hPa might have come from the small elevation of some 8m I am currently at. Sorry for the disinfo. The true data may be similar anyway, but at this moment I cannot verify it quickly.

  • That's okay, don't bother with that.

  • OK, finally I took the time to fix the compiler incompatibilities and changed the graph to plot the difference rawAmbientPressure - ambientPressure. It gives the following result. The difference between the two (apart of the fact that the ambientPressure is smoothed, unlike the rawAmbientPressure) is not that big - just a few tenth of pascal in average. The temperature went from 31°C (on the wrist the first 5mins) to 28°C at the end (on the desk). 

    The same graph with the temperature overlay:

    Frankly told, it does not really look like there is any significant difference in the temperature compensation between the two values. To me it looks more like the ambient pressure only filters out the noise of the raw sensor data, while keeping the internal sensor's temperature compensation. So just a confusing description in the API documentation, in fact. Maybe I'll try another test later, heating up the watch more, and cooling it artificially with ice.

  • It seems that the ambient temperature does not affect the raw pressure value?

  • It seems that the ambient temperature does not affect the raw pressure value?

    Well, the rawAmbientPressure function is temperature compensated, according the API documentation, so that's not surprising. The problem is with the description of the (non-raw) ambientPressure function, which is written in such way that you can interpret it, there is no temperature compensation. This test shows indeed rather the opposite - both of them seem to be temperature compensated.

  • Just a very low I think. I let my watch in the laboratory cool down from about 28°C to 19°C and the pressure (ambient pressure datafield) has changed by 0.2 to 0.3hPa.  

  • So they are all temperature compensated but the values are not same. That's waired. BTW, I asked Garmin China about this again and they didn't reply me yet. I don't think they have an answer to that either.

  • So they are all temperature compensated but the values are not same.

    No, they are the same - the noise of the raw data is in the range of units of pascal, it means hundreths of hectopascal (0.01 hPa or millibar), which is completely neglectable.

  • ... and I think it also clears a bit more the absence of the temperature compensation at the altimeter. Garmin's claim about the non compensated altimeter likely does not mean the ambient pressure is not temperature-compensated, but rather that the conversion to the elevation does not take the air temperature in account. For the accurate conversion of the ambient pressure, they would have to know the air temperature at several levels of altitude, since the air pressure at given altitude is depending on the temperature profile of the entire column of air above the location.