Pressure measurement in Pascals

Issue:Atmospheric pressure is reported over LoRaWAN with not enough significant figures for our needs. The reported data is two bytes yielding for example: “barometer”: 99970. However, in the serial port monitor the device measures as follows: Pressure:999.75 hPa
Any ideas how to send the pressure measurement over LoRaWAN with greater resolution, i.e. down to single Pascals?
Setup:

Server:

Details:

There is no configuration command to my knowledge and, as you may be able to see on this forum, the code base isn’t available for modification.

Do you need lots of these sensors?

Why do they need that extra digit of accuracy - 0.01 hPa is 1Pa and given a human can blow a trumpet at 30kPa aka 300hPa, you are asking for accuracy in free air that would change as someone wafted past - perhaps you need something more custom?

Thanks for your comments, Nick. We are attempting to measure differential air pressure between two rooms, where the delta should be maintained greater than 10 Pa. We hope to find a way to monitor the differential pressure with sufficient accuracy. To filter sampling noise and small variations in pressure, we thought to average measurements in a short period, and using two sensors.

It seems the code base is published on the RAK Wireless developer site, but it is not immediately clear what changes to make to send greater resolution, e.g. more than two bytes over LoRaWAN, or to modify the code if it is rounding the measured value.

Doh, I was thinking about IAQ debate from a few weeks back, too many products that are just numbers, yes, it’s in the Products_practice_based_on_RUI - you need to look at the lines around 153 in app_7204.c.

The code base is using the Cayenne LPP 2.0 format which specifies just one decimal place. Rather than have a hybrid non-standard payload, I’d alter the code to send as you require - no point transferring other data if you don’t need it.

Is this a one off install? And is it safety critical for environmental containment?

Thanks Nick. I had located that part in the code but it helped you saw the same thing. I changed the code to subtract 65535 so that the smaller number would fit in two hex bytes. On the netwokr decoder I added this offset back and now the reading is in Pascals down to single digits.
Just trying to find out if through sample averaging using two sensors we can measure pressure difference between two data centre rooms where the delta should be at least 10 Pascals. If this doesn’t work, which it probably won’t, we will use a single differential pressure sensor.
Cheers.