Bosch BME280 Temperature Compensation for Pressure Measurements

[KVO2]

Hello,

Goal:
I am trying to get very precise pressure readings from the Bosch BME280.

Problem:
I've found there is parabolic looking relationship between temperature and pressure which results in unacceptable skew of pressure measurements due to changes in temperature.

Solutions tried:
1 - Bosch provides code that is supposed to handle the compensation but it is not working for me.

2 - I've also tried measuring the relationship myself and creating my own compensation. This gives fairly good compensation but not good enough and it takes more processing than I'd like. I'm considering putting more time into getting a more accurate measurement of the pressure vs temperature relationship but it is seems wasteful to keep putting time into something that Bosch supposedly already did for me.

Questions:
Any ideas? Has anyone encountered this?

Thank you for any help.

 

[itsmoked]

You should cobble together your test results and confront a Botch field engineer. It's one thing for a user to complain about accuracy and an entirely different thing when the user complains AND shows their data about accuracy. Often that gets very good focus from the manufacturer which can help you get to the bottom of the issue.

Doing the aforementioned has twice gotten me the manufacturer's attention followed by admissions that there were bad production batches that resulted in arm-loads of new replacement chips delivered by hand.

The only other things I can think of for you to stir into the pot are:
1) The data sheet seems to paint the device as something to build into smartphones for non-quantitative reasons, more for 'awareness' type applications not analytical apps. The sheet goes on about idiotic "context" sensing, a bad sign for an analytical application which yours seems to be.

2) Have you verified that your system that is reading the device is not causing the non-linearity? Perhaps the voltage is varying or you have noise on your supply?

Keith Cress
kcress -

http://www.flaminsystems.com

 

[KVO2]

Thank you for the quick reply. It's good to hear that sending results to the manufacturer has worked for you, I am currently waiting for a reply from them for the results I sent.

2) The results are fairly consistent across a few units through a few different methods of temperature variation so I don't think noise is an issue but thank you for the suggestions, I will double check the voltage at various temperatures just to be sure.

 

[KVO2]

Confirmed; temperature does not affect voltage supply.

 

[itsmoked]

You've drawn a line thru some common issues. Excellent.

Depending on where you are I would try to get a Field Application Engineer to your location. It's much better if you can get someone who's a representative of Bosch in front of you, rather than email or letters or even the phone.

Being in Silicon Valley I may be over optimistic in getting on site help.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[KVO2]

Yes, that would be nice. I'm only in the 'silicon vineyard' so field visits aren't too common but it does happen occasionally.

 

[KVO2]

Does anyone have an email or phone number (preferably Canadian or USA) for someone at bosch that replies?