Here is a quote by the guy who developed the program for BF and RTH with GPS.

I was unaware that not having a barometer would affect the functioning of the GPS in such a negative way.  I have 2 quads with GPS without Barometers and 2 with Barometers.  I will be more cautious in the future because of this warning with the 2 quads without barometers.

"Here's the issue. If you have no barometer and less than 10 satellites, when a satellite goes out of view or a new one comes in your altitude will jump ridiculously. The quad could be at 100m one second and at -5 the next. Barometer will mitigate this significantly, because the sensor fusion algorithm I wrote takes GPS precision into account and pays much less attention to GPS altitude when this happens. If you have no barometer, you have to go with GPS altitude because you have no other signal. This is to say: it's very possible that if rescue mode engages with no barometer and 8 satellites, your quad may be coming back and disarm unexpectedly because it thinks it has hit the ground. Don't say I didn't warn you."

Wow! That's not good (but very good to know!)

For people who run quads with both GPS and a barometer, take a note of SJChannel's recent findings with regards to a a bug in versions of Betaflight 3.x and 4.0.x. The bug will supposedly be fixed in Betaflight 4.1.

https://intofpv.com/t-barometer-and-magn...8#pid68778

LOL, I think I can maybe conform this issue. Thinking back on the failures I had when originally testing RTH, the quad indeed disarmed when it got close to its launch point.

https://youtu.be/6CxDYDV9jcQ?t=140

Below 10 sats...

(13-Sep-2019, 11:25 PM)kaitylynn Wrote: LOL, I think I can maybe conform this issue.  Thinking back on the failures I had when originally testing RTH, the quad indeed disarmed when it got close to its launch point.

https://youtu.be/6CxDYDV9jcQ?t=140

Below 10 sats...

I thought of you, Kaity when I read the comment in my OP.  

Funny, my 10 inch has a barometer and GPS and have never seen any anomolies.  I need to see if the Baro is on in BF.  My new 7 inch has a Baro.  I'm going to look at recent DVR and see if the GPS is wonky.  I'm on BF 3.5.7....

(13-Sep-2019, 11:25 PM)kaitylynn Wrote: LOL, I think I can maybe conform this issue.  Thinking back on the failures I had when originally testing RTH, the quad indeed disarmed when it got close to its launch point.

https://youtu.be/6CxDYDV9jcQ?t=140

Below 10 sats...

And that was almost a year ago?  

The BF guys are just now pointing this out?

So I could only find the very first flight of the Super G+ in DVR.  I don't have any old DVR from my other quad with a barometer. 

(Disregard the issue of the battery voltage not changing during the flight.  That has been fixed.)  And yes it is a boring 6 minute flight.  It was the first shake down run.

You can see altitude in the bottom left of the OSD.  It starts at about zero feet at arming, and finishes off at about 9 ft AGL upon landing in the same spot.  So there is an anomaly there I didn't see when the flight was made.   During the flight the altitude readings seem fairly accurate based on previous flights, and changed appropriately with altitude changes.  

So does the error get worse as altitude increases, so if you fly to 400 ft vs my 60 feet would the error grow worse?

Is the fix mentioned in the other post, does it work, or should one wait until BF makes the change?  And will the BF 4.1 change work on BF 3.x.x without messing up the older version of BF somewhere else?

I might just turn off Barometer and not worry about it if I can't be confident of the changes.

From a conversation on Github concerning this problem...

Comment by Mikeller  "@SJChannel: Not sure if idling a quad on the ground is a good approximation of the pressure / vibration conditions you get in a real flight - it would probably be better to test the barometer in a real flight, and check the altitude reading on the ground before taking off, and then at the end of the flight after landing - in my experience a shielded BMP280 can easily yield 10m altitude drift in a 5 minute flight - as opposed to a GPS which may drift around the correct value, but will never actually permanently drift off."

Clearly it would seem a bad idea to blend the two inputs though....

A always found the barometer to be pretty inaccurate even an a quad without GPS (which mine doesn't have). The barometer resets itself to zero upon arming but if you let it sit idling on the ground, you can see the value start to creep up and up on its own, so it can't be anything to do with air movement over the barometer. Looking back through my DVR footage, the barometer reading in my OSD always shows a height of over 10 metres when I'm landing back on terra firma, and looking at one of the DVR recordings from this morning's flying session (snapshot below), the barometer was reading over 20m when I was landing...

I've seen the same inconsistencies in small personal aircraft with barometer drift.

I'm thinking a Barometer may be more of a liability than a asset for a quad at this point.  

I believe the Barometer can be turned off in BF, right?

I think you can turn it off, but now I'm wondering if there is some kind of calibration procedure that could be done to help it maintain accuracy. Does INAV have the same issue?

so what's the current census, use the baro sensor if equipped or don't?

(14-Sep-2019, 03:15 PM)SnowLeopardFPV Wrote: A always found the barometer to be pretty inaccurate even an a quad without GPS (which mine doesn't have). The barometer resets itself to zero upon arming but if you let it sit idling on the ground, you can see the value start to creep up and up on its own, so it can't be anything to do with air movement over the barometer. Looking back through my DVR footage, the barometer reading in my OSD always shows a height of over 10 metres when I'm landing back on terra firma, and looking at one of the DVR recordings from this morning's flying session (snapshot below), the barometer was reading over 20m when I was landing...

hello all, i realize this may be an old thread which caught my attention as it relates to a present situation i may encounter on an upcoming build.

but for starters, if we look at how a BMP280 barometer pressure sensor works, it will measure pressure altitude relative to ambient air pressure it picks up. this will display pressure altitude above mean sea level (MSL) as its reference point much like a real aircraft’s altimeter. so the 21.4m you see above may be your actual field elevation above MSL. 

if the quad resets to zero upon arming then it provides the pilot a more accurate altitude reference relative to the ground upon takeoff which is less confusing to understand. case in point: if you flew from a location that is 1200ft above MSL and claimed to 1500ft indicated altitude via OSD, you are only 300ft above the ground and should land indicating 1200ft, not zero if quad will not reset to zero upon arming.

another thing to keep in mind operating a pressure sensor barometer is the need to adjust for actual atmospheric pressure changes which is always varying relative to present weather conditions. i shall not dive any deeper into this boring read but in a nutshell when the pressure altitude significantly drops or rises without this adjustment feature you get an even more inaccurate altitude readout. all aircraft adjust this value with reference to present conditions before any flight.

now for my use case scenario. krohsis mentioned the risk of relying on gps altitude in the absence of an FC based barometer. but by the same token it was on the premise of having less than 10 satellites. in a similar thread i made i am planning to use a GEPRC TAKER stack for LR to rely on gps altitude. but using M10 chips it will easily lock in excess of 15 sats during cold starts and upwards of 20 sats hot start. hopefully this may mitigate the altitude swings once this thing gets airborne as it picks up more sats.

hopefully someone chimes in on this… cheers!