[Mateusz]

I am going to look at SCL and SDA with a logic analyzer and see what I find. Will report back.

I doubt logic analyzer can show the quality of signal, how the peaks look like, you would need oscilloscope.
Interesting reading http://dsscircuits.com/articles/effects-of-varying-i2c-pull-up-resistors not sure if that is any help though...

[rsk]

You are right.

[nuggetz]

I thought that external mag wasn't supported? Yet I see tutorials on external mag.

[Mateusz]

I thought that external mag wasn't supported? Yet I see tutorials on external mag.

External Mag is supported with OP GPS in released Librepilot. Other external Mags are supported also but with development version of Librepilot, not available in binaries yet, you would need to compile code yourself

[hwh]

..External Mag is supported with OP GPS in released Librepilot. ...

The supported "OP GPS" he's referring to is NOT the one on sale on eBay and the Chinese sites that's called OP GPS.  That one is a tiny gps without a mag.  The supported one is the "OpenPilot GPS Platinum" AKA the "V9 GPS" and was 36mm square.   It's no longer made.

[nuggetz]

Does anyone have a dev version compiled that I can use to run my external M8N Ublox/Mag? I really want to give this a spin.

[hwh]

The project doesn't distribute compiled pre-release versions of the software.

The instructions for downloading and compiling the "next" branch are in the developer section of the wiki
https://librepilot.atlassian.net/wiki/display/LPDOC/Developer+Manual

[Jhinta]

Does anyone have a problem if i host them? Bins that is. And yes totaly free of ads or payments.

Verstuurd vanaf mijn LG-D855 met Tapatalk

[rsk]

I connected an oscilloscope to SCL and true enough, the signal is completely distorted.

I have attached the scope captures. Relevant data:

- actual frequency is 364kHz. Should be 400kHz (fast mode)
- Voltage rises only to 2.6V instead of 3.3V
- It takes the signal 1150ns to rise to 2.6V. Is should be less than 300ns

According to the tables previously provided, Rmin = 1kohm and Rmax = 1kohm

Will add the pullup resistors and report.

[rsk]

I added 1.2kohm pullups and there was definite improvement. See attached scope captures.

- signal rise time was reduced to 115ns which is well within specs
- speed is now about 410kHz

On the first try, I had a successful magnetometer calibration and bias and transform values seem reasonable.  Auxmag health indicator is mostly green. This is a first.

I have not lost data telemetry from the auxmag since the modification was done.

Will do more tests and report. Initial results seem promising.

[hwh]

Glad to hear it's working for you. 

The 10k pull ups on the revo are a bit weak for use with anything other than very short board to board cables.

Out of curiosity, what oscilloscope are you using?  One of the usb ones?   I've been thinking about replacing my big tektronix scope with something a little more portable.

-Hank
 

[rsk]

I am using a Hantek 6022 usb oscilloscope/logic analyzer. It is really convenient to use and very portable and does its job. There are a number of different models that will allow for sample rates according to your needs.

[nuggetz]

The project doesn't distribute compiled pre-release versions of the software.

The instructions for downloading and compiling the "next" branch are in the developer section of the wiki
https://librepilot.atlassian.net/wiki/display/LPDOC/Developer+Manual

I followed those instructions and did a git checkout next and then built the package and ran the installer and GCS crashes on startup.

[f5soh]

Any error ?

[nuggetz]

GCS just crashes. Where would I look for the error? Was hoping there was a log somewhere.