[f5soh]

When calibration succeeded, AuxMagSettings/mag_bias was very high ~10000 and mag_transform values were very low ~0.05

Check the environment, this is the symptom of a magnetic field / magnetic material around the sensors.

[rsk]

Will do. Thks.

[Jhinta]

so got my self a ext-mac , every thing is oke, but when usb is not connected its nearly inpossible to get my esc armed.
any id's?
on mainbuild firmware.

ahhh  fast multi fast-reinsert makes it start.

[kennyevo]

Hi!

My new m8n gps arrived with a mag in it, tried to calibrate as seen in the video (but the whole thing is assembled), the onboard mag works with a 5% error, the aux works with 15%.
Is the "go to last step and rotate your fc in every direction" a good option, or should I recalibrate everything with the instructions on the calibration screen?

Thanks!

[Mateusz]

Is the "go to last step and rotate your fc in every direction" a good option, or should I recalibrate everything with the instructions on the calibration screen?

Mag takes samples from the first step, so if you do something wrong on the way it will affect your calibration. Going to last step and rotating in all possible directions to cover whole sphere around is best I think. Also flight controller has internal mag and by default GCS assume the auxmag is in the same orientation. OP GPS has it in the same orientation, but many M8N units form ebay have mag chip in all different directions depending on model. You need to rotate Mag in GCS settings accordingly such that it matches with internal Mag.

[kennyevo]

Yep I think it's upside down, I'll try to recalibrate the mags on a field, hopefully it will be good

[[email protected]]

so got my self a ext-mac , every thing is oke, but when usb is not connected its nearly inpossible to get my esc armed.
any id's?
on mainbuild firmware.

ahhh  fast multi fast-reinsert makes it start.

Try to set a motor tab correct hz of you esc rspid esc is 400hz.

[rsk]

I checked my setup and this is what I got:

I manually reset mag_bias to all zeroes and mag_transform to ones
I displayed the raw magnetometer data on one of the GCS scopes and found the following:
- data on corresponding axis of on_board and aux mags are perfectly correlated (range -200 to +500)
- except that at a certain point there is a huge spike (+10000 to +20000) on all axis of the aux mag
- and afterwards the data from the aux mag flatlines and GCS stops updating the corresponding data objects
- sporadically (no fixed period) additional spikes can be observed
- only way to reset the aux mag to working condition is by powering off the whole jig.

Is there a way to determine if this is a problem with the I2C interface/driver on the Nano or a faulty RCTimer GPS+mag implementation?
 
Thanks for your help.

[rsk]

I forgot to mention that I also noticed that the refresh interval on the GCS metadata for the auxmagsensor went from the 1000 I manually set to 300 ms by no action of mine.

[Mateusz]

You could try adding pull up resistor between SDL and vcc and scl/vcc not sure value one should check i2c specs. Maybe 10k? Or shorten your wires and keep SDL away from SCL.

[rsk]

Assuming standard mode I2C and a Cb of 400pF, Vcc of 3.3V then Rmin=1kohm and Rmax=3kohm. (attached paper detailing formulas)

I already have the shortest possible SCL and SDA wires. Separating them would be tough though .

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

Thanks

[f5soh]

Revo have already pullup resistors.

[rsk]

Does the Nano also have the pull-ups?

[f5soh]

Revo Nano has also pullup resistors like Revo but without ESD protection.

[rsk]

What do you think then is happening with my setup?