sorry Randy I should have mentioned the Zubax GNSS V2
Yes it’s Chibios…It’s the future lol.
Sure I can produce a longer log no problem at all. What ever I can do to help.
As for the sat count, ya that struck me as odd, on a normal non UAV CAN GPS I get 20 wiht no issues. 4 was in my mind odd.
I wonder if there is a way to update at 5hz. I can ask Pavel if thats possible…
1 Like
This quad flew perfectly with Nutx and 3.5.7
Loaded Chibios and Arducopter 3.6
I had to disable the logging in standard parameters otherwise I couldn’t arm.
Loaded Chibios and Arducopter 3.6
I had to disable the logging in standard parameters otherwise I couldn’t arm.
Reformated the sd card but to no avail. (was working fine with nutx and arducopter 3.5.7)
Is FAT32 ok?
Any help welcome.
Thank you in advance
Henri
Hi Rob, understood. I had planned to switch back to EKF 2 anyway. The innovations that are zero are on XKF8, the issue is that as I begin to fly more agressively the mag interference increases, EKF starts switching and that’s when the problems occur. I get EKF failsafes and other similar behaviour which makes it quite hard to convince myself that all is ok. All I’d really like to understand is whether I can do anything about this. Physically the external mag is quite close to the battery lead, so at low revs things are ok but at high revs I’m guessing the internal compass has less interference, but if that’s the case then the XKF8 innovations are a problem?
The brute force method of stopping the compass and/or EKF switching is to only run one of each. This does remove some redundancy but in case the redundancy is actually making things worse then it could be the right choice.
To disable all but the external compass set COMPASS_USE2 = 0, COMPASS_USE3 = 0.
To only run a single EKF set the EK2_IMU_MASK = 1, EK3_IMU_MASK = 1 (instructions are also here)
Hmph. Someone else has reported that Class4 cards worked fine on NuttX but not under ChibiOS, but moving to a Class10 card fixed things for them.
Odd thing is that I use all sorts of crappy cards in my vehicles here and do not see this 
There are actually patches on master that change the way we decide there are problems with the storage subsystem. If these reports continue we’re likely to want to pull them into the stable Copter release, unfortunately.
777b4c8cfe09ae05795385a1f027b8d57184d959 is one of the patches.
And, sadly, 8b4b38bcb2c1b1aaad8b7563cdeff624c63d0f16 is another. tridge killed everybody’s favourite “no io thread heartbeat” message…
1 Like
@rmackay9 ok I will experiment. At least I would be able to see whether I do actually have a problem or not.
Related question - I have seen the EKF failsafe trigger the failsafe action even when I am in stabilize. Is this right? Surely GPS/mag are irrelevant in stabilize?
The EKF failsafe shouldn’t trigger in Stabilize or Acro mode I think. This came up during 3.6.0 beta testing and we added this line specifically to be sure. The only way it should be triggered while in Stabilize is if the FS_EKF_ACTION parameter = 2. If you’ve got a log showing that it’s not doing what it should, could you post it and we can revisit the issue?
@rmackay9 ok I will check my logs. I did have it set to 2, why 2 only - what about 4 & 5 (which are not documented)?
@andyp1per, OK great, that sounds like it explains it then. I’m not aware of any extra options 4 or 5 for the FS_EKF_ACTION parameter… I think they would be defined here if they existed…
Sorry, getting confused with the battery FS options and smart RTL!
1 Like
Ok Randy.
Here are some logs. I hope I got you more time as requested.
Today I got no HDOP better then 100%
Also got bad AHRS which I wonder is tied to the GPS issue. Also got GPS Signal Health error again…surprise.
Anyway hope this helps.
Also note the strange file names. Odd.
3 2018-10-31 1-07-56 PM.bin (31.4 KB)
3 1969-12-31 7-00-00 PM.bin (256 KB)
2018-10-31 13-07-10.bin (391.7 KB)
6 1969-12-31 7-00-00 PM.bin (268 KB)
@peterbarker Thank you for answering.
I didn’t know you were referring to the sd card! (Thought you were talking about the PixXXX)
Where do I find version 777b4c8cfe09ae05795385a1f027b8d57184d959 patch?
Thanks
Henri
I do confirm this with mine, I thought it was suddenly defective.
@peterbarker Thank you for answering.
Excuse my ignorance, I don’t know what are class 4 or 10.
They’re different classifications of SD card; it’s written on the packet
when you buy it, often on the card itself.
Where do I find version 777b4c8cfe09ae05795385a1f027b8d57184d959 patch?
“git show 777b4c8cfe09ae05795385a1f027b8d57184d959” in a master ArduPilot
checkout.
Henri
Peter
Visit Topic or reply to this email to respond.
In Reply To[45.png]
peterbarker
November 1 Hmph. Someone else has reported that Class4 cards worked fine on NuttX but not under ChibiOS, but moving to a Class10 card fixed things for
them. Odd thing is that I use all sorts of crappy cards in my vehicles here and do not see thisThere are actually patches on master that
change t…
Visit Topic or reply to this email to respond.
You are receiving this because you enabled mailing list mode.
To unsubscribe from these emails, click here.
Peter Barker | Programmer,Sysadmin,Geek.
pbarker@barker.dropbear.id.au | You need a bigger hammer.
:: It’s a hack! Expect underscores! - Nigel Williams
I can confirm that on four different PixXXX same behaviour is showing with four different micro SD cards:
- Arducopter 3.6 on NuttX: No problem
- Arducopter 3.6 on Chibios: Bad loggin and “No IO thread heart beat”
Perhaps something isn’t right in the ChiBios drivers?
Cheers
Henri
Thanks for the logs, unfortunately they seem really short as well… I think in the longest log the board was powered on for perhaps only 27 seconds? At least a few minutes would be good.
The names on the files are because the GPS is sometimes not getting any satellites so it’s not getting a time. We set the flight controller’s time from the GPS… and that time is also used in the naming of the files.
Thanks for that bit of info.
Also sorry for wasting your time.
I think I get whats up with the files so here I try again Sorry Randy, I appreciate you
I put a whack of them in a Dropbox folder. Hopefully its reachable for you.
This appears to be 4 minutes in length
2 1979-12-31 7-00-00 PM.bin
Also I cleared the folder on the pc and wiped the copter. If the logs I posted don’t work I will try again but this time I know it will be fresh logs from that test.
1 Like
Great, thanks for the log. So the good news is that in this log at least, the vehicle is not suffering from the performance issue we have found with another user of the Zubax GNSS V2.
From what I can see, there’s no problem in this log really… except that the GPS has absolutely no lock but that’s probably because it’s inside … and this is not unusual depending upon the building it’s in.
So I’d say take it outside, give it a couple of minutes to get a good GPS lock and then try and arm it.
That was outside.
I had it on my deck, never saw any sats’
Do you see any other issues.
I get the Bad AHRS and VertPos…so I am thinking they are GPS related since to some degree it’s heading based.
@rickyg32,
Ah, ok… so there’s definitely something wrong with the GPS then. I’m not sure whether it’s configuration or what but it’s sick.