Should I expect a crash at some point?

Nervously I took the Spy 750 Y6 up for the first time in over 3 weeks (had a warm spell of 25F today).

In stab it flies great, then switching to poshold it seemed fine, but did the twitchy thing again for a few seconds here and there, then seemed to smooth out (subjective). I brought it down and hovered nearby where it stayed rock steady for a minute or so and landed.

So, what does it mean when INAVerr goes to 255 and stays there? From what I understand it is the FC not receiving data from the GPS properly. Apparently this copter has done it from day one since transplanting the PIxhawk and running on AC3.2 and 3.2.1, with the 3DR GPS and the current CSGShop Neo-M8N. These errors did not exist on the Y6B with 3.1.5 and the 3DR GPS.

Any help is appreciated.

Views: 2523

Reply to This

Replies to This Discussion

Another thing, but probably not related, is when uploading log files using MP, it doesn't seem to do the complete file every time. I tried it 3 times and each time the log was slightly different in size and the data was garbled/incomplete when opening it. That's the reason for the picture above not being a graph. Taking it off the sd card directly in the PC it's fine.

Hmm.

There appears to be gaps in the GPS data. Logs from last year did not exhibit this (see bottom picture).

2/21/2015

August 2014 Mission run:

Is your satellite count good?  I had one copter where the satcount would go to 0 for half a sec or so.  The copter would fly away.  I got rid of that gps to cure the problem, and have started using Neo M8N.  Now I get 16-20 sats and constant reception.

If your connections are good, not routed near interference, twisted.....and I would guess it is not likely the wires, I would suspect GPS reception.  There are definately not-so-good units out there.

They may be other ways to check this, which I dont know right off.  There is a quality term for the gps.  I have not used it.  I simply lost sats.

Saturday the lowest Nsat count was 17. The highest hdop was 1.5.

If you look at the picture, notice there is a pattern of missing data. The Time in GPS is linear, no interruptions, which indicates to me there is no loss of data due to bad wiring but I'll redo the wiring/connectors for good measure. The unit is a CSGShop large size (don't have the model no. handy).

I went over the parameters in UCenter to make sure they were set per instructions at RCG for this GPS. However, according to a newly released GPS M8N at RCTimer here

NOTE: When connected to either APM 2.xx or Pixhawk, the firmware recognizes the ublox GPS 
and sends instruction to set the baudrate to 38,400, rate to 5Hz, turn off NMEA protocol, turn on UBX protocol, as well as other settings suited for use with APM and Pixhawk meaning that the "Default parameters" are irrelevant as they are overridden by the APM or Pixhawk firmware.

 

So now I don't know what to think, but put in a request to CSGShop for help.

I recently had a problem where a quad would not complete all waypoints.  It would start drifting off, and seemed to drop altitude.  It would always do it on the same turn.

It reminds me of your twitchy description.

I cured it.  I switched from the internal compass in the hawk, to the external compass on the M8N.  I think I was getting motor current coupling into the onboard compass.  Oh yea, my alt hold was also dropping out, I raised my throt velocity P and I, and the the alt hold P term.

That quad has a camera and is prob a little heavier.  The default PIDs on alt hold seemed to work, but turned out to be a bit low.  I raised them quite a bit, 50% or more.

For what it is worth.

Oh, also, when I install a M8N GPS, I just changed the baud to 38.4, as I have been told the software will configure the gps, and that seems to work.  So I doubt it is setup, thought the strobing loss of data is odd.  There is not a gps update rate parameter in the ardupilot config is there?  Could be too fast or too slow if so.

The Pixhawk is set for external compass. All settings in the M8N have been reset and changed to the suggested settings, but according to the link above, the M8N automatically changes the settings based on Pixhawk's firmware.

It doesn't lose altitude, just gets an occasional twitch. When it happened 3 weeks ago I thought I was going to lose control of the copter, but was able to land. Someone mentioned the K solar index was high that day, which I usually check before flying but didn't so can't say that contributed.

It holds altitude fine, at least did yesterday. So good in fact I let it hover for a good minute < 5 ft off the ground. The problem is consistency. I wonder if EKF is preventing it from doing really crazy things. I'd just like to get at the root problem causing the hundreds of lines of "APM Speed" errors, which show up as INAVerr @250.

I received a reply from CSGShop. They think it is a Pixhawk issue. Oh the fun never ends.

On the positive side, I ordered an Arsov AUAV X2 FC to test in place of the PIxhawk. If it still does it even after that and replacing the wiring, then fingers start pointing to the GPS. I don't know how else to track down the problem and as of yet I've found no threads on people having this problem and if/what they did to fix it.

I found I have the same problem! Running a Pixhawk with 3.2 and a CSGShip Neo-M8N GPS. In all my logs, INAVErr climbs to 255, sometimes very quickly, sometimes in steps. I checked for correlation with any other parameters but haven't been able to find any. Note that I haven't seen any in-flight "twitches" though.

How fast does it climb to 255 for you? In most of my logs it seems to reach 255 in a few minutes but I found one log where it only climbed to ~60 in in a time span of several minutes. It could be that my problem is not as severe as yours (and I don't notice any twiches as a result).

I definitely have the feeling it's not the Pixhawk HW but somehow related to the M8N GPS. It's well possible there are many out there who simply do not realize that their PM is reporting INAVErr errors...

Wow! I'm not alone. I'd have to look at the logs again, but it climbs to 255+ quite fast. The twitching may not be related, and not knowing what INAVerr really is about makes it worse. I think there is an issue in GPS modes because the copter flies fine in stab mode without simple/super simple mode.

CSGShop did not even attempt to explain what may be the problem, only place the blame on PIxhawk. I linked to DIY, but apparently they are not interested in helping.

I still have not uploaded the data and load into Excel, but the gaps are plainly visible. If you create a kmz file for Google Earth, look at the pattern of data gaps while in poshold.

I'm not ready to toss the CSGShop yet; will put the 3DR GPS back on and test, but I thought it was happening with it on as well; not positive to be honest. In any case, based on what I've observed and Randy's caution that it may be a real problem, my confidence level is about 5% right now to go back up. This copter has not been crashed yet, and I'd like to keep it that way especially due to mechanical failure.

In about a month I'll have another copter put together with a new FC/GPS to see how it goes, but it won't be a GPS from CSGShop this time. 

I use the same GPS, but have not associated any twitchy stabilization with GPS:/

I just looked for inaverr, and could not find it in either log.

Could you advise me on how to locate it, and I will let you know the results.

I have numerous Ardupilots flying, some with M8N, some not.

It's under PM (Power Management)

Hi DG,

I too see "missing" datapoints in the KMZ file. See attached picture. They do not seem to be at the regular intervals you see them though, which could explain why your INAVErr seems to climb to 255 (the max) faster than mine (?).

Hans.

OK, well I looked in the .log I have available from my crash, and INAVErr was straightline 0.0.  That may have been, Im pretty sure it was before I upgraded to a M8N.

Browsing more current logs, it seems to get errors most all the time.  Sometimes it goes up to 255, ramping up in a curving, increasing count.  Less often, it will ramp up to a number, like 7 and stop.

My guess is all M8N ...wait mine is from DroTek.....is that the same as CSGShip/op?  I thought I remembered that name.....????

Anyway, 2 of mine are doing that.  I have 2 more I have not checked, and one uninstalled.

?!  Do you know anything more about what this term means?

Reply to Discussion

RSS

Groups

Season Two of the Trust Time Trial (T3) Contest 
A list of all T3 contests is here. The current round, the Vertical Horizontal one, is here

© 2019   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service