Thanks Spektrum!

[Silverburn]

Witness the handiwork of another quality implementation...
View attachment 16659

Squirting out doing a demonstration of FPV technology to a local SAR rep, and this happened. 500m out, still had rock solid 5.8 video, and the TBS did a flip for no reason. Got control, but it wasn't to last - control was on and off, then fully off. APM completely freaked, and it plummeted to earth from a great height. Failsafe didn't kick in because the radio was still transmitting, but appeared to be transmitting garbage. I've done this route a dozen times without incident...but today, with the SAR operator here to witness "how good MR's are", it decided to freak.

100m up and 500m out...thankfully video was still transmitting, my spotter and I were eventually able to find it again. APM logs show some bizzare "in" signals, consistent with a spektrum freakout.

Impact was significant - damage: 3 motors, 3 arms, GPS, battery, board cam, Gopro, bottom plate must have been massively warped on impact - the RF shield over the "Core" was warped to buggery, and the battery has a 4' bend in it.

My back-ordered Taranis can't get here soon enough.

 

[R_Lefebvre]

Sorry to see that. Can I see those logs? I'd like to see what happened.

I don't know why anybody uses Spektrum anymore. :cold:

 

[Carapau]

I am no defender of Spektrum per se but it is always easy to blame the kit when it could potentially be caused by something else entirely. For instance, is there a loose cable on the aircraft that is arcing a bit- arcing causes horrific interference; if there are any radio towers near by, then this could happen too. Did you fly between a microwave link (you may well not have seen the towers with the antennae on them)- these things can easily produce wild interference in the command link. Of course it could well be Spektrum. I suppose my real point is that after a crash like this we all get pretty mad and it is easy to blame the easiest thing. I have often found that it is worth taking a step back, clearing 'the red mist of rage' and going through everything methodically and with an open mind and then when you reach a conclusion there is a far greater chance that the conclusion is right.

 

[R_Lefebvre]

That's the nice thing about Arducopter. You can go back, look at the logs, and see exactly what happened. So I'm just curious to see what sort of signals we can see.

 

[Silverburn]

That's exactly what I was planning to do right after the crash, but that's proving more frustrating than expected - the USB extension was torn off in the crash (along with just about about everything else), and appears to have damaged the USB port of the APM in the process. It works, if you hold it *just right*, but it never seems to last long enough to establish a stable connection. I've been trying to do the download over mavlink (it should be a small file - I was only 3-4 minutes into the flight), but it keeps crashing.

(I had only been relying on the "tuning" display until I got the quad home). I usually watch all 4 flight Channels (or the spotter did, in this case).

This is a route I do a lot, and the reason I chose it for the demo - I've done this a dozen times with this craft, it's a "clean" route, and I gave the TBS a preflight check the night before. Even when I collected it in the field,all the usual suspects (ESC, props, bullets etc) were all still tight. I've been pretty scathing of some aspects of the APM in the past, but I've never had reason to doubt it's reliability.

 

[Benjamin Kenobi]

Ah, gotta love the Spektrum bashers.

I've met many of the commercial operators in the UK and a large majority of them use Spektrum stuff, without incident I might add. I've never had an issue of any kind with their stuff for the last 6 years of flying.

What makes you think it was the transmitter when it could be so many other factors?

 

[Silverburn]

Because I think I can eliminate the other factors:

- motors - all less than a month old, and never crashed until today. Passed all preflight checks. Although now full of mud, there's still no play.
- frame and arms - all less than 6 months old, and running normally. Core was interference free, even at full power, so no bad electrics. All screws were tight, pre and post-crash.
- ESC's - all motors cut out at once. If one had failed the symptoms would be different.
- Props and adapters - all were still attached even after crash, and there is zero play. Anywhere.
- APM was still transmitting telemetry during the crash, so that had not failed or browned out. In fact, the tuning window was showing strange intermittent troughs on at least 2 CH in's moments before the crash, according to my spotter. GPS was 10 sats and HDOP <2m at time of crash.
- Video and telemetry was still running rock solid, so was not a power failure, and battery was fresh. Had only consumed 1600ma of 4500 pack, according to core.
- all motor bullets, servo leads, power leads etc were all connected pre and post crash
- route is "clean" - no strange transmitters on 2.4, and flown numerous times in the past.

Happy to entertain any other suspect causes - I would rather this was a simply mechanical issue, and *not* a spektrum issue. All my kit is spektrum, and more worrying is that I have a "professional" hex I use for SAR activities, and don't fancy another one of these events while using that vehicle - especially if people are relying on me.

I'm not a Spektrum hater normally - in fact for close work (<400m) it's been error free. But this is now my 2nd medium range control glitch, on two different craft, in two different locations (only common is the shared spektrum radio) and I have to ask serious questions about how suitable it is for the longer range, more mission-critical activities.

 

[Benjamin Kenobi]

When you title a thread something like 'Thanks Spektrum' it sounds like you've made up your mind based on frustration instead of 'help me diagnose this problem'. Your post crash summary uses broad strokes to come to a conclusion which could mask the issue.

If this was a DJI thread a majority would blame the flight controller. Why do you rule out the FC?

 

[Silverburn]

I rule it out because:

1. At the time of the crash, it was transmitting telemetry perfectly, both to MP, and to the minumOSD, and continued to do so after the crash. In fact the telemetry was used in the locating of the quad post-crash.

2. I have spend a further two hours testing. I used the same spektrum transmitter on a fully operational hex on the bench (props off), and noticed glitching on CH3 and CH7, both audibly, and on MP tuning screen. This is consistent with the two CH anomolies my spotter noticed, although he could not tell which channels those were at the time. A glitch on CH3 would explain why the motors just stopped as well. I removed the motor connectors from the APM, and connected them directly to AR8000. Still glitching. I removed the AR8000, and connected a servo tester to each motor in turn. NO PROBLEM. I used a 4 way servo lead, and tested four motors at once. NO PROBLEMS. I took the servo tester out, and swapped in a different AR8000 and an orangerx. The same two anomolies on CH3 and CH7 (strangely though...the orangerx had far less glitching, but it was still there). Switch off the DX8...no CH3 or CH7 anomolies (proves the interference is not coming from the airframe electrics).

I'm still having problems with the USB port on the crashed APM, but frankly, I have enough evidence now to point to an issue with the DX8 transmitter without needing to see the APM logs.

My initial post is based on experience, detailed knowledge of my vehicles, the symptoms of the crash and having a pretty good idea of what component caused it. I'm not dumping on Spektrum without good reason...

 

[Benjamin Kenobi]

Thank you for taking the time to post this. Very detailed analysis.

 

[JoeBob]

Silverburn, I'm newly flying an AR8000 with a DX9, so your post has my attention.

In all your testing, did you happen to use the same jumpers? Those little wires take a lot of abuse with my fat fingers tugging on them, especially since you have to trim the fin off of DJI jumpers to get them to fit into the Spektrum slots.

 

[Bartman]

I rule it out because:

1. At the time of the crash, it was transmitting telemetry perfectly, both to MP, and to the minumOSD, and continued to do so after the crash. In fact the telemetry was used in the locating of the quad post-crash.

2. I have spend a further two hours testing. I used the same spektrum transmitter on a fully operational hex on the bench (props off), and noticed glitching on CH3 and CH7, both audibly, and on MP tuning screen. This is consistent with the two CH anomolies my spotter noticed, although he could not tell which channels those were at the time. A glitch on CH3 would explain why the motors just stopped as well. I removed the motor connectors from the APM, and connected them directly to AR8000. Still glitching. I removed the AR8000, and connected a servo tester to each motor in turn. NO PROBLEM. I used a 4 way servo lead, and tested four motors at once. NO PROBLEMS. I took the servo tester out, and swapped in a different AR8000 and an orangerx. The same two anomolies on CH3 and CH7 (strangely though...the orangerx had far less glitching, but it was still there). Switch off the DX8...no CH3 or CH7 anomolies (proves the interference is not coming from the airframe electrics).

I'm still having problems with the USB port on the crashed APM, but frankly, I have enough evidence now to point to an issue with the DX8 transmitter without needing to see the APM logs.

My initial post is based on experience, detailed knowledge of my vehicles, the symptoms of the crash and having a pretty good idea of what component caused it. I'm not dumping on Spektrum without good reason...

fair enough, you've convinced me you're not just jumping to conclusions. what's the history of the radio equipment? crashes? was the receiver hard-mounted to the airframe or isolated a bit by foam or velcro? have you investigated the prongs of the servo ports to verify they're well attached to the underlying PCB?