I’m running a waypoint flight (yes Andy, waypoint flight, but I don’t think that was the cause). The craft is tracking beautifully. A thing of beauty. At 223 samples (seconds), OSD telemetry tells me BL Missing which I translate as kopterglish for motor failed. But the craft is still tracking very well. Since the program takes me over a very narrow “crashable” path through our majestic but craft hungry Pacific Northwest Doug Firs, I decide to let waypoints bring me home on 7 motors. And that’s going well. Another 15 seconds and I would have taken control and landed. At 364 samples she drops like a stone. The log shows FC I2C. The dreaded I2C collapse methinks. It was a dreary cool winter Pacific Northwest day with the usual 80% relative humidity. Is moisture a likely cause? I’m going to spray the boards, but if anyone sees anything else screaming for attention in the logs I appreciate knowing. The craft has perhaps 3 hours on her, so component failure is more likely environmental/infant mortality than wear. The motors all run fine, and sound fine on motor test post-crash. The damage is at least a boom, some clamps and a tilt servo. Luckily it landed in a monster PNW blackberry patch. Thanks. Mike. Hungry trees:
The cause is magnetic trees, Mike. Go near those damn things and you're doomed! Seriously, yes, if you have not sprayed the Power Distribution Board, that could have been the cause -- the props stir up the air sufficiently that you will see moisture buildup on booms and motors -- and I would be deeply suspicious of the Power Distribution Board and the MOSFETs on it. The real test is to let everything dry out (hair dryer assisted is OK) and then carefully power up the system and see whether all the BL-Ctrls come up and whether you get any bus errors -- you may see some on power up, but the number should not increase. I can see a can of conformal acrylic spray in your future....and some masking tape! Andy.
Since she's broke I can't actually put any load on the motors. But without props, mid speed, they all run and sound fine. No errors. Masking tape it is. Mike.
Hi Mike, can you please post what exact combo/craft setup you built? I will check the log later, I'm on iPad mini now. The moisture doesn't seem to be the cause, there were many more flights in much harder conditions (dripping, flying under waterfalls..) than you describe without a single problem. I also assume you didn't let any possible condensation to build up inside your craft during transport and immediate start with a quick warmup of craft components.
It’s a CS8MK Octo XL distribution boardMK Flight Ctrl 2.5Mk Navi Ctrl 2.0MkGPSQuadroPower Motor QC-3328Xbee 3DR2 Axis + RadiansGraupner MX-20FatShark FPV
Mike: Sorry for the delay in getting to this, but, using MK_GPXTOOL, I agree with what you're seeing -- not that the log file is sampling once every 500 milliseconds, so row 223 is actually 111.5 seconds into the flight. 1. At 111.5 seconds into the flight, motor #4's current draw goes from 8.1 Amps to 13.4 Amps (the motor currents are shown with Motor #1 first through to Motor #8, the current is shown in units of 0.1 Amps, so divide by 10 for Amps): The last good motor current data in row 222 is: MotorCurrent: 70,69,76,81,74,77,49,34,0,0,0,0The first bad motor current data in row 223 is: MotorCurrent: 70,69,93,134,74,77,49,34,0,0,0,0 2. At the same time, the FC board reports the loss of a Brushless Controller. 3. Note: Motor #7 and #8 seem to be reporting currents of 4.9 and 3.4 Amps which both seem disproportionately low -- but they've been doing that since the start of the flight, so this may not be causal). 4. At 184 seconds, the I2C bus starts reporting errors, at which point things go nuts and the copter starts to fall out of the sky. Sadly, the I2C bus errors make the flight log data unreliable as the flight log is recorded by the Navi Control board based on data packets received via the I2C bus. From the video you posted on Facebook, you can actually see one motor has stopped (in the top right hand corner of the frame at 0:36). I'm not sure which boom this is, but it looks like it might be #7 -- hard to interpret this because I cannot sync the flight log to the video without seeing the whole video. If I understand what Jan is saying it is that other people have flown other copters in more extreme conditions and they did not fail. I'm not sure I can make the inferential leap required that says, "and therefore moisture could not be the cause of your failure." The problems is that the successful flights were not your copter and were in different circumstances. Maybe those successful flights were just lucky that the moisture didn't cause a BL-Ctrl shutdown -- clearly *something* caused a shutdown in your copter. If it wasn't moisture on the MOSFETs, what was it? If you happen to have the copter back together to the point where you can test to see if (with the same power distribution board and brushless controllers) you can spin up all the motors (without props, of course), the it will be interesting to see if all of the BL-Ctrls come up and spin up their respective motors -- if they do, then it indeed was moisture (that has now long since evaporated). Thanks Andy.
Was the copter still doing its waypoint thing when it went down? If so, there’s a lot of electronics in the mix, and it’s quite clear that the I2C bus issues were its undoing. I’ve been flying MK for a while now, but I’ve never used waypoints. Is there a way to trigger an emergency Come Home when in waypoint mode? If you were still in waypoint mode and the I2C bus flamed out (or even just got very twitchy...same thing I guess), the copter would have no command and control, and that would explain its crash, right? I have experienced the loss of a BL in semi-automatic flight (hovering in PH/AH) 100 yards off shore over the ocean, and my response was to ditch all the automation and carefully fly her home. I was fortunately (Lady Luck!) able to do so, and (as you saw) the Cinestar 8 was able to fly on 7 motors almost magically. However, if your copter ran into issues where its Navi board was unable to reliably issue commands to the Flight Control and you did not have access to manual control of her, that would lead to bad things. I’m curious what the procedure would be to escape from waypoint flight and take manual control of the sticks. Is that a straightforward thing? If you had it to do over, I would immediately abort the flight and try to set her down as soon as a BL reported in with a serious error. All of which begs the question of what happened in the first place, of course. Was it moisture? Quite possibly. The BLs generate a lot of heat, as do the motors. But carbon fiber doesn’t conduct heat, so it’ll remain cool. I’ve noticed several times (much to my dismay) that after flying in what looks like not so foggy conditions (but obviously high humidity) that my copter will land with (literally) moisture dripping off the CF booms. Easily enough to potentially cause issues. This is why I’m investigating covering my center hub with something. Especially because my new Double Quadro 2XL power distro board and BL 3.0s runs very, very cool with its integrated cooling systems (I have the one with the massive heat sink between the boards). I have no interest (at least now) in flying in actual rain, but I live in a very wet area (fog, etc), and this would give me piece of mind.
Steve, it was flying waypoints. Not only was it flying waypoints, but it was doing so on 7 motors (if you believe the bl missing telemetry). I chose not to disable waypoints as 1) I didn’t know what to expect on manual on 7 motors, 2) she was on the way back to safety, and 3) she was flying beautifully. It is easy to remotely disable waypoints (GPS off), and to fall back to nothing, just AH, and any all GPS modes including CH. I’ve practiced in anticipation of needing just this. Of course that’s not what I did. I think your approach would have been the better one. Why am I using waypoints? Well, I’ve got some planned fireworks shots that will be really hard without. I’m trying to get mileage. ( It’s only due to one of those strange quirks on the road of life that I didn’t end up working with my dad micro brewing beer in Half Moon Bay, 25 years ago. ) Andy, thanks for the debug on sample interval. That explains my pulling of the hair trying to get video to sync with telemetry in MKGPX tool. My current hypothesis is that the electronics are possessed, and that the demon can travel between boards over the bus. I pulled out flight and power, taped and sprayed them. I repaired the gimbal and broken booms, balanced and sensor calibrated everything. I then anchored her, checked motor addressing and rotation, and ran the motors with props up to about 1/3rd power with motortest, for about 8 minutes. I observed no errors. When I tried to fly, she would oscillate violently immediately post lift and then flip. What tiny logs I got didn't tell me much. This seemed like it should be a sensor / calibration thing, or a balance / alignment thing. Or some damn thing. But while I was trying to debug I noticed that about 1 out of 10 times on motortest, a motor would not start. And not always the same motor. I have as second CS8 (ARF) in on order. When it arrives I'm going to spray and swap in flight and power and see if she flies. Mike.
Before you ditch the FC, try recalibrating the ACC, then recal the compass, and try again. If still no joy, then, as Steve says, it might be an FC failure. Remember that the I2C bus runs between all of the boards and all of the BL-Ctrls -- if there is a bad connection anywhere on the bus, all hell breaks loose as the boards cannot talk to each other. On your short test flights, did you see I2C bus errors? Andy.