Hello folks,
I had been flying my 3DR Hexa for about a dozen flights. For the first six flights or so I only flew in Stabilize mode, then when everything seemed okay I went into Loiter mode. I did that for a few flights and it seemed pretty stable. After that I advanced to RTL mode for several flights over the last few days and it would RTL.... Hooray!!
I am going to use my Hexa for Aerial Photography so before I put any gimbals and camera on it I wanted to make sure that it would fly, loiter and RTL reliably. I had been reading quite a few posts about APM crashes and it looked like a lot of the crashes were due to electrical interference from the PDB. The typical recommended solution was to raise the APM/GPS board higher so that it would be farther away from the PDB.
Following a video from Randy, I looked up my Compass Offsets and found them to be +36 for X, +111 for Y and -86 for Z. While those numbers were not too high according to Randy's video, I thought that raising the APM/GPS board could only help lower the interference. The stock standoffs that 3DR used to mount the APM/GPS board are about 25mm high. I got some other nylon standoffs that are about 30mm high so I added them to the ones that are 25mm high, now putting the APM/GPS board a total of 55mm above the top plate of the Hexa. See the attached picture for what it looks like now.
I am also interested in reducing any vibrations on the Hexa so that I can have good picture quality. I have already balanced the motors and props so I enabled "raw" datalogging so that I could see what the magnitude of the vibrations are.
This morning I went out to my flying field to do some flying. I started off in Stabilize mode and it seemed to be okay so I advanced to Loiter mode and it seemed to be okay too. I went back to Stabilize mode and then flew a little farther out and then switched into Loiter Mode again then on to RTL.
In RTL mode, the Hexa rose up to it's 25 meters, turned about 180 degrees back towards me and then started going back to "home". I was surprised that this time it did not come back as straight as usually does, even though there was not any significant wind. It seemed to overshoot "home" by a few meters and the rolls and pitching that it was doing seemed to be more aggressive than before. It then pitched due south and quickly accelerated away from me. It began descending as it was flying away from me. I wanted to see if it would recover so I let it go a little farther that I probably should have because it was at least 50 meters away from me. I finally switched over to “Loiter” mode but it was too late and the Hexa crashed. Luckily it seemed to crash on one arm only and came to rest right-side up but it broke the prop on that arm.
I looked at the log file in MP and saw some strange things. The parts of the flight that were in Stabilize and Loiter mode were correct. The first part of the RTL data looked correct but at 92% of the log file the GPS track coordinates indicate that the Hexa went in a Northern direction at an almost impossible speed. In reality, this is where the Hexa actually went due South, probably trying to correct for the Northern coordinates that it was getting from the GPS. The final track on the MP indicates that the Hexa crashed quite close to “home” but in fact it was about 50 meters South from me.
It looks like the Hexa started getting GPS coordinates that had a significant (about 50 meter) offset and it tried to correct from them but in fact made the Hexa fly off.
What do you think could cause this? The only significant changes that I did was to raise the APM/GPS board as mentioned before. I also had a 2200mAh 3S battery strapped along with the 5000mAh flight battery. The 2200mAh battery was not connected to anything. I was going to use it to power my FPV that I just got working yesterday but I wanted to fly the Hexa first without the FPV to make sure that it would reliably do loiter and RTL modes before I tried the FPV.
Thanks,
Dale
Replies
Hello Folks,
Just got a propeller to replace the one that got broken in the crash on Sunday. I balanced the new prop and put it on the Hexa. I also rechecked all of the props to make sure that they were all tight.
To prepare for the post-crash flight check, I did all of the ground checks, then connected the Flight Battery. I made sure that the APM fired up okay. It got a GPS lock in about 20 seconds, even though I was only going to fly the Hexa in Stabilize mode. Then I spun it up without taking off and every thing looked good. I then got it light on the legs and everything seemed to be good. I then hovered it at various heights up to about 7 meters for about six minutes and everything looked good.
I am going to go out to the flying field (same place that I crashed on Sunday) and fly some more flights in Stabilize mode and then maybe move on to Loiter mode to see if I get any more GPS glitches. There are not any power lines or cell towers in the area of the flying field so I am still wondering what could have caused the GPS glitch. Does anyone have any suggestions on what could have caused the GPS glitch and how to prevent them from happening in the future?
Thanks,
Dale
Hello Folks,
As Randy indicates below the hdop value is too high and the number of sats are too low.
Randy said:
Yes, it's very clearly a GPS glitch. Below is the hdop values which should ideally be around 1.5, you can see they climb to about 2.2. Also the number of sats drops from 13 to 8.
What do I need to do to decrease the hdop values and increase the number of sats value so my Hexa will not crash?
Thanks,
Dale
Does anyone have any ideas on how I can prevent the GPS "glitches" from happening? I thought that I would be helping to prevent magnetomter and GPS issues by raising the APM board by 35mm from it's stock level.
Thanks,
Dale
Hello Rob,
My FPV system was built into my Hexa from 3DR. The sticker on the receiver indicates that it is 1.2GHz. For the crash flight the FPV was not powered up, I had just installed (but not connected) the 2200mAh 3S battery so that I could see how the Hexa responds to the extra weight. The flight battery does light the LED on the telemetry side of the OSD board but I would think that the source of significant interference would be when the video transmitter is powered up.
By the way, I bought the Hexa with the upgraded uBlox GPS. Any fixes or work-arounds would be appreciated.
Thanks,
Dale
Hi Dale
I had some serious GPS glitching when I ran a HobbyKing 900mHz FPV setup that resulted in behavior similar to what you described. I took the FPV gear off and the hexa then flew like it was on rails. Randy warned about this and has a suggestion to roll back some PID setting somewhere (it escapes me at the moment).
Good luck
Rob
Hello Folks,
Thanks for the feedback. Is the "glitch" caused by the GPS reciever, the APM, some sort of interference from the electronics on the Hexa or some environmental condition like terrain (or Burmuda Triangle / Time-space vortex?).
Both of the traces look relatively stable for most of the trace and then seem to become un-stable with increasing frequency at the end. Is it strange that the number of satellites in view would go from 13 and then quickly drop to 8 and then quickly go back up to 13. How many satellites are needed for a good GPS Fix?
What can I do to prevent these "glitches" from happening? The APM/GPS board is already the highest part of the Hexa.
Thanks,
Dale
Yes, it's very clearly a GPS glitch. Below is the hdop values which should ideally be around 1.5, you can see they climb to about 2.2. Also the number of sats drops from 13 to 8.
Dale,
I had a similar incident and every once in a while my telemetry indicates my quad is flying in areas it has not been. I have come to the conclusion the GPS receiver was being blocked so I raised it to the highest point on my quad and I haven't had any issues. I hope this helps? I am still new at this and I am still learning.
Good Luck!
Scott