Chris,
Thanks for the link! And yeah, i havent had a ton if time to devote this week to the unit, so i have just been tinkering per se’. Definatly am getting this isnt something you can just whip through, browsing the full parameter list my eyes glazed over a couple times. Probably will watch some video, read a little more and wipe the unit to defaults and start from scratch again.
One thing that keeps popping up alot in MP while I’m configuring is “bad gyro health”. A reboot doesent fix the issue either. 
Also saw EK blow up on the bench once, but after a reboot it never popped up again with the message.
Maybe I just need to reload the firmware, re calibrate and setup everything again?
Tim
I know Pixhawk 2.1 is going to be different. Evidently the bad gyro health thing has something to do with the heated IMU’s. So it might be better to ask about that on the hardware section of the forum to see what to do about it.
EKF3 in 3.5 is also more complicated. I think it’s disabled by default and EKF2 is used. Having the EKF blow up on the bench is not a biggie. It doesn’t mean much. It just means the two EKF state estmators didn’t agree, and if you have this gyro problem that’s going to happen. I would not worry about EKF at this point until you get the helicopter ready to fly, everything is calibrated, you take it outside and let it get a GPS lock, arm it and spool it up and see what you get for vibes. The EKF (IMU’s) is quite sensitive to vibration. And you won’t know what you got until you run it the first time.
I always prefer to take the blades off mine and spool 'em up with no blades to make sure everything works. And then I do kind of a strange thing - I slip a 8 foot 2 x 6 thru the landing gear and put cement blocks on it. Then I start the engine and warm it up, engage the clutch and spool up the head and run it at full collective, both positive and negative, to test the governor or throttle curve. Test out the cyclic, rudder, etc under power… Then I shut it down and pull the SD card out and look everything over in the logs. If it all looks ok, then I fly it the first time.
Yes, if you’ve been “seat of the pants” playing I guess I would go to mission planner or APM Planner2 in the full parameter list and reload the defaults and start over. You don’t have to reload the firmware. The playing around got you more familiar with the system, which is good. But then follow all the mandatory hardware setup steps including radio cal, accel and compass cals, etc. Then do the heli setup, making sure your rate PID’s and fastforward is set properly so you got enough swash movement to be able to control the heli when you first fly it. It’s going to take a bit to tune it. Repeated takeoffs and landings and adjusting settings in the rate controller. Once you get that fly it a bit. Test it out. If everything seems ok, try alt hold once. If that works try loiter and see how well it “locks in”.
If it doesn’t work quite right pull the logs out it and post them and ask for help.
Just checking, have you set AHRS_ORIENTATION to 0?
Pixhawk 2.1 will comes with 180 which is meant for inverted mounting.
PittRBM,
I set the orientation to 270 i believe as its mounted sideways on the frame of my heli. The horizon looks correct and level in MP.
Tim
Set RC_SPEED to the maximum Hz of your slowest servo and adjust ATC_ACCEL_YAW_MAX. If you are used to FBL-units the default limit of ATC_ACCEL_YAW_MAX is way too low for your style of flying.
I’ve run my 333’s at 333, 270, 200 and 125 and couldn’t tell any difference with ArduPIlot. Well, one difference - they run hotter and deplete the servo battery quicker at 333.
Tim, what do you mean about “teaching the FBL about the geometry” exactly?
There are cases where the Arducopter system is missing features because I never got around to it, no time, or didn’t feel it was necessary for a UAV helicopter.
There are other cases where I did things differently, because I understand how helis actually work better than the hobby helicopter companies.
For example, you can set the servo updates to the max rate supported by your servos. I typically do, just because. But it’s really not a big deal. Nothing on the heli happens at 125Hz, let alone 333Hz. This is all just a bunch of garbage put out by the RC heli industry to sell the new product. The commonly stated “it’s so locked in!” thing is not an engineering term. You’ll notice that most of the rate controls are passing through a 20Hz lowpass filter. In fact, I often set mine to 10Hz. And they’re all doing similar. If you tried to actually run with a 333hz lowpass filter, the servo outputs would be completely saturated by noise. So when passing the gyro signals through a 20Hz filter, the difference in performance due to reduced latency between 125 and 333hz, is negligible.
That being said, most servos, very strangely, run their internal PID position loop and output servo power once per update they get. As such, when they get updates at 333hz instead of 125hz, they actually result in more power and speed. This is what can actually make a difference in the flight performance. But not the decreased latency due to faster signalling. Makes no sense to me why they set it up this way. You should be able to send signals at 50hz if you want, and it would just run the same PID loop 6 times with the same input before getting a new update.
I think they just did it so that if users test 125 vs 333hz updates, they get a change in feel that they want to feel to justify their purchase of “faster” servos.
There’s a lot of snake oil in this industry.
Similarly, I believe that the real reason Futaba invented the narrow-band signaling, 760uS servos with 560Hz updated gyros, was simply to lock users into their ecosystem. Want a Futaba gyro, which are actually a really good quality product? You’re going to have to buy a Futaba servo too…
That being said, you can use these servos with Arducopter. And I do. Not because of reduced latency. But because these servos typically have much higher transit speeds than the low-frequency servos. Doesn’t have to be that way. It’s just what they’ve done.
Futaba servos also are less susceptible to electromagnetic interference. I found this out the hard way after buying a slightly cheaper servo to save money. Then come to find out they were affected by the 3DR telemetry radio.
BTW, I have a heli pretty similar to yours:
I don’t use it much however. The Pixhawk installation isn’t great and it suffer vibration problems. Amazingly for such a big machine, I couldn’t find enough room to mount the Pixhawk properly as I normally do. I could fix it, but just never got around to it.
Not surprised to hear that. Futaba stuff is great, and I am moving towards it almost exclusively. It’s just, it’s great for reasons other than what they are using to market it.
I’ve also had problems with noise from telemetry screwing up cheaper servos.
I’ve had problems with the telemetry radio screwing up a Futaba 9254 too. It’s just common mode on the servo wires from EM radiation from the radio. Simply putting a small ferrite torroid on the servo wires about 2-3" from the servo, or twisting the servo wires, gets rid of it.
Rob,
Firstly, thanks for the information I’m always up for it.
As far as teaching geometry, in all the FBL’s ive set up anyways, there is a step where you show the fbl where 6° is on your particular heli so it gets an idea of how the helicopters geometry is performing. As i understand it, it helps adjust the units performance to your specific helicopter, like servo link lengths, servo arm distances and the different ratios each helicopter has. Take spirit for example, they give a quantifiable number in the setup as a range you need to be in for the FBL to do its job properly. The number on my 700x for example is much different than say my Trex 700V2 as they have very different servo linkage setups. The 700x is direct from the servo arm to the swash and the Trex has the push pull setup with different ratios.
I am likely not explaining this correctly, so please bear with me, but i do know for a fact through testing that if you set that step wrongly, the FBL becomes an ossicilating mess.
I could be way off base, im only regurgitating information i was given, but it made some sence and ive seen what happens if it is out of range, or set to a different degree value than what is called out.
My usual procedure for setting up a heli goes like this…
Set servo arms at 90° to the servo case.
Set the swash levels and at 90° to the mainshaft at mid stick.
Set 0° on the blades at mid stick
Go into FBL setup, trim servos to get perfect 90° on arms and swash.
Teach the FBL geometry to the FBL by setting 6° on blades paralell to the tail boom.
Set cyclic ring and collective range
There is more but its not pertinent to this, like tail setup and gains. Just trying to paint a picture of where that step comes into play. Its been a nessecary step on every FBL ive personally setup across multiple brands.
Tim
Nice heli, i remember that Trex. They also had that Align APS that went to the 3GX if i remember correctly? I dont think i ever read of a succesful use case of that system.
I mounted the Pixhawk on the side of the frame. Hopefully that goes well?
I got the heli running very smoothly already as it was being used with a different controller and gimbal. A very large underslung gimbal to carry a 5D MK3 and an Atomos Ninja 2 SSD recorder.
It was a chore tuning out the slow ossicilations with all that weight hanging under the heli, but i got it after awhile. Hoping i dont have to revisit that with the Pixhawk since the gimbal is balanced already.
Tim
What sort of effect are you having with the servos? With the telemetry radio on im not seeing anything yet, the servos just sit there silently until i move the sticks. Ive never used the BK servos until this, but i will say they seem quiet and smooth. Although im used to Spektrum, not quiet in the least.
Tim
If you’re geting interference from the telemety radio the servos will be “grinding away” on the bench with the heli in Stabilize and not moving. In extreme cases, where you may have run the telemetry cable right alongside servo wires, it may even make them twitch.
If the heli is on the bench and you hear lots of servo noise, simply pull the plug on the telemetry radio and see if goes away. If it does, fix it before you fly it. It’s common mode current induced in the servo wires by electromagnetic radiation from the 915 or 433 MHz radios, which is often times a harmonic of the servo frame rate, and the cheaper servos not having adequate internal filtering of the signal. Killing common mode current is pretty easy. Ferrite beads or torroids will kill it. Using twisted pair will sometimes kill it. In extreme cases you have to relocate the telemetry radio.
Ah, yes. I think this makes sense.
I think what they are doing is pre-scaling the PID outputs to the geometry. This makes sense. Would be a nice feature to add, as it would make it easier for users to get starting PID values.
Disadvantage would be that everybody would have to retune their PID values because of the new setup. Tridge tends to not like doing that to people. I’m more pragmatic if it results in a better system. I think this would be worthwhile doing.
Yep. I’ve had servos actually moving with every transmit burst. I did everything to stop it, but no success.
Actually, I think that was back in the day of Xbees radios. The 3DRadios were much better.
Some of the cheap Chinese radios are no better than the Xbees ones. I’ve looked at them on the scope and they put out spurs and I don’t even know how they’re able to sell them in the US because they don’t meet the FCC spec on emissions for non-licensed radios. I gave up on those long ago because I consider them all to be junk. I’m an Extra Class ham radio operator, so I can run ham band stuff. I built my own telemetry radios running on the 33cm band. My radios have nice tight emissions, way better noise filtering, and 5x the range of the off-the-shelf consumer stuff at 25dBm output.
They can’t be sold in the US legally. They just do it anyway.
The Chinese factories have learned that if they put “FCC” “EC” “ROHS” “UL” labels on things they sell. So they do.
I used to work for a wire and cable manufacturer. We’d see wire coming into the country with our UL and CSA listing on it. They don’t care. They just do whatever it takes to sell product.
If I’m not mistaken, this is a feature in the setup software for the FBL unit, and not in the firmware of the unit itself.