Tim, I have custom radios. I’m an Extra class ham radio operator and I built my own telemetry radios. The radios I have are not legal unless you hold an amateur radio license.
Chris,
Ummm, yeah about the ham radio cert… 
I admittedly need to buck up and make it to the class to be 100% legal with the gear i may or may not use… I believe it only the level 1 technician test i have to take? Isnt that not for all vhf uhf above 30mhz? That would cover me for my 5.8 video downlinks and radio telemetry and Dragon link stuff etc right?
Isnt it that anything not fcc certified or above a certain output wattage or something like that that requires the ham certification?
But I digress, thats awesome youve made your own telemetry radios! I really like electronic and mechanical projects like that, i wish i had the time to do more complete DIY stuff. Ive been looking at a smaller 3D CNC machine for quite awhile, but at the moment it would sit and collect dust as what free time I do have is spent doing projects like this Trex 800 with Pixhawk, building 3D helis or flying. Theres just not enough time in the day to do what I want, and someone awhile back equated Pixhawk to “the rabbitt hole” and until now I dident have any idea what that meant. Pixhawk is now soaking up an immense portion of my available time, and I’m just now scratching the surface i feel.
As to the radios, I see alot of people using the rfd900’s so I guess thats my logical choice?
What baud rate do the radios youve built operate at? 57600? 115200? Im less than impressed with the speed at which the 3dr radios i have connect and download parameters and data in Mission Planner. Is that just normal behavior? Am i expecting too much? Would higher end radios even work any faster and send more real time data?
Tim
Okay, so i was just reading about the “air data rate”. Basically it says you can increase the data rate at the sacrifice of range? Chris, have you adjusted this setting? Is that a viable way to get better performance out of telemetry?
Or am i just off base here? The only telemetry im farmiliar with is not adjustable at all, mostly just in-built radio telemetry in common transmitter systems, Futaba, Spektrum, FrSky etc. And any quad racer telemetry i use is broadcast over the fpv feed in minum osd.
Tim
Yes, just a tech license will get you legal with anything in UHF, and down to 10 meters in VHF. Although there’s only a very narrow frequency spectrum you can use in the 10m band as a Tech. 20dBm is the legal output for Part 15 radios in the US.
I use 57.6K baud. The bandwidth determines the data rate though. And more bandwidth you use, the more power it takes to push a signal. It is not practical to try to get real-time telemetry. A couple seconds latency is fine. And power is not the answer. You can double the output power at the final amplifier stage of the transmitter and you only increase signal strength by 3dB. The antenna is where it’s at. It’s way more effective to build an antenna that has 10-12 dBi gain than it is to crank up transmitter power.
I have no experience with the RFD900 radios.
The amount of time it takes to download all the parameters has increased with every release. Back in the APM days we had like, what, 300 params? Now there’s around 700?
Back to your first auto flight. Notice how when I switch from Loiter into Auto how the helicopter snaps itself around and points at the first waypoint, then accelerates to the target speed while climbing to hit that first waypoint. That’s how it’s supposed to act. If the helicopter seems indecisive, flies sideways, doesn’t stay on the ground track, misses waypoints, or doesn’t hold the proper altitude, you got a problem. May as well abort and bring it back and fix it before it flies off someplace you don’t want it to go.
It’s good to make a flight plan with several turns in it like you see in the video. It’s a good way to make sure the yaw works right, and the helicopter banks into the turns instead of tail turning and skidding around the turns. If you see it doing that kind of stuff, then you have a problem with your tuning.
Chris,
Thanks for the detailed description of proper “Auto” behavior. As an indicator for how well it may or may not fly an auto mission, can you look at the logs for desired pitch and roll vs actual pitch and roll? Is it a sound assumption that if those track somewhat similar it should have a good auto flight?
I was looking at the logs last night and the tail now looks spot on, for desired vs actual, it seems to follow quite nicely. Pitch and roll also look better, but are no-where near as good as the tail it seems? A lot of room for improvement I guess.
Tim
Correct. The helicopter won’t bank into the turns properly, nor will it accelerate and stop properly if those are off. Last log I seen, your roll didn’t look too bad. Pitch was a little hairy for forward flight. And pitch is where you’ll lose control if you start running the speed up. So just be careful with the speeds in auto if you intend to run that low of headspeed.
With a helicopter, Tim, headspeed is your “safety cushion” just like altitude is with a fixed wing. The first thing the autopilot will do in the event of power loss is go to full collective pitch instead of feathering the rotor and getting some forward cyclic into it for autorotation. Running excessively low headspeed, you won’t even have time to react before it stops the rotor and you lose all lift and tail, and are now flying a rock. Maybe practice a few autos with the headspeed you’re running before trying it in auto flight mode to see how much “cushion” you have. To practice auto’s disable the auto disarm param in the FC and just shut the governor off on Channel 8 in flight in Stabilize flight mode. If you don’t want to auto to a full landing, then flip the governor back on. It will skip the ramp and runup time in flight, and should be really smooth transitioning back to powered flight if you kept your headspeed up properly doing the autorotation.
Chris,
At the headspeed i’m running now which is fairly low, I would have to have some good height to even attempt an auto, at least in my experience with the heli, and I would have to catch it immediately to even hope to get it down without issue once I add in the extra packs and with the gimbal on there, I can only hope to slow it down a bit before impact. its a case where insurance is my best friend… A while back I added dead weight to the heli to simulate what would happen in the event of a power loss and I had to auto, and after many attempts I realized that at the heights i’m usually flying at for videography, there is no saving it, i’m in the dead man zone. It’s a veritable falling concrete block. But without the extra weight with those big 800mm blades, its an auto dream. Incredibly easy to float it down to a nice soft landing, and i’m no expert at autos either
So I decided to focus on a stringent maintenance routine with thorough pre flight checks and redundant power systems wherever I can. When I’m using it for work with the gimbal, I have backup power for the flight controller, the servos and the packs are wired in a manner that if one pops and it doesn’t cook the harness it can still continue on with reduced capacity, hopefully… Since I’m not drawing very many amps I don’t worry too much about the esc/drive, but since I am running at partial load always I monitor temps quite religiously and at some point will likely install a Kontronik on it with active freewheeling which would result in cooler running temps at partial load and likely a bit more reliability… It’s just hard to pony up $900 for a darn esc with those. uggh
Tim
I[quote=“ChrisOlson, post:556, topic:16800”]
OK, well I tried it. I get no cyclic response at all on the bench. The swashplate wiggles a little bit and goes the opposite direction of the cyclic stick, and then re-centers back to level.
[/quote]
I just did a bench test using AC 3.4:6 with the following settings
ATC_ANG_PIT_P=0
ATC_ANG_RLL_P=0
ATC_ACCEL_P_MAX,162000
ATC_ACCEL_R_MAX,162000
ATC_RATE_FF_ENAB=1
I didn’t touch any of my PID gains and had about the same swashplate reaponse. No issues. Could you post your param file?
Correction
ATC_RATE_FF_ENAB=1
Bill, I’ll post it in the testing/development thread later. I’m building some code right now to test the DDFP tail fix, and soon as I get that test build loaded in the flight controller I want to test it yet today.
So… Today was the big day, had about 8 more tuning flights and on the 9th i went for it.
AltHold=Works great
Loiter=Awesome. No toilet bowl, no wobble, no wandering just a rock solid in place hover. I tried moving it around in loiter a bit and it feels more responsive than i would have thought.
Ill upload the video a bit later. Definitely excited, feels like it was easily worth it, cant wait to move onto auto flight coming up after a few more successful loiter and alt hold tests. 
Tim
1 Like
Congrats Tim! Look forward to the vid. At this point, seeing it fly tells more than logs, unless there’s a problem that can be identified by watching it.
Thanks Chris, i wondered if id get it a few times, but at this point my confidence in the Pixhawk 2.1 is getting pretty high! I watched it pretty close the second time i loitered it and it looked like it wasent overshooting any movements or anything, and the wind popped up with some gusts and it just sat there. Its a very surreal thing watching that big helicopter sit there!
Tim
Yes, it certainly is. Helicopters are the hardest of all RC aircraft to fly. There’s been a lot of hard work and dedication from a lot of people to make it possible to fly one on RC with an autopilot.
You are right there, outside of the Pixhawk world i had no idea the commitment people have made for this to happen. Now that im in the thick of it its become apparent how much this all entails, and im only scratching the surface.
At this point i am quite thankful to have access to this equipment as the only other choices for helicopters are quite slim these days. Older DJI industrial controllers, or things like the Naza H…uggh had one of those, not a fun experience.
Definalty thankful for all the hard work that has been put into it!
And definatly appreciate the help for yourself and Bill. Still have a ways to go before i have a viable platform though. One flight at a time i guess.
Tim
1 Like
It’s definitely looking good! It’ll be fun to see that first auto flight. Just use the default speed of 5m/s and see what she does.
Chris,
Yeah, its starting to feel pretty good in flight as well. Ill probably keep trying to close the gap between desired and actual along with tweaking control reaponse here and there so it feels more like im used to.
So the accelerometers are being used in stabilize im assuming to limit bank angles? What settings are available if any to tune the anount of bank angle, or just the general feel of the stabilization? Ive used stabilized modes in the past with other systems. In those cases the accels were kind of on top of the normal gyro stabilization and you could either turn the gain up or down to make it feel more like a coaxial or more like a 3D heli if that makes sence along with another parameter to change how far the accels would let it bank.
As to auto flight, i want to make it to a bigger field and do a couple high speed runs/circuits to check for unwanted behaviors, but lets say it has none and flies nicely, i should assume it would behave the same way in an auto mission correct?
5m/s I guess is quite slow so im not worried about that starting out, but once i start bumping the speed up, rbs/pitch-ups aside, what else am i looking for? Is there a parameter to adjust the gain on how it holds altitude? I think i remember reading somewhere about fast auto flight causing the heli to bounce up and down on the z axis?
Another thought i had was, i noticed in your video of the auto flight with your 500 once you hit the switch to start the auto flight it snapped to the new heading facing the waypoint quite quickly, is there a parameter to adjust the speed at which it does this? Im just thinking ahead for when there is a gimbal attached and how yaw-ing that quickly would affect it?
Tim
That .bin should be of one of the loiter tests. I looked through it a bit and didn’t see any issues, although I am also very green still with all of this. The desired pitch and roll vs actual seems to be getting better and better with each tweak. To me compared to logs from last week it seems to be way better. I think I will keep tweaking the P and VFF gains until I get no more significant changes is the logs anymore if that sounds like a good plan?
I feel I am getting close to adding in the gimbal… I had bumped the headspeed up a bit 15 or so flights ago as I felt it would have been too low to carry the gimbal so i’m hoping I don’t have any significant tuning ahead of me with the addition of said gimbal?
Tim