Nicknamed "Laurel and Hardy", for reasons that should be obvious.
Everything calibrated...time to get the PID's done! Well..once it stops raining and snowing here in Switzerland. :-(
Some build annoyances, which I'll go into in an additional post.
Laurels bits (still missing his TM1000 telemetry though)
Hardy's bits...some repaired!
His motor & prop, with the temp sensor in place.
Comments
Euan - Thank you for the build issues and tips! Very helpful
(Spikes in the vibes are landing "events" where alt hold was failing)
Update: Hardy had his re-maiden and flew better than ever. Until I switched out stabilise. Alt hold and loiter were a total mess; slowing climbing or descending, with no 10% "deadzone" in the middle where altitude is supposed to be held.
Config and vibes chart attached...some are saying it could be a THROTTLE MID issue; but could it be my PID's? IS P enough, and is IMAX too high?
The vibes...
Additional: forgot to mention, I stacked the ESC's against the arms; it's a convenient fit, relatively tidy and it help with sorting wiring issues and matching ESC's to motors after you've put it all together.
OK, build issues (for John's benefit):
I should caveat the following by stating that I'm still a newb to quads, and thus my findings may not be issues for more advanced builders.
I'll focus on "Hardy" which is the larger frame with the carbon arms (it has more issues).
1. No instructions. While quad frames are pretty simple, some simple instructions would be nice. There is one thing that's critical information you need IMO - the bolt tension for the arms. Why? Well...
2. The arms are carbon. For those that don't know, carbon can be very strong, and even flexible. It does not, however, respond well to compression. It is also not the greatest material in the world for grip. This makes it a tricky but good material for quad arms. It is therefore critical to never exceed the compression maximum. This tricky when you have 4 clamps binding to it...You just have to guess! Go too far...you'll hear the crack...and the arm is ruined. You need to know the maximum, and you need a torque wrench to ensure you do not exceed that tension.
Also, the material is not grippy. Do some hard maneuvers and you'll find one or two motor mounts have moved. I recommend getting some low/rough grade sand paper, and running down both the arm and the insides of the mounts. It's a common trick we use in cycling when our seatposts don't stay in position. This should help stop any movement.
3. Fundamentally, this frame design and the "top/bottom" arm clamp system can be very tricky. I recommend assembling it UPSIDE down. This will let you thread though the bolts and the "top" of the clamps. Then you drop on the arms (complete with motor wires already run through). Then the "bottom" clamps, then the bottom board (with power board built already in situ). This is where wiring becomes a bit of a nightmare. But it's still the easiest way to do it.
4. The bolts intrude into the battery space. And they are sharp. And batteries don't like sharp....I got round this by adding 10 washers to the top of the inside bolts:
In addition, I used longer landing leg stacks; this gives enough space for a 11,000mAh 6S, and still have enough space to give some wiring room.
I'd prefer to have these extentions rubberised for isolating the legs from vibes (the camera is mounted there), but maybe another day.
5. The legs have no markings, so you need to make sure everything is the right length, and horizontal. I found the best way was to measure the arm length from frame to motor mount. I measured levelness using a spirit level on the arm and then the motor mount: this was the most logical way to do it. It was a lot of back and forth though, as even tensioning the mounts caused a small rotational effect - Once it was level, I had to apply the same number of turns to each bolt, which took ages!
I also needed to buy the DJI board to fit the APM on; there was nowhere to place it without one! (ignore the dodgy level - one leg is sitting on a cable). Also, it's the only way to place the APM in the dead centre of the airframe.
I also opted for the side "platforms". Mainly so I could host GPS on one to the side for guaranteed view of the sky even with dome on, and optoflow, sonar and FPV camera on the front one. That means getting the corrosponding ones as well to maintain COG!
Front one with sonar (outflow sensor still out of stock for months!). Best place for it, or it gets a signal from the legs! FPV will go here too. Also in shot is the LED CAA lighting controller. The TM1000 will be moved to better COG position once my JST inline voltage detector cable arrives.
Ok, maidenened "Lauren", and ran into a few issues.
It should be noted that the conditions were less than ideal. 20mph winds, 5'c, with the finest hint of showers.
I did hand-held the PID training, as per Dave C, and got P rate of 0.22 (pre D tune) then ultimately 0.2390 (still a fraction high I think - there's a few wee jitters), I of 0.08 (not finished), and D of 0.08. All was well. Until I started flying.
1. Stabilize was lethargic. It was like driving a bus. Stuck STABILIZE KP on channel 6, and it's running better at 3.8. There is more to come though. I might try and hand tune - has anyone had good results this way?
2. I seem to be following the throttle with my sticks. Don't know if it was the wind, but I simply could not find a stable hover in stabilize, unless I used *the tiniest* of stick movement. Rapid climb or descent, then slotting into a hover was impossible. I ramped throttle P rate from 6 to 7, which help a bit. I think. Also, very slow to get the props spinning when starting off, and not a lot happens until about 40% throttle. After that...much better, stick chasing aside. Should I ramp more throttle P?
3. Loiter is a bit wandery. Very windy, I admit, but it didn't really seem to commit to maintaining it's position with any sort of vigor. I had to fly it nearly 100% to keep it within 3m. In a calm spot it was ok. Ideas?
4. More worrying. Voltage drops. I run a brand new 11,000mAh maxamps 3S. Fully juiced before flight, but started to get telemetry voltage warning of 10.5V (my set limit on the TM1000) very early in the flight. I landed...and alarms stopped. Looked down - 11.8V. Used the MP Voltage tracker as well...also a 1-3V drop once the props are under load. I pushed on, nail biting...until the load voltage was just about 9V on the TM1000 (about 10.5V landed, which is my "land now" limit). Got home, and all 3 cells on the charger were reading 3.6 (near perfect). Ideas on voltage drop? Not enough C on the battery maybe? Too cold?
5. Also worrying. A number of unrequested "lands". I know they were unrequested, because I don't have land programmed on anything, and failsafe is RTL. I think it is voltage related because land alway proceeded a fast turn, fast acceleration or aggressive climb.
Logs to follow, once i figure out how to attach them.
Crash: sarcasm not required - I *am* that stupid. I put all 4 props on the wrong way once. Labelling should stop another flip...
When I get a minute (and back to a proper keyboard) ill do the build issues.
Yeah, printing the running direction on each arm of a simple quad reveals the real expert. Very nice build BTW.