You have to know your thrust hover value, if for example you have MOT_THST_HOVER=0.35 then you have to set:
PSC_ACCZ_I=0.7
PSC_ACCZ_P=0.35
this is from the wiki page:
http://ardupilot.org/copter/docs/tuning-process-instructions.html
First of all I would like to thank everyone who post in this thread that Wicked has started. I find it to be the most informative reference for building f4 nano Micro Arducopter on the internet. I read it over and over.
This tuning coversesation is very timely for me as I am preparing for maiden first flight of 3" Iflight IH3 with 1105 4500kv on 3s(850Mah), Split cam and m8q gps.
Currently I have Wicked’s parameters from post# 475 on May 4th loaded. When comparing these on the extended tuning screen to andyp1per recent post there seems to be fairly significant differences.
Must admit I am not to crazy about switching to 4.0 master just yet, as I haven’t even installed props .
My flight style would be closer to Wicked’s video than blasting through trees in acro. However, it didn’t really show attitude in takeoff and descent.
So with great gratitude I would like to ask for anyone or everyone to share there latest full parameter file for the safest and most uneventful first flight.
I truly appreciate the time and frustration pid tuning takes.
Thank you again for this great thread wicked1.
1 Like
Hi, thanks for the kind words.
A LOT has changed since May… My latest parameters I posted yesterday are the best flying I’ve had w/ my little copter. And I’m almost positive even that May one was on Master… which was called 3.7 at the time… I don’t think I’ve ever had the official 3.6 release on my micro copters. (I originally built them as test platforms just to test the new firmwares… Then had so much fun flying them, I never get out my big copters anymore)
And you’ll find significant differences between my older params and my latest… I’ve had PID’s ranging from .03 to .12 for P, over the course of this thread.
Anyway, I don’t think you’ll have any problems w/ 4.0… It will be an official beta sometime soon. should be any day now, if things are going to schedule (probably aren’t, though).
And in any case, everyone’s copter is different so you can’t expect anyone’s parameters to get you flying 100%, w/out doing your own tuning, too.
I’d say get your filters increased like Andy’s or my config. Maybe your loop rate, too. Then do an auto-tune. Then start adjusting from there.
I don’t know how much arducopter experience you have in general… But basically, start in stabilize and get it hovering somewhat smooth. It’s difficult to maintain altitude in stab if you aren’t used to it. But be sure it will at least take off and isn’t shaking. Then, when stab works alright, try alt_hold. If it can maintain altitude, then go for the auto-tune. (basically, don’t jump in to alt hold, or any sort of automatic mode until you’re sure it’s ok in stabilize)
I recently got back from a vacation where I got a lot of videos… I will get some uploaded to youtube whenever I get a few extra minutes, so you can see some more flight examples.
Thank you Wicked1, I will go for it with 4.0 as the windy forecast for the next few days doesn’t permit me to fly anyway. This is my first go at Arducopter and mission planner. First build was a wrecked phantom I bought off the internet and gave it a pixhawk using px4 and qgroundcontrol. I realise this will be totally different animal and am ready to take precautions. I would just like the best stuff I can have in there to reduce failure.
Currently I just have single 1/4" by 1/4" piece of foam taped over barometer smashed between fc and 4 in 1. Would you have any pics of your foam around the edge of fc? How did you attach it? What about heat?
My latest copter only has a small cube of foam over the barometer between the FC and ESC, like yours. I haven’t had any trouble w/ this one… It is a different frame… I’m not sure if that made the difference, or if the code changed and handles the problems more gracefully.
But the one w/ the foam around the edges is basically the same, only longer strips of foam. So, about 20mm long strips of foam pushed in between the FC and ESC from screw to screw, on all 4 edges. I’m not at home to get a photo. I can later, but there’s not much to see… It’s just foam along all 4 edges. Start simple, w/ the single square… If it has trouble, add more.
Wow, the wiki is a lot more informative than it was a year or so ago, last time I looked!
Thanks chobitsfan! 
Hi Andrea, Thanks. This is the correct procedure?
But if i switch from stabilize to alt hold, the copter immediately (30 seconds?) lose altitude then land …
Test AltHold
This test will allow to test the altitude controller and ensure the stability of your aircraft.
- Check MOT_HOVER_LEARN is set to 2. This will allow the controller to learn by itself the correct hover value when flying.
- Take off in STABILIZE and increase altitude to 5m. Switch to AltHold and be ready to switch back to STABILIZE. If the aircraft is hovering at a very low hover throttle value you may hear a reasonably fast oscillation in the motors. Ensure the aircraft has spent at least 30 seconds in hover to let the hover throttle parameter converge to the correct value. Land and disarm the aircraft.
- Set these parameters on ground and preferably disarm (A confident pilot could set them in flight with GCS or CH6 tuning knob):
- PSC_ACCZ_I to 2 x MOT_THST_HOVER
- PSC_ACCZ_P to MOT_THST_HOVER
AltHold starts to move up and down the position and velocity controllers may need to be reduced by 50%. These values are: PSC_POSZ_P and PSC_VELZ_P.
Hi Alberto, Yes this is the correct procedure and it worked for me.
What is your value of MOT_THST_HOVER ?
The wiki says to set a value lower then expected as initial setting for this parameter (say 0.25), then when in hover the correct value should be learned.
Andrea, my current values is:
MOT_THST_HOVER,0.35
PSC_ACCZ_I,1
PSC_ACCZ_P,0.5
MOT_HOVER_LEARN,2
Now alt hold worked!!
I tried autotune single roll axis but don’t start. Both as flight mode and with CH7. I set the motors that do not start with armed quad and gas at zero. what can I try to start autotune?
With 450mA 2s it flies almost 5 min 3.7V cells
Thanks!
It won’t start unless you are hovering stably and your throttle is at the midpoint, so you need the hover thrust set correctly.
Hi,
on mission planner, works CLI interface? For me and Hakrc mini f4 no worked
And another question …
how can autotune work if the D gain starts at 0.001 and I’m already at 0.00016?
You need to set min D waaaaay lower. I use 0.0003 on my 3"
Has anyone using 2s (@maroquio @Hugues, and I forget who else) found any new ESC’s that work well? I’m planning my winter build with the new small f7 board Andy’s been using, and a micro 4k camera. There are a bunch of new 20x20 esc’s on the market, but all of them I’ve found so far say the minimum is 3s. Needs to be blheli_32.
The Aikon I’ve been using is alright, but doesn’t do current over telemetry. And it’s a bit heavy…
@andyp1per, the progress you’re making on the filters is awesome!
This one claims 2S. I have used Spedix on other craft with good success. GetFPV or Raceday Quads.
Spedix GS25 25A 4-in-1 2-4s BLHeli_32 Dshot ESC
Tried it… It cuts out at just below 7v. Which is the case w/ most of the ESC’s that said they support 2s last year. (ori32 being another… I tested at least 5 different ESC’s last year that all said they support 2s, and the results were not good on any of them. Other than the ak32. Blheli_s esc’s were all fine… But I want telemetry)
The voltage can droop down to about 5v on 2s. The ak32 I’m using now works down that low and fails gracefully. The motors start twitching a bit before they go, so that’s a sign to ease up on the throttle and land ASAP.
But I’m mad about them because their documentation said, and still says, supports current over telemetry, and it does NOT. It supports current over analog, but not telemetry. But, at least it’s an option. And maybe I can get analog current working on the f7 board… the f4 I’m using now doesn’t have enough ports for it along w/ telemetry.
I think we had this conversation. I’m using an Ori ESC. But its on 4S. That is maddening that the stated specs are wrong. On Raceday Quads they show several claiming 2s but you likely know better than they do!