CranialRectosis' Alien Build

[cranialrectosis]

If I am running ESCs with a heatsink I can get away with putting the ESCs in the frame. When I am running ESCs without a heatsink or where I will push enough power through them that the heatsink must have air, I put the ESCS on the boom under the fat part of the prop so the propwash will cool the ESCs.

I always cover carbon fiber with a non-conductive material (heat shrink) that is ONLY to isolate the CF from electronics. Each boom on this copter will get heat shrink just to cover part of the boom where the ESC will sit. If I individually heat shrink the ESCs, I use two sided tape between the two layers of heat shrink and a zip tie to keep it from shifting. If I use heat shrink to cover both the boom and the ESC at the same time, the heat shrink holds the ESC on the boom and I use a zip tie to keep it from sliding around.

I found in the past that two sided tape under a KISS ESC without a layer of heat shrink between the tape and the ESC causes the tape to melt and get into the ESC and the ESC simply sits on the boom anyway once the tape melts away.

If the ESCs can shift in flight it stresses the solder joints.

If ESCs can't breathe, they may overheat.

If electronics come into contact with carbon fiber it is possible to short out the electronics.

Remember these rules and mod your build accordingly.

 

[cranialrectosis]

Well color me stumped.

I finally got some daylight time and took the copter outside in the cold and the wind to try this autotune thing and the board won't boot. The ESCs don't complete the startup sequence and the light goes out. The Sparky2 likewise lights up and then just goes dark before it completes the boot sequence.

All the video equipment is working fine. Wiggling the connection does nothing whether the board is booted or not.

I brought it inside, took the props off and connected the lipo and it works just fine.

Put the props on and went back outside and the board and ESCs won't boot.

Bring it back in, take the props off and have no problems at all.

I either ruined the Pololu or have botched a solder job somewhere or I have a component that can't work in 30mph wind and 20* temps.

At any rate, I have tried this same cycle 3 times today with the same result each time.


It won't fly so I am taking the copter apart down to the PDB again.

 

[Craftydan]

um . . . I think I see your problem.

Your ESCs hate those props.

you should look into getting nicer ones . . . look for the "KISS compatible" logo

 

[cranialrectosis]

Thanks, Dan.

Work is calling again. I may not have today off after all.

Think I am just gonna go get lunch and put both work and the copter down for an hour.

 

[cranialrectosis]

Work is dealt with. Lunch is dealt with. The copter is dis-assembled. I am uploading proof that this thing flew yesterday.

I put liquid electrical tape on my PDB connections. Once I got the Sparky2 out of the way, I used a thumbnail to scrape the black goop off the PDB pads and the ground lead going to the Pololu immediately broke off with almost no pressure on the wire.

I think after I re-soldered it last time, I wiggled it as I stuffed wiring up under the Sparky2 and in wiggling the wires, broke the negative lead 99% of the way through causing intermittent brown outs.


So far, the hardest part of this build has been the Pololu (as crazy as that sounds). I can see why the MotoMoto board with its dedicated place to put the Pololu is so freaking popular! There is no room for the Pololu under a Sparky2 because of the ports on the Sparky2. I have already broken one Pololu on this build and have probably soldered, de-soldered and moved the Pololus on this copter 10 times now trying to find an out-of-the-way place to put it.

On this frame a 5v Pololu takes up a lot of real estate if you can't put it under the flight controller.

The movement is taking a toll on the wires. I am replacing them all now. Getting out the 2.0 magnifying glasses so I can see it...

 

[cranialrectosis]

I finally stopped trying to be fancy and zip tied the Pololu to a standoff. It ain't purdy but it boots.

It flies pretty well. It is solid in the air like a WarpQuad and on 3S it hovers at just under 50% throttle.

I tried an autotune but it didn't start. I probably have to turn something on in the GUI so I am re-connecting to the PC now. Hopefully I can figure it out before the sun goes down.

I did get in a better, longer test flight. Working on the video now too.

 

[cranialrectosis]

I got autotune working!

It flew well enough I came back and ran autotune on 4S. Pretty nice feature. Easy to use, easy to fly and the copter seems to respond well. I think I found out how to raise the rates. It's too dark outside to fly and test it out though.

Not sure how to post my auto tune results to the forum. The autotune has a button but it fails to update the site.

Anyway, here is the Alien 6" autotuning on 4S.

 

[jipp]

cool. you are making progress. one step forward 2 back then 4 forward right? heh.

well i did what you said. solder the power leads to the PDB. then solder a jst to the FPV and checked with multi meter sure enough 12v.. so im good to go now.. i running the sn20 bleheli ESC. im not sure i have shrink tubing big enough to fit on the arms.. but im not taking the shrink wrapping off them.. i guess ill just make things easier for my self and not use the parachute cable.. and just zip tie them to the arms.. sigh.. i tried to make it neat and i have to make it look like my last one. laughs. but its probably best i guess. ill save the parachute cable for a larger build.. also i was having a hell of a time last time i went to solder.. so i bought some new tips.. 10 pack of different sizes and shape.. holy cow.. it was a breeze to solder up compared to last time.. i guess the lesson here is to use the right tool for the job.. but i guess where is the fun in that. hah.

tomorrow i guess ill consider trying to put the esc in the frame. but i may have to put them on the arms.. which sucks.
thanks for the help. i am using 12mm riser tho. so the PDB will be under there so i guess it will be a little cleaner than my last build.

chris.

 

[jhitesma]

Woo Hoo! Isn't that autotune wiggle fun When it stops it always feels better to me just because it's not jiggling like a bowl of jello aymore The new tune doesn't actually get applied until you hook it up to GCS and accept the results. Partly because the actual calculations are more than the FC can handle so GCS has to do them, but mostly because there's a chance it will lock onto a bad tune so you're expected to review the numbers before using them

There are some interesting tweaks to AT happening on dRonin right now. They're looking at shortening the time it wiggles, and making it wiggle more aggressively so it can find a better tune faster. Mostly theoretical but I just got a sparky the lead dev sent me to replace my flyingF4 so my test mule can try some experimental stuff later this week

Not sure why the autotune share button isn't working for you You could always post a screen shot...or just share the tau which is what I"m most interested in seeing how that setup does on 4s.


Rates are on the Stabilization page Advanced tab. They're listed as "Stick Scaling" near the middle of the screen. You can see the curves adjust as you tweak them. Unlike *flight where rates are a multiplier here they're actually a real unit, it's the number of degrees per second you want it to move at full stick. About 450 or so will let you do single flips fairly easily and is where I like to start on a new build. 740ish with 60-80% expo is a good goal once you're feeling more confident with it, should let you do double flips/rolls pretty easily and still have fairly tame sticks towards the center.

You can also give AcroPlus mode a try...rates kind of go out the window with that since PID's kind of go out the window at high stick movements then Makes things kind of scary at full stick but also kind of fun. I've only briefly tested it so far but want to do some more with it.

 

[cranialrectosis]

cool. you are making progress. one step forward 2 back then 4 forward right? heh.

Yeah. That seems to be the CranialRectosis way today. I powered up the quad this morning and guess what? The board won't boot. The ESCs light up and start to boot and then just quit. I still have some type of power problem that happens when I bring the copter inside from outside. I have one more 5V connection to test. A Chinese one...

well i did what you said. solder the power leads to the PDB. then solder a jst to the FPV and checked with multi meter sure enough 12v.. so im good to go now.. i running the sn20 bleheli ESC. im not sure i have shrink tubing big enough to fit on the arms.. but im not taking the shrink wrapping off them.. i guess ill just make things easier for my self and not use the parachute cable.. and just zip tie them to the arms.. sigh.. i tried to make it neat and i have to make it look like my last one. laughs. but its probably best i guess. ill save the parachute cable for a larger build.. also i was having a hell of a time last time i went to solder.. so i bought some new tips.. 10 pack of different sizes and shape.. holy cow.. it was a breeze to solder up compared to last time.. i guess the lesson here is to use the right tool for the job.. but i guess where is the fun in that. hah.

tomorrow i guess ill consider trying to put the esc in the frame. but i may have to put them on the arms.. which sucks.
thanks for the help. i am using 12mm riser tho. so the PDB will be under there so i guess it will be a little cleaner than my last build.

chris.

Looking forward to it, Chris. Each one is different and each one is a new challenge. Each one goes through a process to become the thing we want. It's the journey I love. The destination is the same for all of these.

 

[cranialrectosis]

Woo Hoo! Isn't that autotune wiggle fun When it stops it always feels better to me just because it's not jiggling like a bowl of jello aymore The new tune doesn't actually get applied until you hook it up to GCS and accept the results. Partly because the actual calculations are more than the FC can handle so GCS has to do them, but mostly because there's a chance it will lock onto a bad tune so you're expected to review the numbers before using them

There are some interesting tweaks to AT happening on dRonin right now. They're looking at shortening the time it wiggles, and making it wiggle more aggressively so it can find a better tune faster. Mostly theoretical but I just got a sparky the lead dev sent me to replace my flyingF4 so my test mule can try some experimental stuff later this week

Autotune was pretty cool. I was wondering about the force of it. I noticed the copter shook harder with 4S than 3S. I thought it jumped a bit harder too. When you set P gain up WAY too high and take off, you get what's known as a 'toilet bowl wobble' where the copter wobbles the motors sequentially in a circle. The wobbles get more and more violent and as the copter tries to contain it, the FC adds more power to the motors to do so. This makes the wobble worse but it also makes the copter rise as if you were increasing throttle without you applying any throttle. Cutting the throttle to 0 and falling is the only option as ANY throttle at all allows the FC to increase throttle on its own trying in vain to offset the wobble. The copter will head for the moon even as you reduce throttle until you chop the throttle completely.

I did Autotune the first time with 3S because I was concerned that an out of control wobble would carry the copter out of my yard in < 10 seconds (ask me how I know all this :black_eyed. I was pleasantly surprised but would wonder how this would work on a tiny little thing on 6S (cause you know it's gonna happen).

Not sure why the autotune share button isn't working for you You could always post a screen shot...or just share the tau which is what I"m most interested in seeing how that setup does on 4s.

The TAU Labs forums have blocked my Avatar (too big by 30 pixels wide) and blocks all my copter photos (too big too) so I just am not bothering with it. I take too long posting as it is. Dinking with the pictures to meet their arbitrary rules is too much hassle (pictures and video are not my thing). Any screenshot I take will likely be too large for their site and will be blocked too.

IMO, if they wanna go mainstream they need a simpler site with fewer restrictions requiring non-topical specific skills to be able to post a photo.

Edit: Today I tried to post the video of the copter running autotune on the Tau Labs site. I copied the link and stuck it between the 'youtube' tags and got a message "You are not allowed to post URLs!". Sorry, done with the Tau site.

Rates are on the Stabilization page Advanced tab. They're listed as "Stick Scaling" near the middle of the screen. You can see the curves adjust as you tweak them. Unlike *flight where rates are a multiplier here they're actually a real unit, it's the number of degrees per second you want it to move at full stick. About 450 or so will let you do single flips fairly easily and is where I like to start on a new build. 740ish with 60-80% expo is a good goal once you're feeling more confident with it, should let you do double flips/rolls pretty easily and still have fairly tame sticks towards the center.

You can also give AcroPlus mode a try...rates kind of go out the window with that since PID's kind of go out the window at high stick movements then Makes things kind of scary at full stick but also kind of fun. I've only briefly tested it so far but want to do some more with it.

I removed autotune and set up AcroPlus as option 3 and I jacked up the rates to 500 as per a thread (your answer) on Tau. The sun came up without snow this morning but I have a ton of work due today and the board won't boot when I connect the lipo again so it will be a few hours or days before I can try again.

Which screen do you want a screenshot of? Not sure where a TAU rating exists.

 

[jipp]

if you spending money may as well pick up one of these too

https://www.multirotorsuperstore.com/quadrevo-powerosd.html

chris.

 

[jhitesma]

Autotune was pretty cool. I was wondering about the force of it. I noticed the copter shook harder with 4S than 3S.

It may have been because you had already run a tune on 3S so your gains were higher. When doing the autotune wiggle it's flying on the current settings. So when you did your 3S autotune you were on stock settings, when you did the 4S you were on the 3S tune which was almost certainly more aggressive than stock.

I did Autotune the first time with 3S because I was concerned that an out of control wobble would carry the copter out of my yard in < 10 seconds (ask me how I know all this :black_eyed. I was pleasantly surprised but would wonder how this would work on a tiny little thing on 6S (cause you know it's gonna happen).

I've yet to have any issues, but I do test under my carport or RV shade just in case it tries to take off on me You can always abort the autotune so I'm not too scared of it anymore.

The TAU Labs forums have blocked my Avatar (too big by 30 pixels wide) and blocks all my copter photos (too big too) so I just am not bothering with it. I take too long posting as it is. Dinking with the pictures to meet their arbitrary rules is too much hassle (pictures and video are not my thing). Any screenshot I take will likely be too large for their site and will be blocked too.

FWIW I have the same problem here I usually grab my photos from FB to re-post here since that's the laziest way I've come up with to resize them without actually having to manually resize them.

And I was talking about posting a screen shot here for us anyway

IMO, if they wanna go mainstream they need a simpler site with fewer restrictions requiring non-topical specific skills to be able to post a photo.

Edit: Today I tried to post the video of the copter running autotune on the Tau Labs site. I copied the link and stuck it between the 'youtube' tags and got a message "You are not allowed to post URLs!". Sorry, done with the Tau site.

Yes, they do have a limit that you're not allowed to post links (and possibly photos) until you've made a few posts. It tends to help cut down on spam posts and is something quite a few forums do. But I do agree their forum is rather frustrating to use and so is their documentation and a bunch of things, and the leadership isn't interested in fixing them...which is why dRonin is happening

Which screen do you want a screenshot of? Not sure where a TAU rating exists.

The autotune page where the results of the tune are shown.

 

[cranialrectosis]

if you spending money may as well pick up one of these too

https://www.multirotorsuperstore.com/quadrevo-powerosd.html

chris.

Got me two of these instead.

 

[cranialrectosis]

Is this what you are looking for?


Edit:
Adding the rates page.


I believe this means I have a TAU of 30 (30C limit on the lipo raising this?) and am running rates of 450/450/500 with a 30% expo. I haven't yet figured out expo on the Taranis so this seems simpler for now.

 

[cranialrectosis]

I dunno what's up with the power but I plugged in a 4S, the Sparky2 and ESCs booted so I flew some laps around the yard.

Raising rates again.


The copter responds with purpose and holds attitude all the way across the yard. Yaw is slow but there is no drift when hovering or when stopping a fast move.

The tune seems to be pretty spot on. Something still feels a bit 'off' though. Since I am playing with both TAU and the Taranis for the first time, I suspect it is just gonna take me a bit to get used to it.

I also connected the GPS module to the OSD. I am acquiring some satellites and may try a simple FPV flight around the yard today.

 

[jhitesma]

Yep 0.0302 is a 30.2ms tau...not bad...but I really would have expected better. What did you get on 4s? Could be the weight of the camera, or just not optimal prop/voltage for those motors. The pack rating wouldn't really make a difference here since the tune happens around hover not at higher throttle where a drop in current would come into play.

Or it could just be tau's autotune not doing as well as dRonin's causing your results to be higher than I expected. I could send you a pre-release of dRonin to try...but your alien is a bit too nice to subject to testing pre-release code I think In a few weeks it should be ready to try and I bet you'll get an even better tune out of it.

May want to boost the yaw rate some more if it's still feeling sluggish to you, or drop the expo on it at least. I know a number of guys who run lower rates on yaw than on roll/pitch so I'm surprised it's sluggish for you. The yaw may benefit from a bit of manual tuning too - the autotune really only does roll/pitch. There's a procedure called "fujotune" that fujin came up with for starting yaw values...that seems to give most people good results...but makes the people who really understand the PID controller shake their head

If you want to try it fujotune is:

(defun fujotune (rp pp ri pi)
(message "yp: %f\nyi: %f"
(+ (max rp pp) (abs (- rp pp)))
(+ (max ri pi) (abs (- ri pi)))))

Basically says for P (or I) on Yaw take the larger of Roll/Pitch's P (or I) and multiply it by the difference between roll and pitch's P's (or I's) And repeat for I (or P).

Which doesn't make a lot of sense really...but apparently works for some people <shrug>


How's your CG/balance? I'm kind of surprised that the gain's aren't closer on that symetrical frame and that the noise in pitch was so much higher (though still plenty low for a good tune.) Looks like you may have a slight bias one way or the other on weight front/rear.

You could also try adjusting the sliders on the autotune to get a bit snappier of a tune if you want - drop the damping down to 1.0 and raise the noise sensitivity to 1.3 and it might perk things up.

 

[cranialrectosis]

That was 4S without the GoPro.

The lipos start to sag at about 2 1/2 minutes and the ESCs start shutting down at about 40% because of the voltage sag. I may be able to sag these lipos on the first punchout with a fresh charge.

I ran autotune without the GoPro. I sat the lipo forward but ran out of cable to plug it in before I got CG where I wanted it (damn good guess, Jason). CG is 1/2" too far back on this tune.

I didn't get a chance to fly again today after I updated to the latest values. They are programmed but not tested. Ran out of daylight.

I may get some daylight tomorrow. Snow and HIGH wind Saturday at the latest so any flying I do this week needs to happen soon.

 

[jhitesma]

That was 4S without the GoPro.

The lipos start to sag at about 2 1/2 minutes and the ESCs start shutting down at about 40% because of the voltage sag. I may be able to sag these lipos on the first punchout with a fresh charge.

Hmm, I'd still expect lower Tau's unless the battery is sagging at hover. But like I said I've spent the last two weeks looking at dRonin autotunes and due to the changes there we're seeing considerably lower Tau values.

I suppose your thinner air could also have something to do with it, props don't have quite as much to bite into so things may not be quite as responsive.

I ran autotune without the GoPro. I sat the lipo forward but ran out of cable to plug it in before I got CG where I wanted it (damn good guess, Jason). CG is 1/2" too far back on this tune.

It's best to run autotune setup the way you're going to fly. So if you plan on flying with the gopro tune with it on there.

As for the CG...it's in the tune numbers The higher noise and lower gain/bias on the Pitch are a sign that your pitch axis isn't well balanced. On a symetrical frame like that I'd expect to see nearly identical roll/pitch values if you were well balanced. Wind could also play a factor, but it didn't look that windy in your video...just cold

I may get some daylight tomorrow. Snow and HIGH wind Saturday at the latest so any flying I do this week needs to happen soon.

Glad to see you've got it flying at least! Can't wait for dRonin to be a bit further along so you can give it a try. They just got the basic page on stabilization settings upgraded so you can set your D term, stick scaling, and expo from there now which makes things a lot less intimidating.

I got the sparky for my ex-flyingF4 and almost have it working...but may have uncovered a flaw in dRonin's Sparky support. Unfortunately I then decided to go nuclear and reflash the bootloader...which didn't go so smoothly. Still have a few options to recover it but will be a day or two before I can stat testing that setup again.

Also may have damaged the silabs chips on the built in ESC's on my 280 Made up a harness so I could use the multi mode on blheli suite to change settings on all four at once - but got my C2D and C2CK lines mixed up. Now even my previously working single ESC harness gets no response from them And still waiting on the replacement replacement motor for my 250. So all 3 of my 32bit quads are down for the count. I could get gutsy and put dRonin on the hex...but the hex is too nice to subject to that just yet...I'm waiting until release candidate status to test on there.

 

[cranialrectosis]

I can autotune anytime. I wanna learn to fly this without the GoPro and when I am confident with it I will re-autotune.

I connected my 4S lipo to the copter first thing this morning, sat down with the goggles to try FPV and the board rebooted before I could arm it.

So, I finished my audits and took the rest of the day off, flew the snott out of my Twitchity mini-Tricopter for peace of mind and tore this thing down to the PDB again. I replaced the Pololu and every 5V connection to the Sparky2 including the pins I soldered to the Sparky2.

The board is booting again now and the intermittent disconnect seems to be gone. For now.

So I connected a 4S lipo and sat down with the goggles and noticed that the video is pretty good. I like the new OSD and I like this cam. Then I armed the copter and spun the motors and see almost total white out from noise.

I have two questions for the forum.
1. I have a cam, OSD and vtx. Which of those should be powered by the LC filter?
2. Does this further indicate voltage sag on the lipo at hover?


Gonna go burn another lipo in the tricopter.