Thanks for the review, i am completely puzzled of what that is… This model flew fine on betaflight for 5 months, very reliable. I then transplanted FC to kakute, flew it on betaflight, it was flawless. Then flashed arducopter and spent some time tuning it, and ended up with those twitches. So you would suggest replacing escs and motors?
Everything on powertrain and signal wires is soldered. Is it possible to tell from log which motor is doing it?
I suspected it is a code issue with dshot implementation. Like command stream to esc “pauses” and then recovers. But i have no idea how to check it.
Other people running dshot aren’t having the same problems, though… I’ve been flying my little copter for at least 1/2 hour a day for weeks now… dshot 600. But, we’ve got different ESC’s and I could see some being more sensitive to certain issues than others.
Like I said in your thread about this a couple weeks ago, you can see the copter dips to the same side in the logs each time this happens, so I’d look at whichever motor that is…
You can tell from the logs. It always dipped to the same side. Roll and Pitch both go the same direction each time it happens… and there is a way to know from the logs which is front/back, left/right… (But I don’t know the directions off the top of my head… But the roll trace going down definitely means either right or left, every time)
And like I mentioned in your thread, I look at Pitch and Desired Pitch, Roll and Desired Roll. From that you can see what the copter wants to do, and what it is actually doing. And you can see it wants roll to be flat, level. But in reality it drops to one side. The trace went down (if I remember correctly), but like I said, I don’t know if that corresponds to left or right.
thx.
it has 1.5 inch signal wire away from everything. and did not interfere with anything using betaflight, but i will look at it and will try to unsolder and resolder it.
the more i deal with this hobby more i get to a conclusion that it cannot be done anymore using simple non-shielded wires. any signal wire has to be shielded and grounded - it seems to be quite obvious now and with this dshot it seems to be the most dangerous part if external interference can actually distort that stream enough to cause motor failures.
rear left signal wire goes close enough to the esp8266 board, but i would not expect it to cause an interference high enough to distort serial flow. but, i will look at it.
it is only every time before i was told it is a fault of hardware - it was not. as i do know hardware, well, pretty well. but, who knows. if all points to a possible hardware fault, i can take it apart.
why i am skeptical - as i updated other model i have - i saw similar twitches on it as well, and that other model never had it before. so, i am not sure, but, that other model uses same exact ESCs as this rooster model and it indeed can be an ESC level issue - if it is too sensitive to the signal quality.
i know it is a complete offtopic here, but, as i just look at the model - the r-xsr comes with those antennas that have an inner core and a thin shield around - does anybody here know if anybody online actually sells such wire in bulk? as it wold be an ideal solution for a single signal wire that would remove any possibility of an interference to affect the esc.
I think the new tiny antenna wires are .81mm coax, but I can’t find an rg designation for it…
And I can’t find any being sold by the spool… only pre-made w/ connectors.
I have bought the longest 2.4 antenna wire I can find and cut it to length when I need coax. I needed it for my video wire in my last build. W/ the cheap prices of imported RC stuff, that’s probably less expensive than buying a spool, anyway.
RG-113 is the slightly bigger antenna wire that was standard a few years ago… 1.13 mm OD.
(I might be wrong about this RG designation… One website had it listed as rg113, but more searching hasn’t come up w/ much w/ that number)
The only issue is I have found that wire is brittle… at least the cheap antennas I was getting it from. I put a dot of hotmelt on each end after it’s soldered in place, for some strain relief.
If you want the smallest possible, you’ll hve to get r-xsr antennas, and use that wire.
If you can deal w/ the slightly thicker wire (I’d say you can), then something like this would be great:
well, smallest available is 28awg, still too thick. my point here is that this cable is needed, ideally, to connect to gps and compass, same cable needed for split2. to ESCs we would need 2 connector. they also seem to have 33awg shielded, but only in bulk. damn.
That’s large cable… If you want small multi-conductor or coax try Cooner. We use many of their products for sensor applications.http://www.coonerwire.com/
right - but you cannot buy, most likely, just 5ft of it from the manufacturer. i see tons of microminiature cables like that, the issue is - none are available on online sales if you need 2-10ft only. if you want a 1000ft spool - sure, no prob.
i decided to get from getfpv an upgrade - a set of Holybro Tekko32 F3 35A and a set of Brotherhobby Speed Shield 2207.5 2150kv motors.
will try to get some 6S lipos as well while they are on sale. will be one crazy rooster after that, but i suspect arducopter will not be able to handle such acceleration - a 6s kick on a 2150kv.
I just received some Brotherhobby 1402 8100’s to try. Can’t decide if they’re going on this one or my betaflight copter… I’m really liking my tiny arducopter, so they might end up going here. It’s flying just about as well as my betaflight copter now! Plus the added security of RTL and ability to do auto stuff… Love it!
I’ve become a fan of the 4 in 1 ESC’s. Most of the issues w/ magnetometer interference are actually from the DC wires leading to the ESC’s. And I’ve found that having the ESC’s directly under the FC/Magnetometer causes less interference than a traditional setup w/ PDB and separate ESC’s. The wires going from the ESC’s to the motors don’t actually cause much interference, so it’s not an issue that those end up being longer. And, if your Dshot issue has to do w/ the signal from FC to ESC, that’s all much shorter, too.
I’m sure yours will be fine… but something to consider next time.
4in1 is fun, but i do not have space in the stack on the rooster for it - all space have been taken already.
plus, i have had an esc to burn on me twice. if 4in1 burns - it will be a complete loss - so i prefer to use discrete single escs - at least on those armattans where wings are ideal for their placements.
on a small model i would agree 100% - center stack location is the best - on the 5" rooster i prefer wings.
on oct 5th chinahobby line will have sale on 5s 1500mah lipos - i guess i get those instead of 6s for the beginning - i have a ton of 4s 1500s - so will add some 5s 1500s to the mix to see how it will go.
then i also want to think of some 3" model to make, but need to find a safe frame that will be flyable in house, if such exists. recent fiasco with this rooster failing to the side kinda scared sh$t out of me when it flipped and made some attempt to smoke itself. no one still explained btw how is it possible for copter code to ignore kill switch - but it has happened.
That’s a good point… I’ve had that happen a couple times, too… Not on a 4in1 yet, but when it happens, that will be a pain to fix… ESC is on the bottom of the stack, so that’s a complete rebuild.
Look at FlexRC frames. (he gave me this frame for this project, so this is a paid advertisement )
But, I’m really happy w/ them! I have a 2" prop “pico owl” frame, which has ducts and is safe indoors.
He’s got a 2.5" called the lora (or something similar) that also has ducts and would be safe inside. You’ve got to get creative w/ mounting everything on them for arducotper, but it’s definitely possible.
3" is getting a little big for indoors, but I fly mine inside all the time.
The frame I used for this project (in the top photo) is the “ascent 3” light" frame.
I can totally relate to the out of control copter scare… That’s why I built this little one. 650mm copter is very scary… 250mm isn’t much better, and is still enough weight to do some serious damage to people or property. These little 150mm, 3" prop and smaller microcopters are relatively safe and inexpensive for beta testing things. This one did cut me a few times, while I was working out my altitude hold issues! But, no damage to my home or anything else
i have now biggest one at 700mm - en enhanced tarot 690 - i have much more confidence in it than in that rooster.
as when rooster flips and pumps 100A into motors and once it darted in the horizontal direction for around of 15m like a bullet - i am not quick enough to kill it and it still had a ton of speed as i killed it mid-movement.
this tarot is much slower and would never get too far as i would kill it quicker than it could move. plus, all bigger models i have are on slow kv motors - 450mm one is on 900kv, tarot is 680kv. it is not fast enough to do any damage, that rooster would probably chop off all meat from my face if it would get into me pumping 30A into each motor while refusing to die. but, so far i had it only once when kill switch did not fire and then altered my codebase to use 100ms insead of 400ms for the ‘sleep’ period - it feels much better now.
so that incident with kill switch malfunction is a grave concern for me now - i do not think devs realize what danger it is if there is some condition in there that may prevent kill switch from working.
About the kill switch…
On your smaller fast copters, you might want to switch that to motor interlock.
You can still flip it to kill the motors almost instantly.
But, it has an extra layer of security. W/ motor interlock, the copter will not ever stop stabilizing itself if the motors are switched “on”. So, you can fly for fun, find yourself way to high, and completely lower the throttle, and it will still maintain level even if it’s basically free falling, and there’s no chance of it accidentally disarming itself mid-air.
It’s similar to “air mode” in betaflight.
as i do know hardware, well, pretty well. but, who knows. if all points to a possible hardware fault, i can take it apart.
)