ESC response test

[Kisssys]

I had planned to do a bunch of motor response tests but I'm not happy with the vibrations I have on this old thrust rig. I did do 4 different esc's on a RCtimer 5010 motor and 15x5 prop. I brought it up to half throttle and with a step input up 300 us.

The ESC32v2 and ESCv3 are in RPM mode. The Kiss and SimonKnfet were of course open loop. The bottom line is half throttle thrust. The width of the pulse is just how long I held down the shift up 300 us button. The only scale that has any significance is the 200 ms per division as it ramps up and ramps down with the step input.

The ESCv3 is easy to pick out.

[Kisssys]

I removed the SimonK code from the tested Hobby King F30 esc and installed BLheli code. Their closed loop and open loop response.

Added two more to the first post.

[Kisssys]

These tests were done with a Flyduino X2208 1100kv motor with a HQ 8x5 prop. I apply half throttle and press a buttion on my tester that raises the throttle 300us for 300ms, then down 600us for 300ms and then back to half throttle. All tests are open loop. No braking on the V2 which is present on the other esc's.

[Mullet]

Steve,

I'm curious what your tests mean and how you'd interpret them. I'm looking into doing a "warpquad" type quad x... Raven 234mm from FreeBird RC... and it looks like their 12A ESC with BLHeli firmware might be the ticket based on your findings in the M4 Swarm thread. I know you used a AfroESC, but I'd think their 12A ESC would be comparable.

[Kisssys]

This got a lot too long but wanted to tell a story and answer a post at the same time.

The Blheli code with lightly damped does a good job on the esc's that have all nfets. Most of
the multirotor esc's now can raise the speed fairly rapidly, only those with active braking can
get the speed back down quickly. The ESC32v3 as you see is the closest to a square wave which would be
perfect response. The Kiss and Blheli and certain SimonK code do a fair job. All have flown the 250 size machines well
with Menno's Hott code that has provisions for using non closed loop esc's with Quatos.

It would probably be worthwhile to post a short version of what I went through getting a ZMR-250 to fly correctly.
When you put a fast esc on a poor frame it will bring out all kinds of problems that will not show
with a slower esc. I had a hard time with the Warpquad type frame until I added a second layer of 2mm carbon
to the booms to lower the vibration levels. From what I've seen the flat boom frames must lack the necessary
rotational rigidity that is necessary to prevent a bad reaction between the flight control, motors and frame. Like shaking a sheet in the wind with the ripples and standing waves all over the place.

I did a H-frame a few month's ago that turned out real well. It had KISS18A esc's, M4 and Emax1804 motors. For this post I just took the frame and added weight to bring it up to the weight of the ZMR-250 frame I am going to show you in a bit. Here is a screen capture of a short indoor hover. Fast Kiss18a esc's running a M4 with Quatos with Menno's Hott code with Hott turned off.

I told Jussi month's ago I would do a build with a common frame and I chose the ZMR-250. That turned out to be a big mistake. I might of gotten a ZMR-250 clone but the workmanship looked great and it is solid carbon fiber. While trying to get it to work I tried 3 different motor sets( Sunnysky,Emax and MultistarElite),3 different esc's(Kiss18A,AfroESC,F-12a Red), 2 different M4 board's and a Flip32 board before I finally added a brace across the bottom of the front and rear booms. This brought the vibration levels down to an acceptable level.

Until I added the brace I consistently got bad vibrations, nothing I tried made any difference and could not get the vibration levels down. They all looked like this.

Before I added the brace I decided to put a Flip32 board onboard and take a look at oneshot125. It was no surprise it was almost exactly the same, I mounted the M4 board on the bottom to read the vibration. I even tried oneshot with a slight improvement. I think the lower resolution of oneshot did more to quiet it down than anything else.

I finally added the brace shown in the next picture cut from G10 epoxy and fastened to the existing bolts and boom holes.

This is a capture of the current vibration levels with the boom brace. Not as good as the above H-frame but isn't too bad.

So after all that I won't wait so long to give up on the frame, just because the other guy uses it doesn't mean it's really
up to snuff. Won't be long before I get rid of the ZMR-250.

Cheers
Steve

[axelnied]

Steve,
Thank you for this really nice summary...!
Can you share some details about the frame you used? I am asking to extract some lessons-learned out of it.
- What is the thickness of the original arms (thickness of carbon sheets)?
- Did they have a noticeable "flex" to it when you would try to bend by hand?
- What is the thickness of the G10 brace material?

[chestnut]

[Kisssys]

Those are load cell readings from a thrust stand. I apply 1500 microseconds or half throttle and center the thrust reading on the scope. Then I apply an increase of throttle of 300 us ( 1800us ) for 300 ms. Then I go all the way down to 1200us for 300 ms and then back to half throttle. I'm only looking at response time and have made no effort to calibrate the thrust part of the reading. The lowering of the throttle is a much bigger command and must be considered in viewing the test. It's just a simple evaluation of what damping does.

Cheers
Steve

[Kisssys]

axelnied

Haven't been home much and couldn't measure the booms until now. The booms look to be of good carbon fiber, very stiff and the inner layers look to be carbon as read with an ohmeter. They are 2.8 mm thick. The strap is made from 1.6mm G-10 glass epoxy.

If the frame hadn't felt so unbelievably solid I wouldn't have played around with it so long, it's hard to believe its that bad.
I did see a post on the ZMR forum that someone felt the frames were vibration prone. This might be one of a kind but I doubt it.

I've included a DXF file of the repair boom parts, remove the ".txt" extension.

Cheers
Steve

[chestnut]