hey,
i use the hobbywing flyfun [also called pentium or fentium] escs in nearly every multicopter because they do their work very well.

but lately ive built 2 quads with which i have some troubles. one quad is built with hobbywing flyfun 18A and the second with hobbywing flyfun 25A. they start to shake after a faster stickinput on roll or nick and also during increasing throttle. i made some research of what could cause the problem.i played with pids, changed fc and changed fc fw. no change in flying behavior after every single change.

then i took a closer look to the escs and found some difference to the hw escs from last year. after they changed the atmega to a SIL chip last year, it seems that they now changed the fets [thats what i think it is]. or made a mix of different types. they also changed the layout and the esc is some mm longer.

i think that the new fet type is causing the problems maybe due a slower reaction time or they changed the fw due the new mix of fets. i cant say that this is the reason by 100%, but this is my suspicion.

the version changed from HW139 V2.0 to V4.0

so please let me know if u made the same experience with new hobbywing escs and be warned that the 2012 hobbying esc could be not suitable for multicopters any more.


hobbywing flyfun 18A from 2011. HW139 V2.0.





hobbywing flyfun 18A from 2012. HW139 V4.0.






hobbywing flyfun 25A from 2012. HW139 V4.0.

I think you're on to something. I just coverted my quad that used older Fentiums (pre dec 2011) 30A to a Y6 setup. I added 2 rctimer motors with the exact same specs as my current jdrones ones (poles, ohms, size, kV-- supposedly the same motors), and 2 newer FlyFuns I got last month and threw the new stuff on the left arm of the Y6. Calibrated and programmed the same.

Upon take off it immediate goes left, like the timing is way off. I recalb and re-leveled the AHRS several times just keeps leaning left, like the newer escs need more puslewidth. I'm going to swap 2 older fentiums in to see if it's the newer escs. Something is definitely different and considering the simonk firmware only works with the atmel set...

Lot's of ESC vendors have changed from Atmega to the SIL chips, it's getting harder to find an Atmega to use with Simonk's firmware.

The SIL firmware might have delays in it.

I think Quax is looking into better firmware for SIL based ESCs you might want to PM him.

i think its not only because of the SIL chips because the hw esc from second half last year also have the SIL chip and work great.
it seems to be because of the new layout/fet mix and/or a new fw for the new hardware.
unfortunately its not possible to test if its only the fw because its not possible to try the simonk fw due the SIL chip incompatibility.
quax made a fw for his esc with SIL chip but its only for these special escs.

yupp I knew something wasnt right I recently purchased the hobbyflyfun 10amp and 18amp for my quads and yupp they were not very smooth like the pentiums i had and have..Ive tried different settings played with the pid's still the same response very jittery response to throttle not very smooth.
then I swapped them out for my 18amp pentiums waaaalaaa all problems solved smooth as silk..throttle response was smooth and precise..

Warthox you are definetly right about the new batch of flyfun esc"s somethings is not right..it would be nice if someone would come out with some firmware for these..

Originally Posted by recharged95

I think you're on to something. I just coverted my quad that used older Fentiums (pre dec 2011) 30A to a Y6 setup. I added 2 rctimer motors with the exact same specs as my current jdrones ones (poles, ohms, size, kV-- supposedly the same motors), and 2 newer FlyFuns I got last month and threw the new stuff on the left arm of the Y6. Calibrated and programmed the same.

Upon take off it immediate goes left, like the timing is way off. I recalb and re-leveled the AHRS several times just keeps leaning left, like the newer escs need more puslewidth. I'm going to swap 2 older fentiums in to see if it's the newer escs. Something is definitely different and considering the simonk firmware only works with the atmel set...

I am just a lurker of sorts here but have a question.

I know you're looking at the ESCs as a component here but I can't help but wonder how well the motors compare to each other. Uneven power delivery and differences in the Kv could be more of an issue than the ESCs as far as how the motor compare for power outputs.

Have you taken anything like a Eagle Tree eLogger and connected it between the ESC and battery (or power board) and looked at the smoothness of the signal traces and excursions or variations in the signals? And compared RPM readings and the smoothness and variation in the RPM between the motors?

The factory winds on motors from HK and the other places with like priced motors are not done that well. I think they might be done by grade school kids. And maybe even grade school kids at a school for the blind.

As an example of a "bad" factory wind, the first image is a comparison of a factory wind with parallel strands and a rewind with a single larger strand of wire.

The unevenly laid strands in the parallel strand wind with crossing turns and variation in the numbers of turns from one arm to another are things that go to smoothness in running and the actual Kv that will result from the wind. So there is always going to be a considerable amount of motor to motor variation even when you have duplicate motors bought at the same time.

If you have one motor that is doing that or two out of three of four that are or are not doing it, I think it could account for some of the issues you are seeing.

I've rewound a number the DAT-750 motors for use on planes with a single motor. The next four images are of a TurboRIX DAT-750 before and after rewinding. It is running an APC 13 x 8 prop in the tests.

The first two are the stock motor full run and the second is zoomed in on the full throttle portion of that run. The variations in the top magenta line seen there show 20W or so excursions in the power level (varying from 70W +/- to 90W +/- or so). That is unusual amount of variation, even for these motors, but it is putting out about it's specified 85 Watts or so and ran smoothly enough. Most buyers would find it a "good motor" and be happy with it.

The last two images are the same motor after rewinding. Again, the first is the full run and the second is 5 +/- seconds of the full throttle portion of the run. Note the significant improvement in the smoothness of the power delivery, that is about as smooth as these motors get.

The RPM excursions seen in the rewind test were due to a poor connector in my static testing setup, they are not in the motor.

Jack

^ Except warthox has half a million copters with equally many reasonably priced motors and no trouble until now.

OK, so I guess the calibration and programming in the controllers is taking care of the differences in the motors OK.

Jack

I have three questions:

1) does this concern "only" the hw-ESC's ore th hw-compatible too?

2) I have hw-compatible Suppo 25A ESC's from Flyduino (and some strange Problems with my copter). Does anyone know wheather Suppo-ESC's are concerned?

3) what ESC's would you suggest to buy now?

i get some suppo 25A hw compatible soon. i dont expect problems due they use atmel chips and so they should also use the 'old' fw. will report.

Following this. I was starting to think I was just unlucky in getting some sketchy Hobbywing ESC's. I've been using the 18's and have experienced exactly what you've described here warthox. I'll pull the heatshrink off of some of my older/newer ones and check.

ok, so I purchased a brand new flyfun and swapped all my flyfuns (V4.0) to my makeshift quad frame, which was flying Fentium V3.0s just fine, actually great considering the frame has so much flex. With these new flyfunss,I can barely hold a hover and it wobbles a lot or makes big drops when applying mild to aggressive stick movements. In otherwords, it flies like crap.

UPDATE: ok, so I set nearly all my PID variables to zero (ki,kp,kd) and it can hold a fairly decent hover, so I suspect some filtering is happening in the ESCs. I'll try to retune everything, but somewhat surprised that I would need to going from V3.0 to V4.0 fw.

On my Y6, I matched all my Fentiums (v3.0s) and balanced the motor layout (3 850kV jdrones on top, 3 850kV rctimers on bottom) and it flies straight and hover is now good (still need to tune more though)-- just a bit of jitter in hover likely due to jackerbes description of having mismatched motors. Sigh. I hate these SiLab chips...

Originally Posted by recharged95

ok, so I purchased a brand new flyfun and swapped all my flyfuns (V4.0) to my makeshift quad frame, which was flying Fentium V3.0s just fine, actually great considering the frame has so much flex. With these new flyfunss,I can barely hold a hover and it wobbles a lot or makes big drops when applying mild to aggressive stick movements. In otherwords, it flies like crap.... what a waste of time!

On my Y6, I matched all my Fentiums (v3.0s) and balanced the motor layout (3 850kV jdrones on top, 3 850kV rctimers on bottom) and it flies straight and hover is now good (still need to tune more though)-- just a bit of jitter in hover likely due to jackerbes description of having mismatched motors. Sigh. I hate these SiLab chips...

So without oscilloscope validation, so far I can attest there's something wrong with the V4.0 and up ESCs.

Yes bummer i have 7 esc's lying around what a waste

crap!

i have 6 of them in my new Y-6 build.......

is the jitters noticeable without a occiliscope?

oh yea