KDE XF esc's

[KDE Direct]

Interesting, do you see this same effect if you run the motor direct off of a RX?

Wonder if the SuperX is causing an intermittent delay, that would cause an effect to be seen at the ESC? The ESC is simply reading the signal and responding instantaneously (at least, as fast as it can read at 600Hz), so it needs to down-sample to the speed the SuperX is communicating on (400Hz?).

I'll do some testing on our side to see if the same effect occurs. The DJI systems also run at 400Hz, but don't seem to show the same effect and the down-sampling is smooth. I wonder if SuperX can increase their response rate and latency to match the XF series, and prevent this from occurring? Any chance you have a contact there I can talk with?

 

[kloner]

I'll have to rig something to check throttle off the rx.... windows thing again....

i'm not sure what the communicating rate is, it has a mode for a different esc mode, might want to get ahold of Drew at xaircraft america and ask specifics

the video is active

 

[fltundra]

Kloner, have you tried changing the esc setting in the SuperX?
333hz standard.
View attachment 18794

 

[kloner]

turned that on and the esc's quit arming, acted like they do when zero throttle signal isn't available. beep, beep, beep

 

[kloner]

Got this esc hooked straight up to the rx and it still does it. same screech, it's easy to get it into it....

 

[KDE Direct]

Hey Steve,

Ran the bench tests on the U7-420Kv editions and no issues, so sounds like the compatibility may be centralized to just the 490Kv editions that you are running(?).

I will post a video showing the bench tests, so you can see the differences in ESC optimizations when it's running the U7 in comparison to the upcoming KDE5215XF-435, it's night-and-day in real-life.

 

[Old Man]

Kloner, have you tried changing the esc setting in the SuperX?
333hz standard.
View attachment 23098

That default 333Hz is what stopped me from buying a SuperX.

 

[fltundra]

That default 333Hz is what stopped me from buying a SuperX.

Never really worried about it after reading this and speaking with Castle when i was doing my research.

All this seems to imply that the higher the refresh rate, the better... without bounds.

Consider a quadcopter in level stabilising mode, the flight controller is making small adjustments in throttle to each motor to keep the quadcopter level. It is making those adjustments based on input from gyros and accelerometers, and you might thing the faster that the control loop responds, the better, and so a faster throttle response in the ESC implied by higher refresh rate is naturally better... without limit.

There is a practical limit to the rate at which the quadcopter can rotate in any axis, a limit imposed by available power and stability of the control system. A high refresh rate helps in the goal of a stable control loop with rapid response. In fact, in some flight controllers the maximum slew rate is a user tuneable parameter of the flight controller algorithms, so you might choose a lower value that the available power supports.

Problem is that the gyros and accellerometers are detecting not only small attitude changes in the platform, but vibration in the frame induced principally by out of balance motor / propeller combinations and exacerbated by frame resonances. This latter signal component is known as noise, it is undesirable components of the total sensor signal. A two blade propeller spinning at 12000rpm creates fundamental vibration component at 200Hz, but there may be significant harmonic component.

So, is there real benefit in extending the frequency response of the control loop (by increased throttle servo refresh rate in the system) when most of the higher frequency component is due to noise.

The situation you may have is that there is rapid adjustment of throttle around the desired target caused principally by noise, by vibration, and adjusting the motor speed will not eliminate that vibration, it just consumes more power in the rapid small variation in motor speed.

In my own experience with MultiWii, I select a low pass filter option of 98Hz to improve stability by filtering noise. The quads are still highly responsive, much more responsive with SimonK than any of the stock FW that I have tried, but with a 100Hz LPF, I don't need 600Hz servo refresh, 200Hz is adequate... but I fly with the MultiWii default 50Hz refresh rate.

So, I would not obsess over different claims of 480Hz vs 600Hz, there are other more important factors which might differentiate the ESCs.

Owen

 

[KDE Direct]

They say a picture is worth a thousand words, so I thought a video would be even better.

Here's the best way to show the performance optimizations of the KDE Direct XF UAS ESC/Motor combination, and how it stacks up against the competition in instantaneous, high-speed flight-response technology.

Also - no commutation errors are witnessed between the KDE Direct XF UAS 55A+ and T-Motor U7 420Kv edition, but from Steve's testing, the 490Kv edition may be posing concerns (?). We'll continue testing this week, but overall, the XF UAS series are optimal and specifically-tuned to the KDE Direct XF Brushless Motor series and are the recommended combination for plug-and-play compatibility.

 

[Old Man]

Hoping not to appear ignorant here, but what's the possibility part of the problem is the ramp up and down speed of larger props combined with slower motors, meaning inertia issues? The next question pertains to the U8-170 using a 28" prop and what's available to obtain similar response match up in the larger class system?

 

[tahoejmfc]

Here are my results with a U7 420 Kv 17" Prop

Maytech - Garbage, Sync errors all day long
Castle Phoenix Edge lite 75A - 1 minute at 60% throttle 106f, 2 minutes 130f - I can tell these act alittle slow in response
Castle Phoenix Edge HV 50A - 1 minute at 60% throttle 97f, 2 minutes 109f - I can tell these act alittle slow in response
KDE UAS 55+ - 1 minute at 60% throttle 114f, 2 minutes 155f - super fast response, seem like best option so far, no sync errors
Castle multipack 35A - 1 minute 60% throttle 160f, 2 minutes 213f - right away you can tell the motors are not getting the power they need

 

[Motopreserve]

Thanks so much for the solid testing and data Tahoe (oops)

 

[fltundra]

Here are my results with a U7 420 Kv 17" Prop

Maytech - Garbage, Sync errors all day long
Castle Phoenix Edge lite 75A - 1 minute at 60% throttle 106f, 2 minutes 130f - I can tell these act alittle slow in response
Castle Phoenix Edge HV 50A - 1 minute at 60% throttle 97f, 2 minutes 109f - I can tell these act alittle slow in response
KDE UAS 55+ - 1 minute at 60% throttle 114f, 2 minutes 155f - super fast response, seem like best option so far, no sync errors
Castle multipack 35A - 1 minute 60% throttle 160f, 2 minutes 213f - right away you can tell the motors are not getting the power they need

What voltage were you running? 16 amps at 65% on a 6S is not a lot for the listed esc's,
View attachment 18797

 

[tahoejmfc]

6S of course. I know that they don't list a high amperage with that motor, but having some safe overhead is smart. I tried a 35A esc and it got smoking hot ever though the motor is using only 16 amps

 

[kloner]

was curious after hearing the big amp loads part on these motors so tried 16", still locked up but at 15" it does not....

 

[tahoejmfc]

todays tests with KDE 55+ and T-motors U7 420Kv


Had to add a 2.5cm fan to the KDE to keep the temps down around 120f


Juicy spots are 0:14, 0:25, 3:50 is at about 50% throttle and make weird sound, 6:08,

<iframe src="//player.vimeo.com/video/98884513" width="500" height="281" frameborder="0" webkitallowfullscreen="" mozallowfullscreen="" allowfullscreen=""></iframe> T-MotorsU7 with KDE UAS 55A ESC from Cinematic Digital Media on Vimeo.

 

[kloner]

that's exactly what mine do

 

[Old Man]

Interesting about the need for a fan.

I don't have any experience with the U7 motors, only the U8-135. However, I experienced a very similar screeching sound at start up during testing with a U8 with Tiger 30 and Castle 50 ESC's using 28" T-Motor prop. That sound might not be generated by the ESC but from something in the motor that isn't quite right. Just a thought.

 

[fltundra]

I think the motors are the problem, not the esc's. Maybe internal resistance increases drastically with temperature or just a bad design.

 

[fltundra]

I ran my tee knee tiny castle 25's last nite at 11 amps continuously 83% on the FC, and never saw more then 95 degrees with my 515kv's on a 4S.