Jive 80HV SBEC Data and Why Torioids Are Used

[EEngineer]

Tested a Jive 80 HV's dual-lead SBEC for its SBEC voltage undershoot(VUS) characteristics...and in the process of doing so....can see why on the Kontronik
SBEC design, their SBEC cables come with their toroids.

My test procedure was such that...with no motor connected....to simulate an SBEC load, the "idle current" was established at ~1A.....and a transient current load of ~5A was applied using high-wattage resistors.

The results of my tests are displayed in Image #1.

As can be seen, the VUS of a Jive 80 HV's SBEC is a mere 320mV.
While the voltage "recovery time" is ~90 microseconds, the magnitude(or lack therof) of the VUS is minimal....and, IMO, is the best that I've ever tested.....

Now, with regards to why toroids are necessary, I refer to Images #2 & 3.

Image #2 was without any toroids....while Image #3 was with the Kontronik toroids leads.

During the test, I first identified the SBEC's switching freq. = ~ 179 kHz, as can be seen in both I#2 &3....

For some reason, there's also a spike at 250 kHz....

Also, at 10.67 kHz, there is a spike....complete with its associated harmonics...at ~ 21.38 kHz....and at 32.08 kHz.....with these associated amplitudes:

179 kHz = ~ -50.4 dBV
250 kHz = ~ -68.7 dBV

10.67 kHz = ~ -62.5 dBV
21.38 kHz = ~ -66.9 dBV
32.08 kHz = ~ -69.15 dBV

Without the toroids.
------------------------------

Repeating the exact same test, this time with Kontronik SBEC toroid leads, Image #3 shows the results:

179 kHz = ~ -82.4 dBV
250 kHz = ~ -85.9 dBV

10.67 kHz = ~ - 84.3 dBV
21.38 kHz = ~ - 91.9 dBV
32.08 kHx = ~ can't see it

With Kontronik toroid SBEC leads.
------------------------------

My point is to demonstrate that such toroids have nothing to do with radio interference at 2.4 gHz...but are used to "snub" electrical noise within the passbands of your FBL's MEMS triple-axis gyros.

Each of the 3 gyros use electronic oscillations to perform their task properly, and if there's external interference above a certain threshold, this external interference can cause havoc with 1 or more of your FBL's MEMS gyro axis's.

Kontronik SBEC's, whose VUS specs are the benchmark of an SBEC, use the toroids to "snub" any such external interference to your particular FBL controller, no matter who makes it.

Other ESCs might not need toroids, but it's well advised to use them with a Kon.

FWIW

[CZorzella]

Excellent work! Thanks for sharing it.

[mkovalcson]

Originally Posted by CZorzella

Excellent work! Thanks for sharing it.

+1

[Groucho_]

Originally Posted by EEngineer

Each of the 3 gyros use electronic oscillations to perform their task properly, and if there's external interference above a certain threshold, this external interference can cause havoc with 1 or more of your FBL's MEMS gyro axis's.

Is there any data as to what this threshold is?

Maybe BEC leads with no toroid cores still give a result that falls well under this value... Or maybe it's dangerously close...

I'm curious to know...

[dl7uae]

There are no numbers which can be named. All depends on too many factors.

Devices applying gyros:
- At which frequency is a gyro oscillating?
- How is the setup of the filtering behind a gyro? (Wouldn't even help in case of a "rectifying effect".)
- How about the filtering capabilities of the voltage supply path in that device?

And then the source itself, SBEC:
We do not speak about sinusoidal waves.
The original waveform feeding the gates of the switching FETs is almost rectangular.
But then many components come into play: The FETs itself (non-linear), the load circuitry, inductance + caps, a Schottky diode (non-linear), L/C filters, and even parasitic elements.

In the end is a crazy waveform, also load-dependent, theoretically composed of many sinusoidal waves of different frequencies.
A spectrum analysis shows these frequencies (hot spots) but does not show why/where from.
A scope shows the resulting wave form plus phase jumps, but not directly the "hot spot frequencies" involved.

All we can do is to attenuate as good as possible with reasonable effort. Toroidal core.

There is no chance to estimate when you might omit this filter action.

(I apologize for the screenshots. Did it "on the fly" in the tohuwabohu (seems to be a German word) of my table. So I just killed the next JLog, naked PCBs everywhere.)

Some load, no toroid:


Crazy wave forms --> many frequencies:


Some load, WITH toroid: Overall attenuated.Still something there, phase jumping, but potential problems reduced to (hopefully) zero.

[EEngineer]

Originally Posted by Groucho_

Is there any data as to what this threshold is?

Maybe BEC leads with no toroid cores still give a result that falls well under this value... Or maybe it's dangerously close...

I'm curious to know...

There is no spec on that.

Invensense makes a wide variety of MEMS gyros

The Blue MiniV FBL controller uses three Invensense ISZ500 single axis gyros.

Here's the spec sheet:

http://invensense.com/mems/gyro/documents/PS-ISZ-0500B.pdf

Supported, but not recommended for new designs.

Some of the answers to posed to dl7 can be found here, but not all.

This MEMS gyro has a frequency passband from 26 to 34 KHz.

Section 8.2.1 has some info on the ISZ500's power supply rejection ratio, and mentions that the PSRR does't apply to the resonant frequency range of the gyro.

Section 8.2.12 mentions that acoustic noise is not a problem, except at the above frequency range.....this is what the double-sided tape that Vstabi uses to prevent
this from occuring, by damping vibes in this range.

As dl7 mentions, there are so many variables.

With a Jive 80 HV, and perhaps the Kosmik 160/200 HV(haven't checked them), if I can see the issue on my spectrum analyzer using no toroids.....and can see the results of using toroids lessens what has been described.....given that Kontroniks insists that they be used........

I'd be foolish not to use them....

FWIW

[taosgraveyard]

Very interesting

In general I use the rings

I think I might one or two helicopters without them

If so I am putting them on.

Thanks for this information

[Seasick78]

Can anyone clarify how many loops around the ring are necessary for effectivity?

Also, is it correct (as is usually described) that it is best to place it as close to the RX or FBL system as possible?

[Wh1teAfr1can]

Originally Posted by Seasick78

Can anyone clarify how many loops around the ring are necessary for effectivity?

Also, is it correct (as is usually described) that it is best to place it as close to the RX or FBL system as possible?

5 loops is the standard.

you always want close to the ESC. because any wire before the toroidal core will be emitting rf interference.

[Danny Dugger]

I am also interested to hear the answer to the question above. I would think placing the core closer to the FBL will do better at reducing EMI or noise from other electronics, but probably would not make much difference regarding the noise from the ESC itself.

[dl7uae]

RF interference is not the issue here.

As many loops as possible. So what..

[Wh1teAfr1can]

Originally Posted by dl7uae

RF interference is not the issue here.

As many loops as possible. So what..

Yeah but any reduction in noise is a benefit if you have the bec wires routed around other electronics.

[dl7uae]

Yes, theoretically, and I also w'd place the core as close as possible to the source. Just out of principle.

But on the other hand with 2g4 as base band the only path of "injection" is by wire, not "through air". The frequencies and the amplitudes we are speaking about are the enemy of only one type of component: gyros. (At least we assume and hope that no other component or circuitry will feel interferes. )

I think I still know what I'm speaking about after already >40 yrs of that. (Although the stupid idea to misunderstand a 2nd level hobby (model helis), to "make in electronics", this stole the main hobby the last bit of time. Damn!)

[EEngineer]

With regards to the placement of the toroid, I maintain that it should be closest to the source....the Kon ESC....where the signals eminate from....I have always needed an extension lead to get to my VBar FBL....and then the toroid is closer to the Kon.

In addition, all toroids are not alike...and the Kon toroids have been selected to be effective for this issue at this particular frequency range....and this issue is, that any electrical interference generated by anyone's SBEC....that falls within the range of whatever your FBL's gyro's passband range is, is going to cause issues.

Torroids, when you spec. them, all have a frequency range...for you DIY guys, you will want a 20 to 40 Khz toroid....which of that physical size, it's difficult to have more than 5 turns.

And, as can be seen in the spectrum analysis pics, the 10.67 khz signal and it's harmonics were
effectively attenuated by using the Kon. toroids(with 5 turns) so as to minimize interference to the FBL's gyros to the extent possible.

The point is......the toroids have nothing to do with 2.4Ghz R/F interference....NOTHING....the toroids are used to reject interference with your FBL controller's MEMS gyros...at frequency ranges of ~20 to 40 khz...
and not at 2.4 Ghz.

When you mod, test your results so that it can be determined if it did any good....or worse, that it made a bad situation worse.



FWIW

[Danny Dugger]

Good stuff, thanks EE and dl7

[pc3associate]

But what would the difference be with the wires in a tight braid without the ring?

[EEngineer]

Originally Posted by pc3associate

But what would the difference be with the wires in a tight braid without the ring?

Noise within the passband of the MEMS gyros.

[pc3associate]

Could you exlain? I have a robot with gyros for a few arms. Was having a lot of problems with jitters etc, braided all the wires in the area and all jitters were gone.

[EEngineer]

Originally Posted by pc3associate

Could you exlain? I have a robot with gyros for a few arms. Was having a lot of problems with jitters etc, braided all the wires in the area and all jitters were gone.

pc, with all due repect, in Post #1, I posted spectrum analyzer images...of how the Kon ferrite toroids work and why they are necessary.

Your "robot" is a completely different "can of worms".....
You have to ask yourself this....would you trust your life to such a "fix".

FWIW

[AndrewJHawes]

If one is a little inebriated and look at Tom's waveforms, you can almost imagine a PQRST cardiac waveform, and believe the esc is ALIVE!
Sorry for the slightly facetious response. Good work though guys!
Andy