Measure spike current in 3 phase motor

[bootloader9800]

Hello Folks! Hope y'all doin great. Newbie here!! Excited to have founfd this forum.

So I got this 3 phase RC motor that is run by an ESC. The entire system runs on a 12V battery.
I would like to figure out the max SPIKE current going through each phase.

I need to know the spike current because my MOSFETs on the ESC keep burning up on me.

I have an OScope and a multimeter handy.

Any idea how I can find the spike current?

Thx 4 ur replies!

[CaptDon]

You could use a very small resistance, in the milliohm range like maybe 10 milliohms (not megaohms) and measure the voltage drop when you first apply full power to the motor from a dead stop. This resistor would go in series with one motor lead. The biggest problem is that these resistors aren't normally stocked. We made these resistors by using a normal piece of 14 gauge wire of perhaps a foot long and then using a variable current source we created a calibration table measuring the voltage drop across that short piece of wire. You should solder your 'taps' directly to the foot long piece of wire so they never change the exact resistance of the wire that they see. Now an example. Use ohms law here, if you put exactly 1 amp through the wire and your wire happens to be 8.33 milliohms you will see 8.33 millivolts at your taps. Now if you up the current to 2 amps you would measure 16.66 millivolts or .01666 volts. If your wire was 20 milliohms at the taps then 1 amp would give you 20 millivolts and 2 amps would give you 40 millivolts. Do you see the pattern here? You would connect your motor directly to the heavy part of the wire and connect your DMM or scope to the taps you soldered to the foot long piece of wire. We sense across a very tiny resistance so we don't create a loss in your motor circuit and the scope would be the best way to capture a true 'peak'. Set your scope for 'single sweep' and do a bunch of individual captures. You will have to play around with the time/div horizontal sweep speed to get the most detailed image of you 'peak'. Also play with the trigger level. What you created with the foot long piece of wire could probably be best described as a resistor with kelvin taps for the measuring points. This gives an extremely accurate way to measure voltage across very low resistances.

[Siwastaja]

The only reason the MOSFETs are burning is because the ESC is pure shit, broken by design.

The ESC is literally the one which controls the motor current, so yes, it's possible it generates current higher than it can take but there is no way to fix it except to redesign it... I.e., design an ESC. Basics; but non-trivial.

I would suggest try another brand of ESC, look for user testimonials before buying.

If you still want to measure the current, I'd suggest look at the shunts and current sense amplifiers on the ESC, and tap to these internal measurements. Oh, it's possible they don't measure current, some do not. Which is exactly why they blow up.

But it will be a futile task trying to measure the current and get the exact moment stored when the MOSFET blows up. Normally this is done in reverse; the designer first designs the current control circuitry and 100% verifies its functionality before running any actual current through it. This way, no MOSFETs are blown, and everything's easier. For example, I tend to inject voltage across the shunt resistor directly and look the control signals on the oscillosscope, before connecting the load when running the prototype for the first time.

But the problem with some Chinese modules is, they may not be even on a pre-prototype level!

[WattsThat]

As already has been mentioned, you’re dealing with a bad design. Even if you are successful in measuring the peak current, what do you intend to do with the information? Do you have full schematics and if it has a processor as I would expect it does,  the source code?

Seems to me you should invest the time and energy in a new controller.

[CaptDon]

If it truly is a three phase motor it will have three leads. If each leg of the ESC is a true bridge there should be 6 MOSFET's. Assuming they are typical mosfets they require a solid 10 volts on the gate to fully saturate and the drive needs to be a very brisk off to on and on to off transition to avoid time spent in the linear region. If the drive is ample you could probably substitute a better/bigger mosfet although gate capacitance may become an issue. You may also discover most chinese made trash is exactly that!!! Buy better!!! U.S.A., the new landfill for chinese made trash!!! The unit may possibly use logic level fets in which case the gate drive would only need to be around 5 volts for full saturation, but LLFET's are easily damaged and not generally found in good quality high current designs.

[Benta]

The OP said RC motor.
These are to 99.999% sensorless BLDC motors for RC planes or speedboats.
The ESC is very likely trash, good ESCs are not cheap.

[bootloader9800]

This is the OP.
CORRECTION: The motor is a BLDC motor. I mentioned in my original post that it was a 3 phase motor

Here is what I did till now. I set up the test according to the diagram below.

I would like to figure out the "Max spike current" going into the motor.

Thank you

[CaptDon]

Looks like a good plan to me.

[Benta]

Ain't no such thing as "max spike current".
You probably mean max peak current. This is dependent on the mechanical load on the motor shaft.

[Siwastaja]

It is a 3-phase BLDC motor. Similar types can be called synchronous motors, or permanent magnet AC motors. Exact winding shape defines whether it's called "BLDC" or "PMSM" but this is just details and doesn't matter much here.

You should measure the currents of at least two phases simultaneously to get the complete image of all phase currents.

What you show works but you need to deal with either low-value shunt, and low voltage over it, likely swamped by noise; or relatively large-value shunt, and a lot of power dissipation, and intrusive measurement (ESC may work differently because of the resistor).

I'd recommend a specific current sense amplifier coupled right next to the shunt with tight layout, and scope that output signal.

A proper ESC itself has such current measurement, so it does not produce too much current. Obviously, this is some shit ESC and a current measurement alone won't give you much information.

There are basically the following typical causes of failure:
1) Overcurrent combined with the lack of current sensing/control,
2) Overvoltage due to poor rating of parts, or poor layout (unsnubbed switch node ringing), exceeding Vds rating
3) Failure to drive the gate properly under all conditions, causing operation in linear range and quick, excessive heating

All of the above can happen at microsecond timescale;

4) Lack of proper thermal design; longer time-scale (tens of seconds to minutes) overheating and finally destruction.

You can kinda estimate the cause #1 with the measurement you propose.

Generally, in a system which provides a lot of inductance (the motor windings!), current changes slowly (no "spikes"), but voltage may change radically and quickly (voltage spikes). This is why controlling the current in a motor drive is almost trivially easy (a few dozen kHz of control bandwidth does fine, just measure and feedback, even a badly tuned PI loop often does fine), and which makes failing to do that properly very awkward.

[den]

Shunt resistor is too small.
Not sure if the way it is connected it will work correctly since scope is grounded, but DUT is floating.