measuring back emf

[solarbbq2003]

anyone know how to measure back emf using a multimeter ( if its at all possible), keen to see what values bmc 500watt motor will give while being used under load with different voltage supplies ( 48v upwards)

 

[swbluto]

I'd imagine you'd start rolling the motor(with controller detached) and stick the multimeter probes in the right "places"(I don't know what a typical brushless motor wiring looks like, but I suspect you'd stick it in one of the phase connections and ground, assuming that exists.) in the motor's connection. By recording the EMF and RPM, you can derive an important motor constant(Mainly K or Kv, although this wouldn't account for torque losses but it'd be approximate enough for "real motors", methinks.).

EDIT:

It appears there isn't a ground wire. I'll have to investigate this more.

 

[fechter]

The back EMF will almost match the supply voltage at no load. It is a linear function of RPM, so if you know what what the speed is at one voltage, you can use a RPM measurment to calculate it at other voltages/loads.

 

[swbluto]

The back EMF will almost match the supply voltage at no load. It is a linear function of RPM, so if you know what what the speed is at one voltage, you can use a RPM measurment to calculate it at other voltages/loads.

You are THUH MASTAH!

But, sometimes.... almost isn't good enough.

 

[fechter]

The back EMF will almost match the supply voltage at no load. It is a linear function of RPM, so if you know what what the speed is at one voltage, you can use a RPM measurment to calculate it at other voltages/loads.

You are THUH MASTAH!

But, sometimes.... almost isn't good enough.

OK, well if you can measure the resistance of the motor, then you can calculate the voltage drop at the no load current and add that to the supply voltage to get the actual motor back EMF.
My guess is the difference will be less than 5%.

 

[solarbbq2003]

ok i'll have a bit of a play with it as time permits, I realise that back emf is linked directly to rpm ( volt/rpm) but I think there is more to it than that, as many of us are aware that damaging mosfets is often linked to a bad connection on the motor phase wires, and that spark ( assuming back emf related) is dependent on load also, eg spin the motor at no load can disconnect the phase wires no problemo, do it while motor is under load and theres a good risk of blowing the fets.
So really I'm looking at determining worst case back emf values ( which i'm guessing are going to be also load dependent not just rpm dependent)

 

[fechter]

Well, when you consider breaking the connections or switching transients, then the voltages are more a function of controller design and not so motor dependent.

If the objective is to determine what rating your FETs need to survive, the BEMF of the motor doesn't matter much. It has more to do with how much energy gets stored in the windings and how good the main capacitors can the absorb pulses.

 

[ZapPat]

ok i'll have a bit of a play with it as time permits, I realise that back emf is linked directly to rpm ( volt/rpm) but I think there is more to it than that, as many of us are aware that damaging mosfets is often linked to a bad connection on the motor phase wires, and that spark ( assuming back emf related) is dependent on load also, eg spin the motor at no load can disconnect the phase wires no problemo, do it while motor is under load and theres a good risk of blowing the fets.
So really I'm looking at determining worst case back emf values ( which i'm guessing are going to be also load dependent not just rpm dependent)

I dont think back EMF is what is killing your FETs during a phase wire breakage, solarbbq. Real back EMF could only get high enough to fry your FETs if you somehow managed to turn your wheel quite a bit faster than it's no-load speed. Now this is kind of difficult to achieve since the FETs body diodes will force regen to happen as soon as the motor BEMF goes over the batteries voltage (plus diode drops). And as Fechter points out, real BEMF is RPM dependent, since it is directly proportinnal to coil/magnet speed. And of course as any voltage source, we have resistance since it's not a super-conductor, which drops the apparent BEMF when current circulates. Just as a battery would, etc.

So I think your real problem is the current circulating in the motor winding inductance while riding at higher currents and high-ish speeds. If one of the driving phases is disconnected while active, the stored magnetic field in the motor causes a high voltage spike to be made across the wire gap, adding to the motor's own BEMF, and I would guess this voltage would appear across the FET too. BLAM!

I've actually never had this happen to me, but it would be interesting to measure and record on the scope such an event... Maybe in my next test mood I could try it.

I think my reasonning here is right...

 

[solarbbq2003]

yes that would explain why backemf is linear with rpm, and no link to load, I must admit I dont understand enough about motor physics to know exactly whats going on, but your explanation makes sense. I guess there must be some difference between a mosfet switching off and disconnecting a motor wire while under load. Starts to get a bit complicated. When a fet turns off ( current stops flowing) field collapses and back emf produced, then will have inductance also at same time, damn it gets complicated!!
Lets say go with fechters suggestion of no load rpm gives back emf approx same as battery voltage.
Then for bmc 350watt ( chinese version ) motor:
No load rpm at 56.5v is 2394rpm
but what to do with those numbers not sure where to go next, because that will be max motor rpm but wont be max back emf value,

 

[solarbbq2003]

just read your last post this thread fechter, think I'm starting to follow this, I had assumed that backemf could be as high as twice the voltage the motor is running on ( not sure where i picked that up from), that was my concern for fets, hence interest in getting backemf values,