9 & 15 MOSFET Controllers, WHY ?

[CompuTutor]

With most common E-Bike brushless motors,
wired in a standard three-wire Delta format,
the controller needed has defined requirements...

Each wire having one MOSFET to sink one motor lead to ground,
the other MOSFET, to source (the same) motor wire to a voltage.

Three of these MOSFET Half-Bridge pairs are needed,
for the three motor wires supplied from the motor.

So we need (at least) six MOSFET's minimum here,
or a mutiple of 6 (12, 18, Etc.) if more power's needed.


So why do many controllers also exist
with an ODD number of MOSFET's then ?

Let's cite the popular "9 MOSFET" version as example.

It makes no sense to me, to double-up on either
the three MOSFET's that are sinking to ground,
or the three MOSFET's that are sourcing to power.

That would leave the other half,
of these half-bridge's imbalanced,
right ?



I've search the internet for an answer,
but have failed to find an explaination.

Where are these last three MOSFET's
being intruduced into the circuit please ?

Posting a schematic (as picture) would be
worth (the proverbial) thousand words...

Search for "9 MOSFET" (and others) yourself,
you'll see why I couldn't find the answer...

 

[amberwolf]

This thread
https://endless-sphere.com/forums/viewtopic.php?f=2&t=19642
with the last post of the first page detailing what loads are on each FET in a cycle to explain how it may help.

This post explains it in a different way
https://endless-sphere.com/forums/viewtopic.php?f=2&t=86477&p=1265144#p1265260

 

[CompuTutor]

Yes amberwolf,
that helps explain it all
Thank you.

 

[stancecoke]

Hm, I thought about that several times. I came to the conclusion, that 9 FETs make no sense for FOC controllers, as highside and lowside switching is completely symmetrical within one electric revolution. Maybe for six step commutation there is an advantage, where one phase is conducted to GND permanently, one phase is doing PWM to the high side and the third phase is floating during each of the six steps.

Kunteng uses different MOSFET types on high side than on low side. So perhaps it's just a matter of costs, to choose two cheap, bad ones on low side, where no PWM is happening (for six step), than one better but more expensive?!

regards
stancecoke

 

[CompuTutor]

Well..., yes, but no...

In Delta, all roads lead to Rome, so to speak.

Am I wrong to see that if A, then B shall be pulsed (PWM) with positive,
while the remaining C will incur two paths of current seeking ground ?

Then add the effects in inductance into this fray,
and it sure seems there would be moments when
the only path to ground would incur even more than
what one part sourced into A, then B, by the time
it sequentially overlaps at C, in a cascading fashion,
yes ?

Similar problems taming Radar's magnatron return path problems
earlier in my marine electronics technician journey reminds me of this,
When certain sets of harmonics would add up to more than expected.

Inductance will surprise us from time to time, reminding us.

 

[stancecoke]

As said, you have to differ between 6 step and FOC commutation.

Six step:

FOC:

FOC commutation uses center aligned PWM where high side and low side are complementary at every time. For zero current all MOSFET are driven with a dutycycle of 0.5

regards
stancecoke