Voltage Boost (and Z-source inverters) versus Flux Weakening

[bananu7]

There have been a few posts about this on ES, namely at least:

INCREASE EBIKE SPEED WITH DC BOOST CONVERTER

How about instead of adding more cells, we just buy a 1800W boost converter. A boost converter is basically a transformer for DC power...... That is adjustable to a fixed output voltage. Plan on using one of these boys to step up my 48v 20ah battery to 72v.... Like here...

endless-sphere.com

I think it really belongs in motor technology, and the more I learn about PMSM motor control, the more interesting rabbit holes I get into. This topic is in fact much older, and I found research dating as far back as 2002!

https://ieeexplore.ieee.org/document/998540?getft_integrator=wiley

https://ieeexplore.ieee.org/document/5544573?getft_integrator=wiley

The relevant point in the original abstract was:

The results show some advantages of the operation by the PWM inverter with voltage booster.

While I don't think that adding a boost converter in front of the controller is a great idea (save perhaps if it was bypassed normally and only used at high speeds, kinda like FW in the first place), if it was integrated in the controller, it certainly shows some potential as an alternative control method.

But then it gets even better. The newer paper from 2015 (https://ietresearch.onlinelibrary.wiley.com/doi/full/10.1049/iet-epa.2013.0291) had an even better idea. Folks who advocate for dual ebike motors, strap in:

When the drive system is operated below the rated speed, the dual motors are operated in parallel to share the total output torque. By using the proposed parallel dual-motor, the efficiency of the whole drive system can be effectively increased. On the other hand, when the drive system is operated beyond the rated speed, one motor is operated at a standstill and its stator inductances are used as three-phase boost inductances.

I am not sure if it's genius or absolute madness. But just imagine a 48V 2WD torque monster that turns into a 96V rear-wheel drive system at high speed... Of course to utilize the entire method proposed in the paper, you'd need much more, perhaps some clutching/etc to lock one of the motors in place (perhaps a bit easier to do with a mid drive combined with rear hub), but still...

What do you think?

 

[bananu7]

For completeness, another paper that specifically talks about boost in the controller:
"Flux-Weakening Control for Permanent-Magnet Synchronous Motors Based on Z-Source Inverters"

https://epublications.marquette.edu/cgi/viewcontent.cgi?params=/context/theses_open/article/1285/&path_info=Li_marquette_0116N_10875.pdf

 

[fechter]

That's interesting. I don't exactly understand how their Z-source inverter topology works, but it looks fairly simple.
In general, the losses in a boost converter are going to be significant so you'd only want to use that intermittently. More typical FOC field weakening will probably have lower losses and less hardware to deal with.

 

[bananu7]

That's interesting. I don't exactly understand how their Z-source inverter topology works, but it looks fairly simple.
In general, the losses in a boost converter are going to be significant so you'd only want to use that intermittently. More typical FOC field weakening will probably have lower losses and less hardware to deal with.

I am way too deep in the rabbit hole now. There's so many papers discussing all different variants of Z-source, "quasi"-Z-source, etc. etc., but _zero_ practical executions I was able to find. Everything is math and simulations. Either it's a pipe dream or a concept that's only nice in theory, or it's actually so good that the companies that have executed on it successfully keep it as a secret.

I agree that the concept looks quite approachable; there's quite a lot of overlap between a boost converter and a sine inverter, so combining them _seems_ like a natural thing to do, but there must be more practical considerations I'm not aware of.

 

[bananu7]

The original paper for the Z-source inverter is I think the most approachable, BTW:

Sci-Hub | Z-source inverter. IEEE Transactions on Industry Applications, 39(2), 504–510 | 10.1109/TIA.2003.808920

 

[marka-ee]

I used one of those 1800 watt DC to DC boost converters on a bike for quite a while. It really helped the top speed. Although one thing to be very cautious about is controllers that offer a regen feature because I blew out a controller when it tried to backfeed power. This is not a good thing with a DC-DC inverter involved in the middle. Of course it helps top speed, if that's the question. Obviously, when I was stuck at 36 volts, and I would have zero load on the wheel whatsoever, it reached a terminal velocity speed that it would not go any faster. This is because of that reverse EMF problem. The extra volts pumping it to 52 volts definitely helped the top speed, and that's the reason why I did it in the first place.

 

[bananu7]

one thing to be very cautious about is controllers that offer a regen feature because I blew out a controller when it tried to backfeed power. This is not a good thing with a DC-DC inverter involved in the middle.

That's another reason why putting the voltage boost stage inside of the controller makes sense. During braking, it could bypass that and simply push the rectified voltage back through the power cables.

 

[mxlemming]

The zsi is using the controlled shoot through state of the h bridge with series inductance in the supply bus to generate a higher bus voltage. There's two obvious huge issues here...
1) this is effectively giving the DC link a huge inductance (many probably hundreds of uH) where normally the design of power stages focuses on extreme inductance minimisation down to single digit nano Henry level.
2) there is an extra inductor and diode on the high and low side added to the circuit. This incurs extra losses. The h bridge still has to be rated for the maximum voltage across it which becomes the motor back emf at the max rpm.

A general realisation after spending sufficiently long in the field doing this practically is that it's better to put extra iron and copper into the motor and therefore decrease the resistance and gear ratio required rather than add it to cunning external circuitry.

 

[E-HP]

If the boost converter is more efficient than field weakening, and you only need the extra speed occasionally, rather than continuously, then it might be cool to have a controller with a wide voltage operating range, and a circuit to switch to higher voltage boost converter in when needed, and back out for normal riding. Also, that way when you need more than the current than the converter can handle for accelerating you do so without the converter, and then flip the switch to have the converter feed the controller when you’re up to speed.