openinverter forum

Hi Guys,

Has anyone tried using an OI with a Hyper 9 motor?

I am designing a trike and would like to use a Hyper 9 direct coupled to a 3.45 diff but i have a dilema.
I want an AC motor for regen and easy (neat) reversing, but I also want to run a high voltage pack, say 280-300V so that I can utilise DC charging via CCS or Chademo. I understand this has a lower voltage limit where a 144V pack is likely to be a problem for most chargers.

I have a Transwarp 9 already but the Hyper 9 will allow me to delete ugly reversing contactors and fit more batteries due to shorter length etc. Not to mention gaining regen.

I'm not sure if the Hyper 9 is primarily an IPM motor that also has SR characteristics or the other way round and whether this will work with OI.

Do you think it would be possible to run a 300v pack but limit the max volts to the motor to stay within its limits?
Are the Hyper9 and Hyper9HV internally the same and only controllers different or are there stator differences?

Im guessing required torque and rpm will govern what AC voltage will be applied to the motor anyway?
My calcs look like about 3500 motor rpm at 100kmh. Planning on small diameter rear tyres to keep rpm as high as i can but this is about as good as it gets with a 3.45.

Thanks.

The Hyper9 is rated for DC voltages up to 130V and the Hyper9 HV for up to 180V.
That means it will have quite low inductance so the control loop might go unstable at higher voltages than that. A way to counteract this somewhat is running a high PWM frequency. This rules out the Prius inverters. The Tesla M3 inverter with SiC FETs is designed for higher frequency, 17.6 kHz should be possible. But it is yet hard to hack, see Damien youtube channel.

All others, Leaf, Ampera, i3 will rather max out at 8.8 kHz

Apart from that it's a regular IPM that provides torque through permanent magnets and reluctance combined.

Awesome. Thanks for that information.

Is there a way to limit max AC motor volts while running a higher DC bus so that it doesnt go over 180V AC or am I missunderstanding how AC control works.
I wasn't thinking of running the motor at 280-300, just the battery.

I'm still stuck in DC pwm world where you can run say a 360V DC bus but only do a max 50% pwm to keep motor volts at 180v.

Am I right in thinking that the low inductance will affect the stability of the current control loop rather than voltage?
So current will rise too quickly to get control of?
This is probably still an issue in DC motor control too now that I think it through.

Thanks for your help on this.

johu wrote: ↑Sun Apr 16, 2023 5:23 amThe Hyper9 is rated for DC voltages up to 130V and the Hyper9 HV for up to 180V. That means it will have quite low inductance so the control loop might go unstable at higher voltages than that. A way to counteract this somewhat is running a high PWM frequency. This rules out the Prius inverters.

So, dumb question I should have asked a long time ago...

I'm planning on using a forklift AC motor (48v), and a Prius inverter. Am I in trouble? What voltage pack should I be designing for?

mikejh wrote: ↑Sun Apr 16, 2023 8:07 am Awesome. Thanks for that information.

Is there a way to limit max AC motor volts while running a higher DC bus so that it doesnt go over 180V AC or am I missunderstanding how AC control works.
I wasn't thinking of running the motor at 280-300, just the battery.
......

Well sine OI used to mplement a sort of buck - boost converter for the voltage. You could use 300Vdc from the battery and motor would see 180V all the time. Or if you have less voltage some boost would let motor know it has 180V from 150Vdc battery. This was usefull in winter when cells would sag but motor would retain torque...

Caution voltage on motor would still be 300Vdc, just PWM chopped...

MattsAwesomeStuff wrote: ↑Mon Apr 17, 2023 5:35 am I'm planning on using a forklift AC motor (48v), and a Prius inverter. Am I in trouble? What voltage pack should I be designing for?

Not a dumb question, but perhaps you are in trouble...

arber333 wrote: ↑Mon Apr 17, 2023 5:50 am Well sine OI used to mplement a sort of buck - boost converter for the voltage. You could use 300Vdc from the battery and motor would see 180V all the time.

Thanks Arber.
Yes I think this is what I was thinking. We have a PWM inverter so we can make the AC voltage anything we want providing the switching elements can withstand the DC bus voltage, and the processor has sufficient PWM resolution to enable 50% to be max?
But I guess the lower the inductance the closer the motor looks like a dead short to the power stage and that means fast current rise time @ 300V.
I need to learn more.... Just when you think you know how this stuff works....

Would it be worth using an F334 with the high res pwm module or is that just pushing IGBT's past their upper F limits?
The IXFN170N65X2 Mosfets look like a possibility. They're not SiC but I don't need 1200V or 1700V Mosfets for a 300V bus either.
Thinking they're a bit slow though (60ns TdON) and the gate C is 10 times higher than a similar IGBT.

Any reason why we can't achieve 17-20Khz pwm with an IGBT that says its good for 20-60Khz switching?

I'll do some more reading of this forum... Guessing Sine OI is a different firmware of the original OI?

Thanks.

mikejh wrote: ↑Mon Apr 17, 2023 1:53 pm Not a dumb question, but perhaps you are in trouble...



Thanks Arber.
....
I need to learn more.... Just when you think you know how this stuff works....
....
Any reason why we can't achieve 17-20Khz pwm with an IGBT that says its good for 20-60Khz switching?

Well i am still running Mazda to that firmware and it keeps the motor happy in winter when the battery sags.
My system dV change would be something on delta of 50V which is why motor feels comfortable at any doable voltage of my battery. If you go lower there are switching losses to consider.

Also 6kHz or 8kHz is the optimal choice for normal IGBTs. Maybe maximum 14KHz beyond that switching loses rise so high that inverter is heating up a lot more. That is a lot less robust that it should be.
Beyond that i would suggest you use small avalanche IGBTs with very low resistance. But that would mean experimental power section...

arber333 wrote: ↑Mon Apr 17, 2023 5:13 pmBeyond that i would suggest you use small avalanche IGBTs with very low resistance. But that would mean experimental power section...

Nope nope and nope. I'm a solder monkey that took 3 years to get a motor to spin.

Can you explain more about the practical effect of me trying to use the Prius inverter with my low voltage motor (or to a lesser degree, the Hyper 9), in terms of, what I as a driver would notice or be limited to, and how it would affect what the car can't do?

arber333 wrote: ↑Mon Apr 17, 2023 5:13 pm Also 6kHz or 8kHz is the optimal choice for normal IGBTs. Maybe maximum 14KHz beyond that switching loses rise so high that inverter is heating up a lot more. That is a lot less robust that it should be.
Beyond that i would suggest you use small avalanche IGBTs with very low resistance. But that would mean experimental power section...

Thanks Arber,
I'm not against an experimental power stage, just not sure I want to destroy $1000 - $1500 SiC mosfets often...
Yes I hear you on the switching loss issues. Maybe those 650V mosfets would be Ok even though they are slightly slower.

MattsAwesomeStuff wrote: ↑Tue Apr 18, 2023 1:00 am Can you explain more about the practical effect of me trying to use the Prius inverter with my low voltage motor (or to a lesser degree, the Hyper 9), in terms of, what I as a driver would notice or be limited to, and how it would affect what the car can't do?

Hopefully someone more knowledgeable than me will explain it but I am not sure it will be anything you will notice performance wise.
I think it will be more of a case of the inverter may not live...
If current control is unstable then the power stage will die quicker than the controller can reduce the demand current. Unless there is some robust hardware current limiting built in.
Best case would likely be erratic driveability in the higher rpm range? Not easy to control acceleration or regen? Please correct me if I'm wrong.

I'm not sure how far from original spec you can go with AC motors.
Older EV conversions with DC forklift motors (low voltage) worked OK at 144V because you actually want them to pull a heap of current. Current = torque. Like the racing guys that pushed 2000A through them and then switched to parallel at mid track to make it pull even more.
And you had no worries of demagnetizing the rotor with stupid current. Just melting the comm... Good times....

mikejh wrote: ↑Tue Apr 18, 2023 12:45 pmI think it will be more of a case of the inverter may not live...

Mhm. Well, that does sound like it would affect driving performance.

Unless there is some robust hardware current limiting built in.

I don't know how robust "robust" is, buy the Toyotas are generally quite forgiving about such things. Not sure they're forgiving at a microsecond scale though.

And you had no worries of demagnetizing the rotor with stupid current.

Well...

I've actually been mistaken this whole time. I was thinking the Hyper 9 was an ACIM. It's a synchronous permanent magnet motor. My forklift motor is an ACIM running on the SINE firmware. So, does that change prospects for me?

MattsAwesomeStuff wrote: ↑Tue Apr 18, 2023 1:48 pm
I've actually been mistaken this whole time. I was thinking the Hyper 9 was an ACIM. It's a synchronous permanent magnet motor. My forklift motor is an ACIM running on the SINE firmware. So, does that change prospects for me?

Not at all. Back in 2014 i was running the first code by Johannes that was able to V/Hz the weakening region so much into the 400Hz that is was able to run ACIM motor wound for 75Vac. Car was sluggish but it worked! I was even using a DIY power section with elcaps and 300A block IGBTs! And no protection!!! Of course i broke power sections at a rate 1 per year.
Of course i went and had it rewound for 185Vac and that complemented really good with 96S pack. You might try that too if you want speedy car.

Today i notice most OEMs have 600A IGBTs in their power sections and plastic caps. Very low precharge time is needed.

arber333 wrote: ↑Tue Apr 18, 2023 7:33 pmOf course i went and had it rewound for 185Vac and that complemented really good with 96S pack. You might try that too if you want speedy car.

Nope, definitely not interested in rewinding a motor. If this won't work, I'm throwing it all away and deciding between an Outlander rear with a DeDion, or just putting a gas engine back into it.

In fact, I'm only pursuing this path because it's the simplest and quickest way forward to a running vehicle. The motor is 265lb, which is more than the original engine even weighed, and capable of an order of magnitude more torque than the differential can handle without shaving its teeth.

...

I'm trying to avoid doing that thing where I keep asking the same question hoping for a different answer because I don't like the one I was given. But, I still don't understand in practical terms what the limitation is.

Here's my motor spec:



It's a 6-pole motor.

And, 4.5 years ago when I asked whether this would be suitable for EV use at highway speed, no one really identified frequency problems or I wouldn't have started down this path. Lots of that is me not knowing how to ask the right question the right way (still is).

What I figured is, it's rated for 1000 RPM at 48v, if I need it to rotate ~5000 RPM, I just give it more voltage, tell the controller higher frequency, and it'll go. There was moderate skepticism that it would handle 200-400v typical of an EV, but Major figured it would easily handle 10-12,000 RPM, so, that wasn't a problem.

I've only ever heard about voltage limitations for speed (too low and you won't get enough speed, too much back-emf), voltage limitations for arcing (mostly commutator issues, not an AC problem), current limitations (wires will melt), and RPM limitations (balance and/or core will eventually fly apart). This is the first time in 5 years that I've ever heard anyone mention anything about a frequency limitation of some kind, and I don't really grasp what or how that limitation is coming into play.

In order to get an induction motor that's made for 50hz to run faster you have to make that AC 100hz, 200hz, 400hz etc to get the higher rpm. Giving it more volts is a DC motor thing.
That's where the frequency comes into play.

Now you have to synthesise an AC sine wave at 400hz using a chopped DC (pwm) frequency at say 8khz.

I think there is a limit to the upper sine wave frequency for a given pwm frequency and inductance of the motor starts to affect it too.

You also have to have a current control loop that can run fast enough to keep the change in current under control.
Think that's where Johannes mentioned the low inductance might be a problem.

I read a great post from Pete in another section last night that explains it well. I'll see if I can find it and link here.

MattsAwesomeStuff wrote: ↑Wed Apr 19, 2023 5:13 am
I've only ever heard about voltage limitations for speed (too low and you won't get enough speed, too much back-emf), voltage limitations for arcing (mostly commutator issues, not an AC problem), current limitations (wires will melt), and RPM limitations (balance and/or core will eventually fly apart). This is the first time in 5 years that I've ever heard anyone mention anything about a frequency limitation of some kind, and I don't really grasp what or how that limitation is coming into play.

Frequency is generaly not a problem for motor, more for inverter. I think with current IGBTs you can get up to 400Hz in frequency using 14kHz carrier PWM. I tried to use 18kHz with Volt inverter to describe 800Hz for EMRAX motor and inverter started to overheat.

If your motor is wired for 50Hz then you can expect more heating losses if you go higher in Hz. This is why i suggested to rewind the motor.
But i think you can still use this motor with some 120Vdc and there would not be much consequences. Maybe you can rewise you intended battery from 96S to 32S3P to get 115Vdc? Just an idea...

The problem is i am not sure your Prius inverter drivers would work that low. I think that would merit a test?

Just curious - what is the attraction of the Hyper9 over Leaf or Outlander rear motors? They should produce similar power, are cheaper and have already successfully been used with OpenInverter?

MattsAwesomeStuff wrote: ↑Wed Apr 19, 2023 5:13 am What I figured is, it's rated for 1000 RPM at 48v, if I need it to rotate ~5000 RPM, I just give it more voltage, tell the controller higher frequency, and it'll go. There was moderate skepticism that it would handle 200-400v typical of an EV, but Major figured it would easily handle 10-12,000 RPM, so, that wasn't a problem.

I've only ever heard about voltage limitations for speed (too low and you won't get enough speed, too much back-emf), voltage limitations for arcing (mostly commutator issues, not an AC problem), current limitations (wires will melt), and RPM limitations (balance and/or core will eventually fly apart). This is the first time in 5 years that I've ever heard anyone mention anything about a frequency limitation of some kind, and I don't really grasp what or how that limitation is coming into play.

Can't comment on the RPM the motor can physically stand but the inverter should cope. On the voltage I think the main limiting factor is the winding insulation. You're increasing the voltage both in terms of the sinusoidal amplitude plus the flyback voltages that you get with inverter drive (if you look on machining forums there is a lot of debate about using inverters on older lathe/mill motors and whether non-inverter rated insulation is up to it - I've converted mine without any trouble though). I'd be temped to reduce the battery voltage as suggested by arber33. This should still work fine with the Gen2 but will reduce available power. If you parallel MG1 and MG2 could you get 500A? If so then 115V would give 57kW (I might be tempted to push a bit higher on voltage to get a little more power).

Edit - This is my take on how ACIMs work, might help? viewtopic.php?p=52155#p52155

On an AC induction motor, with the standard OI firmware (stm32_sine build), there are no current control loops so loop stability shouldn't be an issue.

On a PM motor low inductance is an issue. If considering a Hyper9 it would be worth asking the manufacturer what the winding inductances are (ideally Ld and Lq rather than phase inductance). Anything below a few tenths of a mH is going to be problematic.

Thanks Pete.

I read your post on page 12 of the IPM motor simulation and FOC SW thread last night (about PM motors).
Thanks for your effort in writing all this up. Your wealth of knowledge is impressive.
I think I can understand why you are saying the low inductance is an issue after reading that.

Not sure I understand when you say there are no current control loops. What is protecting the power stage from melt down?
Are you able to explain what will be problematic if it is not the current control loop, so that I don't go off on a tangent...?
I was thinking the individual PWM switching could cause current to rise too quickly with low L.

My reason for the question about the Hyper 9 is related to my design of a motorcycle trike. Originally I had planned to use a series wound Transwarp 9 because I have one, but they are quite long and takes up a lot of real estate that could be used for batteries.
Plus I'd really like to have regen and the ability to reverse easily without having to resort to reversing contactors.
But using the standard controller with the Hyper 9 limits me to 144V and I would also really like to be able to do CCS or Chademo charging (min 200-250V?).

Leaf and Outlander motors are not so easily available in Australia due to our so far low uptake of EV's, and not sure about packaging.
Leaf motor looks pretty small but my plan was to direct couple to a Commodore/Camaro style IRS diff at 3.45:1
The output shaft on the TransWarp9 fits the output coupler on the 6speed auto which bolts straight up to the guibo on the diff. See pic.
Obviously the Hyper 9 is not splined but I could machine up a coupling to match the 1 1/8 keyed shaft.

So my thoughts are mainly due to the off the shelf availability of the Hyper 9 rather than anything else.
Guess I could look at importing a Leaf motor too...
Hyper 9 vs Transwarp 9 gets me another 14 cells.

EDIT: So with the Hyper 9 I was hoping there was a way of running a 280 - 300V DC bus but limiting the AC output to 144V AC to the motor, but not sure if that is possible. That way the motor would still have it's expected voltage. But I realise that the stator will still "See" 300V pulses which might be an issue with the low L?

Thanks.

Thanks for that and sorry the simulation thread became such a long winded document. Really didn't know what I was doing at the start so it meanders quite a lot. FYI - there is a set of links on the first post viewtopic.php?p=44299#p44299 that provides a shortcut to the most useful stuff (including the one you mention) without having to wade through the rest!

mikejh wrote: ↑Wed Apr 19, 2023 2:05 pm Not sure I understand when you say there are no current control loops. What is protecting the power stage from melt down?
Are you able to explain what will be problematic if it is not the current control loop, so that I don't go off on a tangent...?
I was thinking the individual PWM switching could cause current to rise too quickly with low L.

Sorry again, wasn't too clear in the above post (will edit to clarify). It's just the AC induction motor control with OI that doesn't use current control loops (unless you use catphish's experimental code). The PM motor control does use the loops.

To protect against meltdown, on both AC induction and PM motors you either rely on built in protection (Prius Gen2 or Gen3 inverters) or the over current comparator on the OI control board (which monitors the current in hardware) to shut down the PWM if current exceeds the safe limits.

mikejh wrote: ↑Wed Apr 19, 2023 2:05 pm My reason for the question about the Hyper 9 is related to my design of a motorcycle trike. Originally I had planned to use a series wound Transwarp 9 because I have one, but they are quite long and takes up a lot of real estate that could be used for batteries.
Plus I'd really like to have regen and the ability to reverse easily without having to resort to reversing contactors.
But using the standard controller with the Hyper 9 limits me to 144V and I would also really like to be able to do CCS or Chademo charging (min 200-250V?).

That makes perfect sense. Very nice trike design BTW

mikejh wrote: ↑Wed Apr 19, 2023 2:05 pm Leaf and Outlander motors are not so easily available in Australia due to our so far low uptake of EV's, and not sure about packaging.
Leaf motor looks pretty small but my plan was to direct couple to a Commodore/Camaro style IRS diff at 3.45:1
The output shaft on the TransWarp9 fits the output coupler on the 6speed auto which bolts straight up to the guibo on the diff. See pic.
Obviously the Hyper 9 is not splined but I could machine up a coupling to match the 1 1/8 keyed shaft.

Might still be worth considering the Leaf or Outlander options, both are fairly compact. Details of the Outlander motor, including dimensions, are here https://openinverter.org/wiki/Mitsubish ... Drive_Unit. From memory the Outlander spec is 195Nm (although I'm hoping to get 250Nm out of it) and the Leaf is 280Nm (which might be better for driving the diff directly) so both are comparable to the Hyper9. Might also be worth contacting m.art.y on the forum (no idea whether he can ship stuff that far though?).

Edit - Leaf motor details here - https://openinverter.org/wiki/Nissan_leaf_motors, torque corrected above, some graphs here viewtopic.php?p=52750#p52750.

mikejh wrote: ↑Wed Apr 19, 2023 2:05 pm EDIT: So with the Hyper 9 I was hoping there was a way of running a 280 - 300V DC bus but limiting the AC output to 144V AC to the motor, but not sure if that is possible. That way the motor would still have it's expected voltage. But I realise that the stator will still "See" 300V pulses which might be an issue with the low L?

As long as the winding insulation in the motor can handle the full pack voltage and the inductance is reasonable I can't see why this approach wouldn't work. Really need a handle on the motor inductance to have any degree of confidence though.

mikejh wrote: ↑Wed Apr 19, 2023 6:07 amIn order to get an induction motor that's made for 50hz to run faster you have to make that AC 100hz, 200hz, 400hz etc to get the higher rpm.

Yes, I'm following you.

Now you have to synthesise an AC sine wave at 400hz using a chopped DC (pwm) frequency at say 8khz.

Right. So, the faster your frequency, the less... I dunno, "resolution" you have on the inverter side, to imitate a sine wave smoothly? Is that the problem?

8800 / 400 = 22. So in that example we have only 22 steps per cycle (or, half that, to hit both halves of the cycle?), and that starts getting too lumpy? And... things happen... when lumpy?

And, for example the OEM Prius motor would never need that high of frequency? Or, it doesn't count, because things are different there?

I think there is a limit to the upper sine wave frequency for a given pwm frequency and inductance of the motor starts to affect it too.

Mmm, yeah, I think you and me both hit the confidence wall here as to what happens next.

Something something, too fast, inductance, no worky no more.

...

arber333 wrote: ↑Wed Apr 19, 2023 7:43 amI think with current IGBTs you can get up to 400Hz in frequency using 14kHz carrier PWM.

This is where I've lost you. Why? I understand "inverter gets too hot", but... why?

Also, do we know that the 1000 RPM my motor gets is at 50hz? Is it linear to go faster?

The diff is a 3.44:1 ratio, I'll direct drive. With stock everything, 3500 rpm gives almost exactly 70 mph in an Opel GT. That's probably the upper end of where I'd drive it. Maybe I'd like to hit 80, but I'm not a teenager, I'm not street racing with it and I won't have the battery capacity to get home if I did.

So, 3500 RPM looks like a nice goal. Is it true to say that I need 3.5x 50hz = 175hz? (plus some extra, to account for slip?).

And, if so, is 175hz something a lot less intimidating to the Prius inverter and something it can handle?

But i think you can still use this motor with some 120Vdc and there would not be much consequences. Maybe you can rewise you intended battery from 96S to 32S3P to get 115Vdc? Just an idea...

Oh, I'm a special case, I've got way more flexibility than that.

Lower voltage is easier in every way. I was almost headed the other way, since I'm one of the very few who can have a small kWh but high voltage pack, as I'm making my own from 18650s. I'm not constrained by the minimum building block size like most other people are. I have thousands of cells, not just 96. I can pick any voltage I want.

It's why I was still half-considering the Toyota Highlander MGR rear end, since no one else has a 600v pack to ever run one in a car light and small enough that the MGR could still push it. I kinda want to get this motor working (as the "easy" option), and while driving, work on replacing the whole rear end and do a swap. So at least I get some joy of driving rather than endless featurecreep.

I was even thinking, since the Prius inverter can handle 600v, why not just give it 600v, or, maybe 300v now, and, in case I want more range but can't get more 18650s (my free supply went out of business 3 years ago), build them to an AH rating that could be supplemented by larger form factor cells in the future (Leaf or whatnot), continuing in series. I guess that's shot, and I should stop at 200v at most. Any reason not to just target the ~208v that the OEM Prius pack had?

The problem is i am not sure your Prius inverter drivers would work that low. I think that would merit a test?

Well I ran it at 12v. It spins. I don't know if that answer the question or if I'm missing something.

It needs ~190v minimum to power the DC/DC but, that's a tertiary concern (I'll just use the onboard boost converter to boost the 120v to 200v so that the DC/DC can buck it down to 12V! ... ). Or, umm, a just buy a DC/DC.

...

Pete9008 wrote: ↑Wed Apr 19, 2023 10:00 amCan't comment on the RPM the motor can physically stand but the inverter should cope.

Hurray. That was easy. Pete said so, no take-backs!

Just because no one else has mentioned yet and it might too be a game-changer... how bad is my efficiency going to be by overdriving this motor compared to its comfortable level of 50hz?

On the voltage I think the main limiting factor is the winding insulation.

Hmm... I'm less concerned about that. It's a fairly modern motor (2005-ish?) and I'd imagine all enameled wire is going to be thick enough to be "enough". I'd doubt they used special thinner enamel than they would've used for, say, I dunno, 600v industrial motors. It probably all comes off the same reel, right? Maybe?

Maybe I'll give everything a laquer spray to it all to discourage arcing.

If you parallel MG1 and MG2 could you get 500A? If so then 115V would give 57kW (I might be tempted to push a bit higher on voltage to get a little more power).

I don't think I can parallel MG1 and MG2. My board wasn't built for that. Unless I can just shove all the MG1 stuff in parallel with the MG2 wires? (MG1 shutdown, MG1 U/V/W phase signals, GFIV fault).

viewtopic.php?p=52155#p52155

When I read that, I understand it, but I know I do at a shallow level. I don't think I could apply it and correctly problem-solve anything.

...

Mike if I'm hijacking your thread too far let me know and I'll fork off to my own.

Thanks Pete,

I did see your table of contents in the first post. I'll have to work through them. A lot to learn there.

Pete9008 wrote: ↑Wed Apr 19, 2023 2:45 pm To protect against meltdown, on both AC induction and PM motors you either rely on built in protection (Prius Gen2 or Gen3 inverters) or the over current comparator on the OI control board (which monitors the current in hardware) to shut down the PWM if current exceeds the safe limits.

So does the AC controller still work as a current controller with a PI loop for torque control?
From my experience with DC controllers the throttle sets the demand current (torque) rather than speed (rpm). PI loop controls PWM to achieve the demand current. So hopefully the PI loop keeps current inside reasonable limits and the hardware over current protection only kicks in if the PI loop goes nuts or doesn't catch it quick enough.

Thanks for the compliment on the trike. Its a work in progress. The beauty of CAD is you can try hundreds of combinations before ever cutting any metal. Might have to try to model up a Leaf motor. You've certainly got me thinking about it.
Shame about the tiny output shaft. Still I can machine up a similar coupler and get the splines wire cut to suit.
Trikes only looking like about 350-400Kgs so I think these motors will be fine even with the lower gear reduction. If it's sluggish I can always go with 4.11 gears due to the higher RPM capability of the motor.

Matt: All good about the thread. We can all learn a lot from all the replies.

This is where I've lost you. Why? I understand "inverter gets too hot", but... why?

IGBT's have quite high switching losses compared to mosfets. They take a long time to turn on and longer to turn off. That's why you don't want to go too high in frequency. The faster you switch them the more the switching losses become the dominant heating factor. Too high and the switching causes more heat that the current going to the motor.
Mosfets are better but historically couldn't go to high volts with decent current ability. Until the SiC mosfets came along at 1200V.
That's why the Tesla controller can go to 17Khz and not break a sweat.

Yes generally the name plate will show the specs relative to each other. Some might be maximums but a figure of 961 rpm is very specific and is probably at 50hz rather than a maximum. The problem with 50Hz motors is not usually the windings but the laminations. They use thicker lamination steel (its cheap) as they don't need to go to high frequency, where the motors designed for 400+hz use really thin laminations and different silicon content to achieve really low eddy current losses. Pretty much means that it just might not be as efficient as a motor designed for that frequency. Some of the magnetic flux will be absorbed into the stator as heat instead of driving the load.

Iulian talks about lamination thickness and frequency in his PM motor designs.
https://www.masinaelectrica.com/130kw-m ... struction/

Only way to know is to test it when using an unknown motor.

mikejh wrote: ↑Fri Apr 21, 2023 2:05 pm So does the AC controller still work as a current controller with a PI loop for torque control?
From my experience with DC controllers the throttle sets the demand current (torque) rather than speed (rpm). PI loop controls PWM to achieve the demand current. So hopefully the PI loop keeps current inside reasonable limits and the hardware over current protection only kicks in if the PI loop goes nuts or doesn't catch it quick enough.

I think part of the confusion here may be coming from using the terms AC controller. Both induction and permanent magnet motors are AC but the control algorithms used by OI to control each are very different. The standard AC induction motor control doesn't use PI controllers for current control, it essentially runs open loop controlling voltage based on speed and load and uses slip angle to control induced current and so torque (not an expert so don't quote me on this!) .

The Field Oriented Control used for the permanent magnet motors does use PI controllers and is pretty much as you describe. Essentially the phase currents and motor position are measured and used to calculate the actual direct and quadrature currents. The desired direct and quadrature currents are also calculated based on the requested torque. The PI loops then use the error between the measured and desired currents to servo the direct and quadrature voltages that need to be applied to the motor. These direct and quadrature voltages are then converted back to phase voltages and output as PWM values.

The over current trip, whether it is built into the inverter (Prius) or hardware comparator on the OI board, is separate to this and is just there to act as a fail-safe in case the software doesn't manage to keep things under control.

mikejh wrote: ↑Fri Apr 21, 2023 2:05 pm Might have to try to model up a Leaf motor. You've certainly got me thinking about it.
Shame about the tiny output shaft. Still I can machine up a similar coupler and get the splines wire cut to suit.
Trikes only looking like about 350-400Kgs so I think these motors will be fine even with the lower gear reduction. If it's sluggish I can always go with 4.11 gears due to the higher RPM capability of the motor.

Have you considered using something like the complete Outlander motor transaxle? It might fit between the rear wheels (depending on your suspension setup) and leave the section of the frame currently occupied by the Hyper9 free for batteries too?

mikejh wrote: ↑Fri Apr 21, 2023 2:05 pm They use thicker lamination steel (its cheap) as they don't need to go to high frequency, where the motors designed for 400+hz use really thin laminations and different silicon content to achieve really low eddy current losses. Pretty much means that it just might not be as efficient as a motor designed for that frequency. Some of the magnetic flux will be absorbed into the stator as heat instead of driving the load.

Iulian talks about lamination thickness and frequency in his PM motor designs.
https://www.masinaelectrica.com/130kw-m ... struction/

Only way to know is to test it when using an unknown motor.

Excellent point on the laminations! Thanks for the link too, some really interesting/impressive stuff there

Pete9008 wrote: ↑Sat Apr 22, 2023 6:07 pm I think part of the confusion here may be coming from using the terms AC controller. Both induction and permanent magnet motors are AC but the control algorithms used by OI to control each are very different.

Thanks Pete. Yes I hadn't considered the difference in control between induction and IPM. Makes more sense now.
I spent a few hours yesterday re-reading my Electric Powertrain book and the math did my head in. I still like the idea of switched reluctance motors. So much simpler...

Have you considered using something like the complete Outlander motor transaxle? It might fit between the rear wheels (depending on your suspension setup) and leave the section of the frame currently occupied by the Hyper9 free for batteries too?

I hadnt really considered the entire transaxle due to the transverse mounting arrangment of the motor. The frame is only 320mm wide (centers) so it might stick out the side. I like symmetry.

Excellent point on the laminations! Thanks for the link too, some really interesting/impressive stuff there

Yeah I've been keeping an eye on Iulian's site for a while since he designed his first 45kw open frame motor, hoping for a release date.
He's now formed Joule motors but not much information as to when they will be available. I'd love to be able to use the 130kw motor. That would leave plenty of room for batteries and abundant torque for direct drive. Front battery pack comes up to 147Kg though. And pack voltage of 313V nominal.

I do like that last render

That motor looks pretty much perfect. Have you tried contacting Joule motors to see when they might be available?

Thanks Pete,

Yeah it does. Not as much torque as a Hyper 9 or Leaf motor but I don't need it to be a rocket ship.

I have contacted them in the past via their site contact form but didn't get a reply.
Just today I found another email address that Iulian has on his site so I gave that a try. Fingers crossed.

Maybe I'll try to fit another two cells in the rear box and go for 100 cells all up. Nice round number and 320V.

Received a reply from Iulian on the Joule Motors 130KW motor.

They are available now and one off pricing is around 7000 euros.
With our crappy exchange rate in AU that converts to about $11500 AUD before shipping.

He did indicate that the price could get better with a larger order.... but I only need 1.

Still the Hyper 9HV used to be $6700 here but they are now listed at $8200.

More thought, design work, and permit to build required first...

Ouch, that's more than I'm planning on spending on the entire conversion (including the donor car)!

I think I'd be looking at importing Leaf motor. It wouldn't look as good as the Joule motor but you could always put a dummy case round it.

Another alternative, if you want a bit more of a project, is to use the MG2 rotor and stator from out of a GS450H gearbox mounted into a custom machined housing. Damien's dyno run indicates that it can do around 225Nm and 150kW on a 96s pack and it should be around the same dimensions as the Joule motor.

Regarding the battery I hadn't realised the back box was batteries too - what kWh capacity do you have and what range are you aiming for?