Low voltage, high amps, can it scale > 100kW?

[zombiess]

The topic of running low voltage at high amps has come up, often from LiveForPhysics so I wanted to start a separate thread on it to further discuss.

Right now the status quo is to run mid to high voltage <700V to get high power levels. It's been suggested that the same is achievable at lower voltage and higher amps, say 100V 1000A for a 100kW config. Lower voltage has some advantages such as a less complex battery management and less danger. I'm wondering what a low voltage high amp +100kW motor looks like? What are the challenges in making the motor? What does it's inductance look like? I'm familiar with the controller side challenges and they are difficult to overcome, but 1000A is doable. I'm not sure how far beyond 1000A it can go before there are major issues.

Low inductance/resistance motors are difficult to control, the answer here is running higher switching frequencies, but higher switching freq means more controller losses and heat to dissipate. Another large issue is the required dead time (higher Fsw = larger percentage of dead time) creates all kinds of low frequency harmonics which muck about with the motor and can cause control loop issues and lower overall efficiency. There are ways to minimize dead time to well under 1us, but it's still a challenge.

Where do you think the realistic limits are on how far this mostly unexplored solution could go for EV drive systems?

 

[Arlo1]

I think the sweet spot is between 28s and 98s

Motor inductance helps with motor controllers as well as things like field weakening. Systems with higher voltage have motors with higher inductance.
A motor wound for 2x the voltage is going to have 4x the inductance. This makes the amperage rise time 1/2 with the increase in voltage so the controller is able to control the current easier.

As for current I think there is limits either way. If you want 1000hp you will find it extremely hard to do with 50v but if you have ~500v it will be easier but now you have the problem the high voltage will bring. And its not just safety that changes but things like corrosion and Battery Management will become much more complicated.
Voltages above 120 scare me as they should.
I will make an attempt at as much current with 150-200 volt rated components as I can but I think the current will be something that will add noise to the system as you increase it.

There is always more then 1 way to skin a cat. multiple power stages etc. But everything has its trade off.

Tonight I got Zapped with 220v DC while hooking up the contactor on the second side of my battery for the leaf motor controller Its crazy how sneaky high voltages can be.

 

[zombiess]

Now try adding in RF energy into that HV and watch how much sneakier it can get.

Is the answer multiple controllers + windings on the motor to keep things in check?

 

[Arlo1]

Now try adding in RF energy into that HV and watch how much sneakier it can get.

Is the answer multiple controllers + windings on the motor to keep things in check?

I don't think there is an answer there is more then one way to do this and time will show who finds better luck at what they do.
Multiple controllers take up a lot of space and add complicity. But some are doing it because a size 6 is not enough

 

[Arlo1]

I think the future of this stuff is going to be when someone turns on the switch to artificial intelligence and lets a computer/robot figure it out.
What takes use years as humans will be sorted out in minutes by robots!

 

[jonescg]

[youtube]Nnjh-zp6pP4[/youtube]

Had to be done

 

[flathill]

The reason we are stuck with thinking we need high voltage is because we are still thinking stationary where it costs more money to transport electrons using low voltage.

For land/air/sea electric vehicles low voltage and high amps is the obvious solution, even for a nuclear powered aircraft carrier. If the battery or power source is next to the motor why use a lethal system voltage?

The controller and the motor must be designed as a system. The controller should bolt onto the motor. There should be one battery per motor for redundancy and to keep the battery to motor wire as short as possible. You might even want one battery per phase as pulse discharge is better for battery life. Complexity can equal redundancy. We now have magnets double the strength of neo. Ironless hub motors are going to happen. They need permanent magnets to enable the torque density we need. Permanent magnet motors are not fault tolerant and a winding failure can cause a locked rotor and your vehicle to crash. You want one controller per phase. The number of phases should equal the number of coils. The master controller is a self learning fpga driving all the sub controllers, but each sub controller can also work independently if the master link goes down. This is the end game system design.

 

[jonescg]

Right, so in the context of this thread, where can we see examples of 100 kW+ electric vehicles using less than 150 volts? I want to exclude big DC motor machines here, as these are rarely at 100 kW for very long.

And for those that are putting out 100 kW, what is the weight of the power train? The 150 kW power train in Voltron Evo weighs in at 53 kg. The battery to power it is anywhere from 53 kg to 85 kg, depending on the race demands. So about 100 kg to 130 kg. Lets round it up and call it 1 kg/kW as a package.

Pairing up multiple motors and controllers is one way of doing it, but it adds epic complexity, at least on par with a 168s battery and associated cell management. Are there any examples out there?

My view is that 360 volts is a great Vmax for cars and motorcycles putting out 100 kW+. Our race bike is only running at 700 V because the motor-controller-battery combination meant that the best power and efficiency dictated it. But like nuclear power - you can engineer many safety features into a system making it inherently safe, despite the obvious risks.

 

[liveforphysics]

If everyone only made things that already had been built and well proven, we would still be banging rocks together.

HV for traction drive applications will be something only laughed about in less than a decade when EVs mature away from legacy industrial motor drive nonsense.

Things that won't be parts of future performance EVs.

Voltages > 100vdc
Motors with only 3p.


Get a big hydraulic crimper, mine was like $80 shipped on amazon. Cold-forge-crimp some appropriate lugs onto appropriate sized cable, pass what current through it you designed it to handle.

The reason it hasn't yet occurred is only because the controller tech available today is largely built from copying legacy industrial junk never intended for an application with the battery and motor a couple feet apart.

 

[jonescg]

I can crimp lugs as big as my fist - but there's no point because there's currently no performance components for me (or the EMUS guys) to bolt them to.

Is this a huge research area going wanting? By the sounds of it you could write a research grant based on this immensely important need. I think the notion of a performance EV running on under 100 V DC is an awesome idea, and I'm sure the entire EV industry would be stoked to see it happen. Surely from a safety and complexity perspective, Tesla, GM et al. would be jumping at the prospect.

So why isn't it happening? Is it because it's expensive to manufacture multi-phase motors and low inductance controllers? Or is it cause it's a solution to a non-problem?

 

[Gab]

Isn't the whole reason to reduce the amps to do with heat dissipation and keep the system lightweight? higher voltage and lower amps means less heat for the same power. To Dissipate heat you need thermal mass and this adds weight to motor but also very thick cables. So i think high currents are possible today but people do not do that as it means you need very heavy motor with high thermal mass to handle this huge current and not overheat very quickly.

If the motor was used in a stationary machine the weight would not matter much, but in a bike or a car every Kg matters hence this is not done.

 

[Arlo1]

Wrong.
The motor will weight 100% exactly the same for the same power rating it will have the same copper fill only be rewound for a different KV nothing changes here.
The wires from the controller to the motor will be bigger for lower voltage and higher amps so they can flow more current with out getting hot.
So the weight will be very close you will have a very small weight penalty with phase wires and battery wires but you will have a much safer system and easier to control with a BMS. One thing to point out is the cost of the components rated for high vs low voltage.
Everything from the caps to the power switches in the controller will be cheaper.
The part for the build like the fuse and contactor will be cheaper with lower voltage.

 

[liveforphysics]

Isn't the whole reason to reduce the amps to do with heat dissipation and keep the system lightweight? higher voltage and lower amps means less heat for the same power. To Dissipate heat you need thermal mass and this adds weight to motor but also very thick cables. So i think high currents are possible today but people do not do that as it means you need very heavy motor with high thermal mass to handle this huge current and not overheat very quickly.

If the motor was used in a stationary machine the weight would not matter much, but in a bike or a car every Kg matters hence this is not done.

Time to learn the basics Gab.

Running high voltage low current or visa versa:

No added cell heat, no added motor heat.

Only the conductive paths between the battery to controller and controller to motor need to become proportionately larger.

Unlike in some industrial factory where equipment can be thousands of feet apart, these distances are often only a couple of feet in an EV. It's quite practical and simple to use a couple aluminum bus bars or for the DIY'er to crimp some stout little cables of appropriate cross section to have no additional loses drawing say 2000-- 10,000A or whatever the application requires.


Dr. Jones- You continue to ask why the mature future of EV design isn't here yet. The answer continues to be that someone needs to build it.

With each steady new release of higher current density semiconductor designs, the task of making the missing piece of the puzzle, the controller, gets easier.

 

[jonescg]

Dr. Jones- You continue to ask why the mature future of EV design isn't here yet. The answer continues to be that someone needs to build it.

With each steady new release of higher current density semiconductor designs, the task of making the missing piece of the puzzle, the controller, gets easier.

I'd say my question, more accurately presented, is why is there currently no demand for a low voltage high power system.

 

[jonescg]

Tonight I got Zapped with 220v DC while hooking up the contactor on the second side of my battery for the leaf motor controller Its crazy how sneaky high voltages can be.

I'm very pleased you're still with us, Arlin! But how on earth did you get a zap from your battery pack while hooking it up? Don't you have isolation or at least a service disconnect somewhere in the system?

 

[Arlo1]

Dr. Jones- You continue to ask why the mature future of EV design isn't here yet. The answer continues to be that someone needs to build it.

With each steady new release of higher current density semiconductor designs, the task of making the missing piece of the puzzle, the controller, gets easier.

I'd say my question, more accurately presented, is why is there currently no demand for a low voltage high power system.

There is but at the moment the EV world is still quite small.

 

[Arlo1]

Tonight I got Zapped with 220v DC while hooking up the contactor on the second side of my battery for the leaf motor controller Its crazy how sneaky high voltages can be.

I'm very pleased you're still with us, Arlin! But how on earth did you get a zap from your battery pack while hooking it up? Don't you have isolation or at least a service disconnect somewhere in the system?

Yes the service disconnect you mean the contactor I was connecting? 220 into my hand was just a nasty tingle.
But I do have plans to make this safe and I think I will make it so I can brake it down into 204v (nominal) sections with a Anderson or something. I am still working out how to put this together in my head.

 

[zombiess]

I'd like to see some data gathered on silicon cost, voltage, amps, number of phases and motor inductance to see what the sweet spot might be for a multi winding multi controller config. At some point we reach diminishing returns on paralleling lots of small devices which can make modules a better choice even if the cost is higher. I'm not sure where that point is yet.

I'd like to see the insides of a 1000A controller. I've seen the Sevcon Gen 4 size 6 and it's surprising to me that it works at the current level it does.

Once the decision is made to go multi phase/winding, getting a motor produced is the next step. Controllers seem to have the ability to be ganged together with some success. What does it cost to have a custom motor produced, then what is it put into? How many phases, how many controllers?

 

[Punx0r]

Surely it's not even a case of a custom motor, but an existing one rewound for your desired number of phases (assuming the motor has more than 3 poles )?

 

[zombiess]

Say it is rewound with 2 pairs of 3 phase with high current, now one has to contend with 6 thicker phase leads exiting the motor in the same space 3 thinner ones previously occupied.

This might not be an issue on car sized motors, but rewinding is a low volume solution, but not a bad place to start.

 

[flathill]

Remember to maximize torque you don't actually want a sinusoidal back EMF, but if you only have 3 phases you need it to minimize torque ripple.

Multi-phase machines are much more common on military craft that use low rpm direct drive permanent magnet motors.

Direct drive motors, whether in-board or in-wheel, are the future of EV's

You can use two 5-phase star windings and stagger one set of windings 18 electrical degrees to effectively eliminate torque ripple with trapezoidal back emf motors. What you end up with is a 10 phase motor which is effectively a 20-phase motor but with a little more losses due to recirculating current, and can be driven by two 5-phase controllers for redundancy.

 

[liveforphysics]

A topology as you described Flathill would make the controller side of putting huge power into a motor with low voltages very easy.

These things will begin to become commonplace once the EV world develops beyond the industrial nonsense based status-quo composing the current market offerings.

 

[flathill]

5-phase double star motors also have less endturns so the motor is more compact and has less losses than the equivalent 3-phase motor (end turns of one coil do not cross end turn of another coil)

Also you need 3 times less phase current than an equivalent 3 phase motor, so yes a sevcon style power stage we already make today would already scale well beyond 100kw

All we need is two "size 6" 5-phase sevcon controllers that don't yet exist

Even without going to 5-phases it is still possible to do, but like I said you need more than 3 phases if you are serious about preventing motor faults that can produce enough torque to cause a wheel lockup on a motorbike or the more tricky case when you 2 or more motors driving each side in a car for torque vectoring. If one motor fails it can is some cases cause the car to turn and crash in an instant with no time to countersteer. With a 5-phase double star motor you can actually detect the fault and correct for it so the motor feels like normal until you have time to get it serviced. Each phase of each star winding is physically separated and one slot only contains one coil, so the chance of a coil to coil or phase to phase fault is almost 0.

 

[parabellum]

Flathill, can you please link some 5-phase double star motor docs or even better advantage explains. I just do not get why 5.

 

[Danny Mayes]

I have always wondered why we don't see 4volt systems. Eliminating the BMS altogether. Also the battery pack would not have to be serviceable, if a cell goes then you just loose some capacity. This would make construction way simpler (just cast the cells in resin) and reliability would increase not to mention safety.

Maybe not feasible for > 100kw but certainly for < 100kw applications.

I am not so strong with electronics so there may be obvious reasons but i can't see why there is not more interest in this. Maybe there is and i am unaware.

I would love to make a 4 volt power stage for Labowski's brain but I am a mechanical guy, that is way out of my league

D