Maximum Hub Motor Voltage?

[Kingfish]

I have been searching the Sphere for a thread on this; perhaps it’s been discussed by another set of keywords – though I haven’t found much on the subject. Most folks are constrained by amps, and I wanted to talk about another way to look at the problem with getting power efficiently to the motor.

So instead of stuffing a bunch of current at < 100 volts or even 150 volts, does it make sense to turn the problem around and go big on the voltage and low on the current? I know Luke and bigmooose weighed in at one time or another and suggested that the next convenient voltage break for controllers after 150V is upwards to 300-400 range like what they use for eCars and eMotorcycles (based upon switching devices).

My concern specifically is that voltage dictates top-speed if we’re single-geared, w/o reduction, direct drive - aka hub motors. Out of a lark does anyone know the voltage limit of a typical 9C hub? If someone develops a HV-controller could the 9C handle it or would the dielectric of the magnet wire break down?

Curiously KF :|

 

[amberwolf]

The limit is going to be the insulation on the coil wires themselves. I don't know what that voltage limit is. Could be a couple hundred volts to several hundred. Probably someone would need to test that.

The kV itself doesn't matter as long as you are using them as a middrive, but if you intend to keep using it as a hubmotor you'll wanna probalby rewind them for a different kV, and possibly use a smaller wheel than originally intended.

If you do rewind it, then you can use whatever insulation class wire you want, and so have it as high a voltage as you can find wire for.

 

[John in CR]

A 9C will require balancing and proper bearings to run much above stock rpm. I did that with mine for running as a mid drive at 150V, and it is ok up to about 1500rpm no load. My balancing was fairly crude, but 300-400V would require the job to be done properly. Another consideration if going that high is the controller's processing speed with such a high pole count motor. 24 poles at 3krpm is 72k electrical rpm.

 

[Arlo1]

Your over thinking it Kingfish.

I run my x5 at 100v if I wanted I could likely run it at 1000v and get 10x the top speed. But......... thats 850km/h
If you want run 200v... The thing that hold most people back is the IRFB4110 is an amazing fet for its size and price! You want to run higher voltage then you have to find a controller to handle it! My x5 might start arcing at some poin as I up the voltage due to the worn phase wires from taking it apart a few times ot the windings them selfs are black so at some voltage it may arc out... But that doesnt matter because Its limited by other things!

 

[ZOMGVTEK]

The biggest issue is the MOSFET availability. Compared to the IRFB4110PBF, there is nothing close to its capabilities at 200V. Even assuming you can cut the current in half with a double in voltage, you can not make up for the Rds losses and current handling limitations. Plus, 200V realistically means 48S LiPo, which is tricky to charge, and almost forces you to have a large pack to help the cells balance. You could make a 48S1P pack, but chances of getting near the rated capacity before hitting LVC on any cell are going to be low. 48S2P and beyond are starting to be really sizable packs for what most people run.

The IRFB4127PBF is rated for 200V, 17mOhm Typical, 76A.
The IRFB4110PBF is rated for 100V, 3.7mOhm Typical, 180A Silicon limited, 120A package limited.

Thats 4.6 TIMES higher Rds, and 2.6x lower current handling, all for twice the voltage. This is a problem. Also, realistically, 200V into a 200V FET is really on the edge. Don't even bother with the 250V FET's, they're horrible.
Throw in the fact that with the kV of most motors out there, 200V would be stupidly unrealistic speed, and you have the reason.

That said, I'm trying to get a GM 15 FET to run on 48S... So far it exploded a little.

 

[Kingfish]

The speed I am targeting is upwards to 75 mph. I suggest the 9C purely for non-distructive testing. I would not try going that fast on a bicycle For my next bike I need to start someplace and I am looking for the upper limit of the present hub motor.

The calculations for motorcycles suggest that decent powered motors begin around 20 hp - and at 100 or 150V that is still a lot of current.

If I had 150V, 1 hp is about 5A & 10hp = 50A. The idea that I had was to reduce the current to reduce the heat, although switching would create heat too. I think that the 9C is limited to 5kW without cooling.

Anyways, I'm just toying with ideas to figure out the next path forward and sort out the best options that will get me from here to thar

Zomgvtek: I didn't plan on using MOSFETS if going HV cos of what you said is quite true. I feel like I have taken the bicycle model about as fast as I care to go safely. So the next level is eMotorcycle. I just don't think I could get away with pedaling at 40 mph :lol:


Best, KF

 

[ZOMGVTEK]

My current GM is in sad shape, so I will probably do some higher voltage experimentation on it when I get a 5403.

I entirely agree that current is bad, and voltage is good. The only issue is things start to get seriously tricky when dealing with HV DC.
Near and above 100V you start getting to be a bit dangerous. The arc through 45A PP's, charging up resistors to 200V is BRUTAL.
With how most people more or less throw bikes together, its no wonder they run 36-48v. You should see my brothers build, its really sketch stuff.

The main reason I would want to go higher than 100V would be to increase efficiency, which means I can turn the power up.
In theory, it works, the only issue is it does not appear to really work out in practice.
I don't think I want to be comfortable on a cheap bicycle going 60-80 MPH. In time, I probably would, but this isn't necessarily a good thing.
It does bother me a bit how hard it is to get a REALLY solid BLDC controller for high voltage. Even the HV Zilla only takes a bit under 400V.
To make a 1,000 HP electric car at 350V, you need a few THOUSAND amps. To me, it sounds a lot easier to run 1,000V than 1,000A.

Regardless, I'm going to be messing with 200-400V stuff in the near future. I'm a little worried about pushing my GM past 200V, the windings charred appearance is far from comforting. So far it spun a few revolutions at 200V and didn't fail.
Nothing terribly serious, just trying to see how far this off the shelf stuff will go.

 

[Doctorbass]

Kingfish, are you targeting 75mph with a ....9c ??? :|

If so, it will need to hold 10-15kW minimum until you reach that speed.
I bet it will overheat before you reach that speed due to the low stator mass..

anyway.. i'm just curious.. i'm at 70mph now with an X5 5303 and it work well but the continuous power is 10-12kW and motor temp at 120 celsius constant during the 40sec i was at that speed
I agree on the mosfet.. the 4110 are the best way...

Otherwise the kelly KBL 220A 120V is giving me like 25% more power than a 18 fets infineon and can take up to 136vdc

My estimate is that the X5 can take acceleration uop to 88mph (140km/h) without overheat ( above 160 celsius)... with a kelly programmed to 100% current and running at 32s (134VDC)


There is a kelly 72V ( max 90V) with 350A on ebay right now at 400$.. that might be the key for you with higer KV motor

Doc

 

[wojtek]

soon i will test this baby: http://www.endless-sphere.com/forums/viewtopic.php?f=31&t=19719
with modified HS and later if all goes well with group buy, i will put him on 5404!

 

[liveforphysics]

The idea that I had was to reduce the current to reduce the heat,

It absolutely 100% does not work this way.


If you're stuffing the slot with as much copper as you can fit, it makes no difference if you're running 48v or 480v, the motor can't make even a tiny bit more power on 480v than it can on 48v in the same RPM range.

 

[Arlo1]

The idea that I had was to reduce the current to reduce the heat,

It absolutely 100% does not work this way.


If you're stuffing the slot with as much copper as you can fit, it makes no difference if you're running 48v or 480v, the motor can't make even a tiny bit more power on 480v than it can on 48v in the same RPM range.

Unless the controller is what was limiting it in the first place right?

 

[neptronix]

You absolutely have to take the total amount of watts you'll be pushing into account..

More volts = higher top speed. Higher speed = more torque needed to sustain that speed. More torque needed = higher amps. That makes for more watts needed. And those motors are 80%-85% efficient, so the more power you dump into it, the more waste heat you are creating.

I have yet to take my 9C/MXUS DD over 39mph for any extended period of time without reaching the edge of blowing it up. I am shocked that it survived the 2800w i was pumping into it constant @ 80v/37a.

 

[ZOMGVTEK]

The idea that I had was to reduce the current to reduce the heat,

It absolutely 100% does not work this way.


If you're stuffing the slot with as much copper as you can fit, it makes no difference if you're running 48v or 480v, the motor can't make even a tiny bit more power on 480v than it can on 48v in the same RPM range.

Would you agree that its still a better idea to run 100V instead of 10V, assuming the motor is wound for the same RPM's on both setups, and the current limit was set for the same power on both?

 

[Kingfish]

Firstly, I am not going to run a 9C hub up to 75 mph on a bike; I am curious about the upper limit though so I know when to stop dumping more power into it. Allow me to rephrase: I have zero-intention of driving an electric bicycle to 75 mph, at least not a stock bike! Maybe a custom-built bicycle… but that’s a pipe dream at best. :wink:

I leave it to Doc to break all the ebike records and support those who are so willing 8)

My path is that I wish to make my own motor. I want to use the 9C hub to create a profile and then compare the new motor design(s) against it because it is a known quantity. The new motor will be used either on a bicycle frame, a custom frame, or adapted to a motorcycle frame.

I understand that voltage can raise the hub motor’s top speed, just as the gearing configuration can change it. The goal is to define the reasonable limits of the 9C through various common voltages < 100V cos I don’t plan to push a 9C hub higher than that with the controllers I have in stock. Then compare the new motor designs at like voltages and currents to evaluate the efficiencies.

If the motor performs well (and actually I want it to perform better) and I can reach my targets of sustained 75 mph at STP @ sea level (Seattle more or less) without major cooling issues (huge investment in weight/$) then I will be happy and call it good.

However I am speaking figuratively and I haven’t built diddly. The question remains – what is the maximum voltage of a hub motor, and that’s the part that I am interested in.

Best, KF

 

[liveforphysics]

The question remains – what is the maximum voltage of a hub motor, and that’s the part that I am interested in.

Best, KF

You could certainly make a 20,000V hubmotor, and switch it with a vacuum tube based controller if you wanted to just waste efforts on something.


The single advantage you gain from increasing voltage and winding a motor for high voltage, the single solitary advantage, the only single thing that improves, is not needing as large of wire between motor to controller, or controller to battery.

Every other aspect remains exactly identical in respect to heating, torque, power, etc etc. Except when you cross out of mosfet range into IGBT range, where you pay a Vf penalty from the IGBT's body diode which hurts controller efficiency a bit.

So, yes, if you want to increase complexity of the battery system by a large amount, and run a less efficiency less robust controller setup all to arrive at the same place you end up if you just range a 48v system (or whatever), you can run into the kilavolt range if you feel like it, the wire with proper insulation is certainly available.

 

[Doctorbass]

I think that what KF is asking for is if we woudl take a conventional hub motor like the HX, 9C or X5...

My answer to that is that the voltage will be as high as the weakest insulation area can take in the winding.

A winding that is arcing is the weakest point for voltage i think so it depend on the insulation class and the air presure.. let say 1 atm :lol:

let see teh worst case: ( two different phase winding touching together but that have a notch in the insulation exactly at teh same place so the arcing distance remain total the insulation tickness wich is the total tickness of the insulation of both wires
Arcing depend on the distance so if the winding is made of 21AWG copper wire and that the insulation tickness. size of each strands: ( 0.029"dia) and insulation is around 1.5mil


1.5 x 2 = 3mil (7.6um) . now If we need 10kV per cm, that mean: the arcing will occur at 10v per um wich is 76V for 3mil

so if you have two wire without insulation on a notch that are together, the spacing between the copper that is 7.6mil on the X5 winding insulation will need 76V to arc

BUT.. THAT IS IN THE WORST CONDITION

Doc

 

[liveforphysics]

The way a motor is wound, its very tough for a pair if wires to even get to a 10v difference somewhere inside the motor, even if running a 150v battery or whatever

 

[John in CR]

Isn't the real limit for something like a 9C going to be an electrical rpm limit where the frequency starts getting high enough that those cheap laminations are getting too much heat from eddy currents.

Since Kingfish asks for maximum in the title, but his last post mentions the 100V practical and cheap controller limit, then I'd have to say that that any direct drive hubbie can go up to 100V.

 

[Kingfish]

Isn't the real limit for something like a 9C going to be an electrical rpm limit where the frequency starts getting high enough that those cheap laminations are getting too much heat from eddy currents.

Since Kingfish asks for maximum in the title, but his last post mentions the 100V practical and cheap controller limit, then I'd have to say that that any direct drive hubbie can go up to 100V.

Understood; hub motors (at least the big 3 mentioned) can tolerate 100V - gotcha :wink: I feel safe then for my series of proposed tests; thank you!

Out of a lark, and because I am still very curious, what is the highest voltage anyone has ever taken a hub motor, and if possible, please post the associated current (or wattage). Embellishments accepted (share the story of why and how).

Thanks for all the kind replies, KF 8)

 

[ZOMGVTEK]

For some reason, not many people are willing to push stupid voltages into hubs. I suppose there is no real benefit, even 200V on my GM would push it over 100MPH unloaded. Probably a entirely unrealistic speed, given that it will burst into flames before hitting 80.

That said, one of these days I will make a controller that actually handles 200V, and I plan on doing a few WOT runs on it. My motor was technically propelled by a controller being fed 48S for less than 1 second, and it still works fine on 24S.

 

[amberwolf]

One good reason to push a hubmotor to really fast speeds (if you have a controller that can take the voltage and commutate that fast) is to use it as a middrive of some type, so that you are reducing that much higher speed down into much higher torque instead.

 

[John in CR]

I've run a 9C with my 150V controllers with the 4115 fets. That was as a mid-drive. It worked fine though I really never got to push it (just 2 short test rides), because I had some sprocket and chain alignment issues. I put my big scooter hubbie on that swingarm for a test and haven't gone back to working the kinks out of my mid-drive. Brute power is just too intoxicating.

Ideally I'd like to push the ventilated 18 fet 150V controller to 50-60A with either my ventilated 9C 2807 that I've modded for mid-drive only use, or do the same with one of my smaller BL36 Aotema hubbies that are more of a speed wind. I believe it's an interesting route to well over 5kw power in a lighter package than my 20lb hubbies. With a swingarm mount, but the hubbie more than half way to the pivot, the unsprung weight is much less of an issue.

At some point soon, I'm going to go even further with a scratch build for off road, and put the swingarm pivot at a hubbie axle to totally spring that weight. It will also make the chainline simple for combining the pedal chain.

 

[Arlo1]

Isn't the real limit for something like a 9C going to be an electrical rpm limit where the frequency starts getting high enough that those cheap laminations are getting too much heat from eddy currents.

Since Kingfish asks for maximum in the title, but his last post mentions the 100V practical and cheap controller limit, then I'd have to say that that any direct drive hubbie can go up to 100V.

Understood; hub motors (at least the big 3 mentioned) can tolerate 100V - gotcha :wink: I feel safe then for my series of proposed tests; thank you!

Out of a lark, and because I am still very curious, what is the highest voltage anyone has ever taken a hub motor, and if possible, please post the associated current (or wattage). Embellishments accepted (share the story of why and how).

Thanks for all the kind replies, KF 8)

In all reality they can probably tolerate ~1000v! But you will find reasons we have stopped in the 100-200v range. For one my bmx goes 54mph!!! that's a x5304 on 100v for two the irfb4110 is a very good very cheep mosfet and once you go to a 150v mosfet you gain controller heat beacuse of the higher rdson. Then look at how bad something ~200v is for mosfets...
There is no real maximum voltage for a electric motor when you have other limiting factors like mosfets caps, electrical rpm limits of the controller and speed limits of a bicycle!

 

[John in CR]

I'd want to be behind a concrete wall if a 9C was running at 1000V. 10krpms at that diameter sounds like shrapnel to me. We need video. Who has a 1kv controller?

 

[Arlo1]

I'd want to be behind a concrete wall if a 9C was running at 1000V. 10krpms at that diameter sounds like shrapnel to me. We need video. Who has a 1kv controller?

I bet it would last but who cares.... As if you are going to ride a bicycle at ~500mph!!