Want to build 500Hp electric turbo, people with experience please chime in

V

Vetal

Member
Hello to all forum members!
I plan to build cheapest possible 500Hp electric turbo for my street V8 (and possibly move it to other engines since it is so vehicle-independent).
But Speedmaster P-2 charger would cost about $1100 in Europe so I don't want to spend that. I want to use compressor wheel + housing from conventional turbo (cheap ebay ones). Like this:
1748270886358.png
and somehow direct drive it by a motor.
This will make system much more flexible (able to swap wheel/housings).

I have couple of questions:
1) am I right in selecting the largest wheel possibly? Reasoning is, they usually spin slower, so might be useful at some 30.000 rpm already. I think Alex even with 40KW motor could not spin compressor faster than 32.000rpm.
On the other hand, I might get Castle 1721 2400KV motor (with 33.6V LiPos) which spins to 70-80K in hope that it will spin some SMALL turbo wheel at that speed. But it's only 10-12KW AT MOST
2) Why do we use high voltage and LiPo batteries? Why not use car 13V system? Battery provides current up to 500-700A so that should be enough, and if not - put second car battery.
3) does compressor wheel put some meaningful axial load on the shaft? If so, it would require some additional thrust bearings and complicate the system
4) Not looked into it but I remember gamers use all kinds of liquid cooling their systems, might cannibalize something from that area too to increase power

Guys, please advise :)
 
brettus8 said:
I haven't seen a converted turbo setup that works as well as the P2 yet. Maybe you will be the first .....
Click to expand...
That's time, money, experience ) I'm sure Alex would have got the same results if he went this route.
BTW, Alex, have you had any in-car results with TP Power or Castle motors? @AlexLTDLX
 
Vetal said:
Found that water blankets for these small motors are plentiful on Aliexpress
Click to expand...

I bought these on Ali Express, I can fit two comfortably on a motor, and will find some suitable heat transfer material. I don't think CPU heat paste will work because of the constant vibrations. I haven't worked out how to get air flow around them yet but I am sure I will work something out

They are a very tiny bit too small, but cutting one of the grooves lets them expand and clamp on nicely

1748401661437.png
 
r.brown.bayliss said:
I bought these on Ali Express, I can fit two comfortably on a motor, and will find some suitable heat transfer material. I don't think CPU heat paste will work because of the constant vibrations. I haven't worked out how to get air flow around them yet but I am sure I will work something out

They are a very tiny bit too small, but cutting one of the grooves lets them expand and clamp on nicely

View attachment 839
Click to expand...
Yeah, passive cooling... But in engine bay it will be hot as hell even without motor running, so water cooling is a must
 
I initially had a water jacket cooling system on my setup, but found that it wasn't necessary with adequate airflow. Heck, even without adequate airflow, these things don't really run long enough to get much more than warm.

As to why I push the P2 so hard (I get nothing from people buying these things, I have no affiliation with Speedmaster at all). It strikes the perfect balance on MANY levels. First of all, trying to get a brushless motor spinning to 70k rpm under a significant load is very hard on the ESC. You need to consider electrical RPM - also known as eRPM - that's the mechanical rpm times the number of pole pairs in a motor. So if you run a 4 pole castle motor to 70k rpm, your eRPM is 70,000 x 4 = 280,000 rpm. That's a lot of very fast switching from the power MOSFETs in the ESC. MOSFETs suffer from internal capacitance - effectively, you have to charge and discharge this capacitance every time you switch a MOSFET on and off. This causes all kinds of voltage and current spikes. The faster you ask the MOSFETs to do this, the harder it is to control these spikes. Eventually, you end up letting the magic smoke out (I've done that many times). The P2 doesn't have to spin that hard because of the larger impeller diameter.

Secondly, the hardest part to manufacture in a centrifugal compressor is the impeller shaft. This may seem counter intuitive at first, but it must be extremely strong and extremely precisely made - otherwise, you'll see the impeller start wobbling, even the tiniest bit, and it'll hit a harmonic causing instantaneous catastrophic failure. One of the forum members (who's a machinist by trade) has been having exactly this problem making a shaft. They need to be made from some sort of extremely hard and durable steel - 300m or vasco or something similar. These materials are very difficult to machine because of their physical characteristics. Again, the P2 comes with a shaft that you don't need to modify (though I do wish they had an HTD drive machined into them rather than a T5 drive, but I'll take what I can get).

Finally, the P2 is suited to an extremely broad range of engines - from 200hp to 500hp naturally aspirated engines. This covers the vast majority of engines out there. And with the pulley drive, you can tune the pulley ratio to optimize the impeller speed for a particular engine.

If I had my druthers (and access to a manufacturing facility), I would design an integrated motor/compressor combo - it would be more compact & efficient (possibly much more compact if I could use the back of the impeller as the rotor in an axial drive configuration - but that has other potential issues), but then you'd give up the tunability aspect of the pulley drive.

And lastly, to answer your question about using the car's battery to power the thing. It's all about power - in this case, watts. P = IV. The more voltage you have, the less current (amps) you need. The Sledgehammer needs 34,000 watts to run. With a 12 volt system, if you could somehow eliminate voltage sag under load (a typical lead acid car battery sees about 3 volts drop on a 300 amp load), to hit 34,000 watts, you'd need to draw 2,833 amps from the battery. Again, that's with NO voltage sag. Practically speaking, it's impossible with a 12 volt system. You also need wiring, contactors and connectors capable of handling that current - also, in a car, practically impossible. I'd love to run even higher voltages - something around 80-100 volts; but those ESCs and motors aren't readily available (for two reasons - the high speed switching losses I mentioned above are harder to control at higher voltages) and also, anything above 80 volts is considered lethal. So there's a safety issue. As if these things aren't dangerous enough.

Hope this helps.
 
Vetal said:
Yeah, passive cooling... But in engine bay it will be hot as hell even without motor running, so water cooling is a must
Click to expand...
I am fortunate in this regard, I am able to put the unit in front of the left wheel and wall it off from the engine bay and get easy access to fresh air from the front
 
I would not put the unit in a hot engine bay directly; particularly over any exhaust components. The magnets in a brushless motor can start to demagnetize above 180* F. Low mounted in front is fine (mine is next to my radiator), pulling in air from the fender well, or you could mount it in the fender well directly.
 
AlexLTDLX said:
I would not put the unit in a hot engine bay directly; particularly over any exhaust components. The magnets in a brushless motor can start to demagnetize above 180* F. Low mounted in front is fine (mine is next to my radiator), pulling in air from the fender well, or you could mount it in the fender well directly.
Click to expand...

Yea, the fender well is a better name for the location it is in. I 3d scanned the well to help design all the parts I need to build. In the photo the left is the wheel cover, the right is the inside of the bumper cover. The last photo gives a good idea of the setting, the green line I drew on it is the bottom of the headlight. The opening in the bumper is where the factory fog light was. The thin plastic dust guard in the foreground I am replacing with a full-height aluminium sheet.

The engine in this car sits quite far back, the front pistons will be behind where the wheel cover is in these photos, the exhaust headers are well away from the fender well.

1748473941662.png1748473974615.png1748474001057.png1748474194869.png1748475663108.png
 
AlexLTDLX said:
I initially had a water jacket cooling system on my setup, but found that it wasn't necessary with adequate airflow. Heck, even without adequate airflow, these things don't really run long enough to get much more than warm.

As to why I push the P2 so hard (I get nothing from people buying these things, I have no affiliation with Speedmaster at all). It strikes the perfect balance on MANY levels. First of all, trying to get a brushless motor spinning to 70k rpm under a significant load is very hard on the ESC. You need to consider electrical RPM - also known as eRPM - that's the mechanical rpm times the number of pole pairs in a motor. So if you run a 4 pole castle motor to 70k rpm, your eRPM is 70,000 x 4 = 280,000 rpm. That's a lot of very fast switching from the power MOSFETs in the ESC. MOSFETs suffer from internal capacitance - effectively, you have to charge and discharge this capacitance every time you switch a MOSFET on and off. This causes all kinds of voltage and current spikes. The faster you ask the MOSFETs to do this, the harder it is to control these spikes. Eventually, you end up letting the magic smoke out (I've done that many times). The P2 doesn't have to spin that hard because of the larger impeller diameter.

Secondly, the hardest part to manufacture in a centrifugal compressor is the impeller shaft. This may seem counter intuitive at first, but it must be extremely strong and extremely precisely made - otherwise, you'll see the impeller start wobbling, even the tiniest bit, and it'll hit a harmonic causing instantaneous catastrophic failure. One of the forum members (who's a machinist by trade) has been having exactly this problem making a shaft. They need to be made from some sort of extremely hard and durable steel - 300m or vasco or something similar. These materials are very difficult to machine because of their physical characteristics. Again, the P2 comes with a shaft that you don't need to modify (though I do wish they had an HTD drive machined into them rather than a T5 drive, but I'll take what I can get).

Finally, the P2 is suited to an extremely broad range of engines - from 200hp to 500hp naturally aspirated engines. This covers the vast majority of engines out there. And with the pulley drive, you can tune the pulley ratio to optimize the impeller speed for a particular engine.

If I had my druthers (and access to a manufacturing facility), I would design an integrated motor/compressor combo - it would be more compact & efficient (possibly much more compact if I could use the back of the impeller as the rotor in an axial drive configuration - but that has other potential issues), but then you'd give up the tunability aspect of the pulley drive.

And lastly, to answer your question about using the car's battery to power the thing. It's all about power - in this case, watts. P = IV. The more voltage you have, the less current (amps) you need. The Sledgehammer needs 34,000 watts to run. With a 12 volt system, if you could somehow eliminate voltage sag under load (a typical lead acid car battery sees about 3 volts drop on a 300 amp load), to hit 34,000 watts, you'd need to draw 2,833 amps from the battery. Again, that's with NO voltage sag. Practically speaking, it's impossible with a 12 volt system. You also need wiring, contactors and connectors capable of handling that current - also, in a car, practically impossible. I'd love to run even higher voltages - something around 80-100 volts; but those ESCs and motors aren't readily available (for two reasons - the high speed switching losses I mentioned above are harder to control at higher voltages) and also, anything above 80 volts is considered lethal. So there's a safety issue. As if these things aren't dangerous enough.

Hope this helps.
Click to expand...
I've been waiting for your input patiently (almost :D) Hope you got your dyno and other stuff sorted :)
My bad, when I was thinking about 12V system, I was thinking about some 600Amps, so only about 7KW with an optimistic V drop. Not enough.
Miscalculation at my side.
BTW I think I have seen motors upto ~100V or so, I don't remember where: either bikes or scooters probably. probably not e-foils :D
 
I found some videos on Youtube with guys pushing 20-25KW of power with Castle/TP motors. Asked what this guy (25.5KW, looks like TP Power) runs:

2 more guys running 20KW+ on Castle 2028 1700 (can post if you like to see)
 
AlexLTDLX said:
I would not put the unit in a hot engine bay directly; particularly over any exhaust components. The magnets in a brushless motor can start to demagnetize above 180* F. Low mounted in front is fine (mine is next to my radiator), pulling in air from the fender well, or you could mount it in the fender well directly.
Click to expand...
One more thing I'm forgetting to ask. Did you run your car with e-charger switched off? I wonder how much restriction it is for V8 to suck air through compressor when it's off (whether I need a bypass valve when charger is off).
 
On a 500 hp NA engine, it's equivalent to 1 psi boost "loss." So it's not actually noticeable on the street.
 
Good to know. I had been thinking of an intake bypass but decided I didn't want to increase the complexity. I decided that so long as the car is not a hindrance in traffic that I could live with the loss. 1psi is something like 6% at sea level I think, I doubt anyone would notice that in rush hour traffic 🤣
 
Vetal said:
Hello to all forum members!
I plan to build cheapest possible 500Hp electric turbo for my street V8 (and possibly move it to other engines since it is so vehicle-independent).
But Speedmaster P-2 charger would cost about $1100 in Europe so I don't want to spend that. I want to use compressor wheel + housing from conventional turbo (cheap ebay ones). Like this:
View attachment 837
and somehow direct drive it by a motor.
This will make system much more flexible (able to swap wheel/housings).

I have couple of questions:
1) am I right in selecting the largest wheel possibly? Reasoning is, they usually spin slower, so might be useful at some 30.000 rpm already. I think Alex even with 40KW motor could not spin compressor faster than 32.000rpm.
On the other hand, I might get Castle 1721 2400KV motor (with 33.6V LiPos) which spins to 70-80K in hope that it will spin some SMALL turbo wheel at that speed. But it's only 10-12KW AT MOST
2) Why do we use high voltage and LiPo batteries? Why not use car 13V system? Battery provides current up to 500-700A so that should be enough, and if not - put second car battery.
3) does compressor wheel put some meaningful axial load on the shaft? If so, it would require some additional thrust bearings and complicate the system
4) Not looked into it but I remember gamers use all kinds of liquid cooling their systems, might cannibalize something from that area too to increase power

Guys, please advise :)
Click to expand...

You didn't provide any specs on the compressor and housing you posted a picture of and the proper way to assess the practicality of that combination is to calculate the cubic feet of air (CFM convert to lbs/min) that your motor will require to meet your goal and then plot that value on the compressor map of the compressor wheel you are considering.

I have completed a prototype of exactly what you are planning and can say the proper size compressor wheel and housing necessary to meet the requirements of your goal, that will allow use of a relatively low rpm rotating assembly (35-50k rpm) to work within the parameters of a reasonably priced ESC to the tune of around 2-7 psi and 1 or 2 motors putting you in the 15 kW and up power range... will need to be in the range of a GTX5533r compressor wheel (88/133mm)and housing and since the tariffs hit that combo has gone from ~$500 on ebay to ~$1300, making the P2 or V2 the best option now especially since it is a finished product.

Go to Garrett turbo's website and look up the compressor map for the GTX5533r so that you can see compressor wheel rpm vs flow. It would have been perfect for my 3.6L DOHC motor and the high rpm it's capable of, requiring about ~40k rpm to deliver 7 psi at peak engine rpm, with a base motor spec of 304 hp and 278 lb/ft of torque, 1.5 bar x 304 = 456 hp minimum as that's traditional exhaust driven turbo power where Alex has demonstrated that the efficiency bump is much higher, closer to double as electric boost delivers nearly the same power as an exhaust driven turbo at half the boost pressure.

Here's the link, scroll down and look for the maps next to the wheel inducer options:
https://www.garrettmotion.com/racing-and-performance/performance-catalog/turbo/gtx5533r-gen-ii/

I haven't had any problems so far with cutting a good shaft on my lathe. I saw the thread regarding the problem Alex mentioned and a lot of it is rapped up in design, an extra long shaft having the dreaded over hang outward away from the output end bearing and no nose bearing to counter it. A better design for direct motor drive would be to include a coupling point at the rear for the motor (As Alex has done) so that the compressor wheel is almost flush against the output end bearing. That leaves very little room for wobble harmonics if the shaft is fairly straight.

I've only spun my compressor up to about 18k rpm but that's because I don't have an ESC capable of delivering the 60ish volts I have and when you connect a charged hybrid battery at 68 volts to a 2-6s ESC and manual speed tester, it goes Pop! before you ever get a chance to synch the motor and apply a little voltage to see what happens.

What I did accomplish using practical voltage ~25v was pretty intense and I fully understand why Alex was standing so far away during some of his testing. The videos do not do the experience any justice.
 
Last edited:
You must log in or register to reply here.
Top