Open Source Electric Supercharger

[ciro]

Originally Posted by neutron256

I had some time today so I took a stab at the three motor concept. This is nothing like the power that @rusty959 is planning but it's the same general idea. It's a simple 2:1 gear ratio so nothing outrageous. I'm using 39mm Turnigy SK3 EDF for the motors in this design which would give it a no load output of 72,000 RPM.

I have no plans to actually build it but if anyone else is interested it's a start.

You will burn up those motors when they are linked link that. brushless is not a fan of being linked together. brushed is the only way to effectively link multiple motors on the same gearbox. i cant remember the exact reason but in college our robotics team avoided this kind of setup like the plague.

[rusty959]

Originally Posted by ciro

You will burn up those motors when they are linked link that. brushless is not a fan of being linked together. brushed is the only way to effectively link multiple motors on the same gearbox. i cant remember the exact reason but in college our robotics team avoided this kind of setup like the plague.

Do you think you could see about finding more information about this? I have done searches and can't find any reports of issues, but I may be looking for the wrong thing. I could definitely see some issues with cogging on startup from a dead stop which if done repeatedly could cause issues, but I plan on always running the motors at at least a low speed which would decrease if not eliminate this issue.

Seems like everything related to multiple brushless motors involves trying to run two with the same ESC.

[mid_life_crisis]

In audiophile music they go through a process whereby they make sure that amps of the same type actually produce the exact same output when fed the same signal to keep the sound matched.
Could it be necessary to test the motors to make sure the same input produces the same output? e.g. power, rpm? Given the way the controllers work, the speed should be the same or the motor just won't run, but theory is one thing, reality is often something else.

[spyingwind]

Originally Posted by rusty959

looks good neutron. That could be a nifty setup.


once again im referencing the rc world habits here. With a composite spur no lubrication is needed. With a metal on metal setup some people run them no lubrication with no ill effects and some run a dry lube such as teflon or graphite powder.

ill probably go for dry lubes and see how they work. If I have to lift the hood every time I fill up and relube while waiting on the pump so be it. I'd probably do it for inspection purposes anyways. There are also "permanent coatings available.
I'm not going to mess with oil at all though.

If the motors are balanced and the heat produced from the gears, barrings, and what not are cool enough, then very little is need to keep it from over heating.

I would guess very light oil, and very little oil is needed. Again balance might be the biggest issue.

Never know till it is tried, tested, and broken.

[rusty959]

Originally Posted by mid_life_crisis

In audiophile music they go through a process whereby they make sure that amps of the same type actually produce the exact same output when fed the same signal to keep the sound matched.
Could it be necessary to test the motors to make sure the same input produces the same output? e.g. power, rpm? Given the way the controllers work, the speed should be the same or the motor just won't run, but theory is one thing, reality is often something else.

Could it be necessary? Maybe. I don't think it will though. However, since so many people have brought it up, maybe I should buy some pinions for the 4 motors I have laying around in my room and do a test.
The difference between this and "audiophile music" is with speakers, if they aren't matched cancellation will occur. With the motors, I expect them to self regulate to share the load because without load they would all be spinning faster. It won't be possible for one to be driven by another motor.

Originally Posted by spyingwind

If the motors are balanced and the heat produced from the gears, barrings, and what not are cool enough, then very little is need to keep it from over heating.

I would guess very light oil, and very little oil is needed. Again balance might be the biggest issue.

Never know till it is tried, tested, and broken.

I'm just not convinced oil will stay on the gear at 100kish rpm. Maybe enough of a film would, but I'm not so sure of it. Which is why I would first try a dry lube. We will see. Ill be doing enough bench testing that Ill be able to get some good observation of what happens with some different stuff.

[mid_life_crisis]

Originally Posted by rusty959

I'm just not convinced oil will stay on the gear at 100kish rpm. Maybe enough of a film would, but I'm not so sure of it. Which is why I would first try a dry lube. We will see. Ill be doing enough bench testing that Ill be able to get some good observation of what happens with some different stuff.

I'm picturing an oiler from my shop that has a needle for precision application to router bearings. The oil is very light and pretty much stays put at 30,000 rpm. Trust me, you don't want oil flying off your router when you're working on an expensive piece of figured wood. You might try that as an experiment as well.

[rusty959]

Originally Posted by mid_life_crisis

I'm picturing an oiler from my shop that has a needle for precision application to router bearings. The oil is very light and pretty much stays put at 30,000 rpm. Trust me, you don't want oil flying off your router when you're working on an expensive piece of figured wood. You might try that as an experiment as well.

Thanks Ill look into that. I know ill be getting some light oil for the bearings, it might be a good type to try on the gears as well.

[ciro]

Originally Posted by rusty959

Do you think you could see about finding more information about this? I have done searches and can't find any reports of issues, but I may be looking for the wrong thing. I could definitely see some issues with cogging on startup from a dead stop which if done repeatedly could cause issues, but I plan on always running the motors at at least a low speed which would decrease if not eliminate this issue.

Seems like everything related to multiple brushless motors involves trying to run two with the same ESC.

I couldnt find anything on it but I just remember that being something the team had tried and burned up a bunch of motors. It could have been a variety of things that made them burn up or I am backwards in my thinking and its brushless that was ok and brushed needed to be separated. its been a few years you will have to forgive me if i am wrong.

[rusty959]

Originally Posted by ciro

I couldnt find anything on it but I just remember that being something the team had tried and burned up a bunch of motors. It could have been a variety of things that made them burn up or I am backwards in my thinking and its brushless that was ok and brushed needed to be separated. its been a few years you will have to forgive me if i am wrong.

Alright. Well, I'll probably just set up some tests with the brushless motors I have laying around. I can do fun things like putting different gear ratios on and forcing them to run different speeds and such.

[neutron256]

Sorry there haven't been any updates recently, I've been dealing with health issues and waiting for parts to arrive (I love McMaster-Carr).

I have been starting to test motor/controller/battery setups. One note for those who are planning to use the charge in parrallel/discharge-in-series concept. Many (maybe most) of the RC brushless controllers have a startup delay. For the one I'm using this is about three seconds. Mine also won't start the motor until it sees a neutral input signal.

Three seconds is nothing if it powers on when you start the car, but if it only gets power when the system activates then it could be an unacceptable delay. The good news is that the capacitors in the brushless controller can keep it from rebooting if power is removed and restored very quickly (like a relay activating) so this might help with a work around for this issue.

[rusty959]

Thanks for checking on the quick power removal, at least we know one controller can handle a quick period without power. I still need to be careful about finding one that can handle not only short times without power but also a 3x change in voltage.
They aren't the controllers ill be using in the end, but the ones I have laying around have the same neutral input/3 second thing going on as I remember from using them previously.

Ive bought some fun stuff to assist in my testing etc, such as some 1/4 inch aluminum, aluminum rod for the motor/gear adapter, and some gears. Assuming everything is here by weekend as it should, I should have a chance to get on a mill over the weekend and make a test setup to throw some motors on. I also picked up an infrared temp sensor which may be of use.

I need to get started on my cnc mill soon, thats a big project in itself and is what is going to hold up my e-charger setup from getting very far. But, Ive been busy as most people are.

[rusty959]

Alright so Ive got some good news and bad news.

The good news is that I got some parts in the mail:





The other good news is I may be buying a house a bit sooner than expected, so ill have a better work area.

The bad news is I may be buying a house a bit sooner than expected, so I need to be putting as much money as I can into things like a down payment and other necessities, which hinders this project quite a bit. Ill be continuing what I can with this though, we will see what happens. I'm still planning on setting up that test this weekend.

[wulfgang]

This is an interesting topic. I'd like an electric supercharger for DE days, but the math doesn't work out for a lap-averaged energy draw (the alternator doesn't make enough power even if it operates at 100%). A really good alternator might be able to give 2 kW average power, so if you are on the throttle roughly 50% of each lap, then you have a max of 4 kW to drive a supercharger. So not very much.

With regard to the gears, keep in mind that some superchargers are gear-driven and they have a lot of gear whine. Then there are belt-driven units that can be unreliable because of the belts.

...and then there is Rotrex, with their cool non-newtonian fluid in a planetary traction drive. If you had the money to fabricate parts to direct drive a Rotrex, then your electric motor options would greatly increase because you could drive at 10-15 krpm and still have the compressor wheel turn 90 krpm. None of the other superchargers even come close to this speed, so it seems likely that a complex Rotrex-like traction drive is necessary to get such speedups on the drive.

You might also want to look at sourcing the compressor wheel from a supercharger rather than from a turbo. Although in principle the wheels should be equivalent, it seems like the supercharger manufacturers get the maps that have the plateau a lower speeds.

[wulfgang]

Originally Posted by solma

Third, be careful about technical spec on component, MOSfet are not good for most motor control appication Ton-Toff are normally to long(so they heat up to much), use IGBT its like MOSfet but design for this kind of application.

The IGBT always has a 1.8-2 V junction drop, so they only make sense in higher voltage designs. If you drive your motor at 12 V, then the IGBT will lose 2/12 = 17% efficiency right off the top, before switching losses. At 24 V, it's 8%, so still quite a bit.

Yes, MOSFETs switch slower, but as of around 2005 it is not really an issue below say 10 kHz, and at low battery voltages they have much lower voltage drops than IGBT. This is why MOSFETs are still used in so many ESCs. Switching noise will most likely not be an issue, especially if you drive at 20 kHz, and there is no point in going to higher frequencies for this project.

The bottom line is that you should let the design dictate the details. If you want to drive at 70 V, then maybe an IGBT looks pretty good, but if you are driving at 20-30V, you can use automotive grade MOSFETs and get higher efficiencies.