Open Source Electric Supercharger

[rusty959]

I went ahead and snagged some pics of my wiring that I have, just so you guys have a good idea.

You can barely see the wire here. Its sitting just above floor level on the side.


This is up under the dash, passenger side


Coming out into the trunk


Going through the firewall, left of battery


I don't have it hooked up ATM because my stereo isn't in. (I pull it out at times for autocrosses and the like)


Give me a bit and I'll try my paint skills at a wiring diagram, and explanation of how I'd wire it.

[rusty959]

Originally Posted by AN

IMHO , I think it will be a good idea to keep the dump pack in the hood and move the starter battery to the trunk ...

For a simple setup, I'd agree with you. The more complex you get, the more reasons there are for putting it in the trunk. Such as size, bunch of wires, heat of engine bay, etc. The benifit of a trunk setup is there is literally nothing more in the engine bay than a wire. Just hook it strait to the esc.

Originally Posted by AN

I thought we need to make use of diodes to balance charge each 12v Capacitor pack to balance charrge each of the 6X 2.7v .

Could you please elaborate the wire design? Thanks ..

However you do it, you will need a balancing circuit. If you do it with diodes, cool. But something needs to be there.

Ok so I made a diagram. This is most likely how I would start if I were to do this setup right now. Its crude, in paint, a little busy, etc, but it gets my point across. Any questions just ask.
I don't have fuses in there, but they are definitely needed.
Line thickness roughly correlates to size of actual wire needed.
Obviously some things such as connection points on a node can be moved slightly with no effect.
This is about the max current I'd suggest flowing before running a dedicated ground line from the front as well. Id potentially suggest going ahead and isolating it, as it is possible for weird things to happen.

[AN]

Yep, Thats pretty much what I had in mind. However, I believe, I can shuv 4x 21Ah Powersonic batteries in the stock battery area with a larger tray on my Miata. With a 3d printed battery cover, it should look clean. I will have to be creative with MinnKota charger location. I still hope we can find a more compact charger. I wonder if we can have someone custom build one ...for a reasonable price.

Also, I am dropping the idea of the capacitors for now. Any reasonably sized boostcap pack will hold very little energy .. So its better to keep things more compact.

With the eCycle motor being kind of pricey ... I think I will go with the Leopard motor (5500W) will give me 38k rpm.

Not making much progress with the compressor / wheel selection. I know Phantom is using Mitsubishi TD06 housing ... not sure which wheel .. Largest wheel available for that housing on 25G.

[neutron256]

Originally Posted by AN

Yep, Thats pretty much what I had in mind. However, I believe, I can shuv 4x 21Ah Powersonic batteries in the stock battery area with a larger tray on my Miata. With a 3d printed battery cover, it should look clean. I will have to be creative with MinnKota charger location. I still hope we can find a more compact charger. I wonder if we can have someone custom build one ...for a reasonable price.

Also, I am dropping the idea of the capacitors for now. Any reasonably sized boostcap pack will hold very little energy .. So its better to keep things more compact.

With the eCycle motor being kind of pricey ... I think I will go with the Leopard motor (5500W) will give me 38k rpm.

Not making much progress with the compressor / wheel selection. I know Phantom is using Mitsubishi TD06 housing ... not sure which wheel .. Largest wheel available for that housing on 25G.

I think Rob is machining the TD06 housing to accomidate wheel from a larger compressor. I don't see it there anymore but at one point there was a pic oh the phantom site of a compressor housing on a lathe. Of course that could have also just been something he was testing.

This is what makes me think the SeaDoo supercharger compressor may work great. It's a much larger wheel which means slower speeds but more torque are required. I wish my parts/tooling would get here. I'm excited to start bench testing.

Edit: I would suggest possibly using a Garrett T3/t4 compressor. In the Garrett style turbos the back plate to the compressor is separate from the center bearing hub. This would allow you to adapt it rather then having to machine a completely new backplate. Coupling the motor to the compressor wheel in a solid compact design has been the trickiest thing for me, but I have some new ideas on that if you're interested. I also have CAD models I've made of the TD06 and T3/T4 back plates if anyone is interested.

[Target70]

Originally Posted by neutron256

This is what makes me think the SeaDoo supercharger compressor may work great. It's a much larger wheel which means slower speeds but more torque are required...

would the larger compressor/wheel add latency? Shiv has tested the phantom tq250 and the tq300, and found the 250 to be much more responsive.

[neutron256]

Originally Posted by Target70

would the larger compressor/wheel add latency? Shiv has tested the phantom tq250 and the tq300, and found the 250 to be much more responsive.

There will almost certainly be more latency/lag with a larger wheel, but this also depends on the motor being used. These motors get to full speed very quickly even under load and I suspect there is more to the issues Shiv was seeing with the TQ300 then just lag.

I think a big key to mitigating this is variable seed control of the motor, rather then the full throttle all or nothing system that the Phantom system uses.

-Intelligent speed controllers that can sense and adapt to load conditions and do things like soft starts

-Closed loop control to reach a target boost value rather then max boost possible

-Starting partial boost at partial throttle (say 80%) and raising as throttle increases.

The main reason I started this thread is because people were wanting to experiment with more advanced control and power features like this. I'm hoping that in terms of the motor/compressor hardware we can come up with a really solid easy to build design that people can then use to experiment with.

[Target70]

Originally Posted by neutron256

There will almost certainly be more latency/lag with a larger wheel, but this also depends on the motor being used. These motors get to full speed very quickly even under load and I suspect there is more to the issues Shiv was seeing with the TQ300 then just lag.

I think a big key to mitigating this is variable seed control of the motor, rather then the full throttle all or nothing system that the Phantom system uses.

-Intelligent speed controllers that can sense and adapt to load conditions and do things like soft starts

-Closed loop control to reach a target boost value rather then max boost possible

-Starting partial boost at partial throttle (say 80%) and raising as throttle increases.

The main reason I started this thread is because people were wanting to experiment with more advanced control and power features like this. I'm hoping that in terms of the motor/compressor hardware we can come up with a really solid easy to build design that people can then use to experiment with.

It sounds like you will need a pretty complex motor control unit, needing a variable input/output table, and either a pressure sensor or tie in to the MAP sensor to read and adjust the output, and either an integrated or dedicated charger. I also don't know what type of tuning would be required to be able to adapt to all these variables, most of this is above my pay grade. All I can say is more power to you, and everyone involved in their own experiments, you guys can't do anything but advance the technology. Unfortunately I don't have the equipment or knowledge experiment myself, but I'll keep watching, looking forward to whatever you come up with. Really all I would want is to see a constant psi (say 5) held through the whole rev range with a smooth analog output, instead of a sine wave of off/on pulses.

[AN]

Originally Posted by neutron256

Edit: I would suggest possibly using a Garrett T3/t4 compressor. In the Garrett style turbos the back plate to the compressor is separate from the center bearing hub. This would allow you to adapt it rather then having to machine a completely new backplate. Coupling the motor to the compressor wheel in a solid compact design has been the trickiest thing for me, but I have some new ideas on that if you're interested. I also have CAD models I've made of the TD06 and T3/T4 back plates if anyone is interested.

Hi, could you please point me to where I can pick up a T3/T4 housing ? couldnt find it at Kinugawa ... Also, whats a good resource to find out all the wheels that will fit that housing?

Thanks

[neutron256]

Originally Posted by Target70

It sounds like you will need a pretty complex motor control unit, needing a variable input/output table, and either a pressure sensor or tie in to the MAP sensor to read and adjust the output, and either an integrated or dedicated charger. I also don't know what type of tuning would be required to be able to adapt to all these variables, most of this is above my pay grade. All I can say is more power to you, and everyone involved in their own experiments, you guys can't do anything but advance the technology. Unfortunately I don't have the equipment or knowledge experiment myself, but I'll keep watching, looking forward to whatever you come up with. Really all I would want is to see a constant psi (say 5) held through the whole rev range with a smooth analog output, instead of a sine wave of off/on pulses.

Closed loop control really isn't that complex, and doesn't require tables. I will be taping into the stock MAP sensor for feedback. Then the controller just has to adjust the motor speed up or down based on whatever the target boost is.

As long as boost iis limited so it doesn't exceed the range of the stock MAP sensor tuning isn't that complicated either. The ECU has the data it needs already. Tuning will mostly be to maximize performance.

I'm using a trollng motor battery charger from MinnKota which so far seems to work perfectly for this application.

[neutron256]

Originally Posted by AN

Hi, could you please point me to where I can pick up a T3/T4 housing ? couldnt find it at Kinugawa ... Also, whats a good resource to find out all the wheels that will fit that housing?

Thanks

You probably won't find just a T3/T4 housing and wheel. I had to get a complete Turbo off eBay. They can be found for as little as $100. I usually wouldn't say this but go cheep. The quality you'll get from going more expensive is mostly going to be in the bearing which is useless for our purposes. From what I understand unless you get into the really high end billet vs. cast wheels there isn't much difference in the housing/wheels.

[neutron256]

Thanks to a bout of insomnia I did a quick flow chart of my basic concept for closed loop control. This is very simple and all the inputs are 0-5V so it should be very easy to program on most microcontrollers. The output of course would be a simple PWM server signal.

[Target70]

The last time I looked at programmable micro controllers (~12 years ago) you did everything with ladder diagrams, and they were not very cheap. I have no idea where things have gotten to in that time but if you are able to get OTS components, and just hook them up, that could save time and money on a custom motor controller, plus have the added features.

[SloS14]

Originally Posted by Target70

The last time I looked at programmable micro controllers (~12 years ago) you did everything with ladder diagrams, and they were not very cheap. I have no idea where things have gotten to in that time but if you are able to get OTS components, and just hook them up, that could save time and money on a custom motor controller, plus have the added features.

You would be amazed. You can get a capable MC for about $15 now. I've got three of them laying around.

[neutron256]

Originally Posted by Target70

The last time I looked at programmable micro controllers (~12 years ago) you did everything with ladder diagrams, and they were not very cheap. I have no idea where things have gotten to in that time but if you are able to get OTS components, and just hook them up, that could save time and money on a custom motor controller, plus have the added features.

They have come a long way and are very focused on ease of programming. Most inputs/outputs need little or nothing in the way of supporting components.

Originally Posted by SloS14

You would be amazed. You can get a capable MC for about $15 now. I've got three of them laying around.

I favor the PicAxe which can be had for a few dollars for the most basic versions. It's not as powerful as something like an Arduino, but for most projects I think an Arduino is overkill.