I have designed an electric supercharger consisting of a Rotrex C30 driven by a 14HP electric motor. The blower and motor are installed inside a stock air box. Seemed like a fun and different project a year ago when I started. It uses a VESC speed controller with an Arduino controlling the boost. My goal is a flat 6psi boost curve from 4k rpm to redline. My bench testing shows it should work. I managed to fit a water to air intercooler between the air box and front of the engine. There is an 80V LTO battery in the trunk. There are more details on Youtube. For my first test drives, I turned the boost down to 2psi (using the Arduino controller), and I'm running straight 100 octane Sunoco unleaded, for now. The A/F is running around 12.0. I have an FMU for fuel and the MAP sensor is clamped, so I'm trying to use the stock ECU. One of the advantages of this setup is it leaves the car almost complete stock - fairly easy bolt on stuff. The blower seems to work fine. But the ECU is apparently going berzerk when the electric motor starts. At 6k rpm in 3rd gear, the timing is reported to be about 24 degrees with the motor off, but it drops to around 8 degrees when I switch on the supercharger. The clamped MAP sensor is reporting +1 psi of boost with the blower on, and about -1psi with it off. There is no bypass valve so the blower is a restriction when not running. Obviously, the massive timing reduction is killing power. All I can think of is the knock sensor is confusing the supercharger noise with knock and killing the timing. I've heard the AP1 knock control is very basic and sensitive.

Anybody have any thoughts or suggestions?

 

You mentioned some data on timing which does seem to suggest knock control is taking over. On 100 octane I honestly doubt you would see real detonation at 24* even with 6 psi at 6k rpm. Do you have any further data logging you can share?

 

I've only done 2 test drives and this problem just jumped out at me. I simply switched the motor off and on and watched the timing on the obd scanner. I was also watching IAT as the car heat soaked the sensor. The timing might have moved a couple of degrees with IAT, but that is minimal in comparison. I'm thinking the next move would be to disable the knock sensor and see what happens. Should be zero risk at only 2psi with 100 octane. I can either unplug the sensor (which is hard to reach), or short it out with a capacitor near the ecu (easy to do). I'll report back tomorrow hopefully.

 

It seems doubtful its noise from sc. When kits bolt the same sc directly to block, there aren't knock sensor confusion issues.

Yours is fully isolated from block, from sensor.

Something else is pulling timing I think. Will be interesting what disabling knock sensor reveals.

 

I found the problem. It was not the knock sensor. When I disconnected it, I got a CEL and the ecu pulled about 8 deg of timing (with or without the blower running). Then when the blower was enabled, timing dropped another 14 deg or so. So now the ECU is pulling timing for 2 reasons. I hooked the knock sensor back up, then changed the MAP clamp from 3.1V to 2.9V, and the timing problem went away. I read somewhere 3.1V was the correct number, but it's too high. I'm using a very precise clamp. I'm surprised the ECU pulled all that timing because of a high MAP sensor voltage and didn't throw a code. The MAP sensor is now reporting 14.5 psia under boost (I'm at sea level in Florida). The boost is running about 2psi from 4000 to 8000 rpm, which is as high as I have run it so far. The FMU quit working, so I'm not going to 6psi of boost until that is straightened out. I have a little extra fuel because I faked the ECT sensor into thinking the car was running a little warm, which gives about 10% more fuel and pulls a couple of degrees of timing. The ECT tweek is active only when the blower is making boost.

My early impressions of an electric blower on an s2000 are all positive so far: 1) I was worried about how it may sound, as unlike an engine driven centrifugal, this one runs at pretty much constant speed as engine rpm rises. I had to have ear protection when I was bench testing it. But in the car I could barely hear it. Intake and exhaust noise completely dominate blower noise at 2psi. 2) For only 2psi, the car feels stronger than I expected. The mid range pull through VTEC feels really good. The engine doesn't drive the blower, so there is some extra crankshaft power. 3) unlike a normal supercharger, this one is not turning until needed, so it takes a little time to spool. I didn't feel or hear the blower turning on, it was very normal. With more boost, it may be more noticeable. The turn-on can be tuned in software. 3) I drove around for about 20 minutes and did maybe 20 pulls of 2 or 3 seconds each. Typically in 3rd gear from 4500 to 7500 rpm or so. The 84V battery voltage didn't move at all. The charger runs all the time and kept up with the power drain from the small pulls. The battery can run down to as low as about 74V. 4) I saw about -2psi of boost at high rpm's with the blower off. I'm thinking this is probably ok and I won't need a bypass valve. There will hopefully be enough battery capacity so that running without boost will be rare, and -2 psi should be ok occasionally. 5) IAT barely moved during boost although its only 2psi. My gauge runs on its own sensor before the TB that shouldn't heat soak as bad as the stock one.

I would appreciate your comments and suggestions. I don't know of anyone else running this configuration, so this is all new to me.

 

Cool project

 

Super cool project! Love the creativity of going electric on the Rotrex. Glad you sorted the MAP clamp voltage issue - that was a good catch.

Some thoughts as you move toward 6psi:

Regarding the battery / power, your setup is probably good for 2-3 minutes of continuous WOT before the battery gets low. For street driving with short pulls you'll be fine, but track use might be tough. Consider adding real-time SOC display so you know what you're working with.

The -2psi restriction when the blower is off is costing you 10-15hp - might be worth adding a simple electronic bypass valve controlled by your Arduino.

Honestly, ditch the FMU before going past 4psi. They're just too unreliable for this kind of setup. A Haltech Pro Plug-in would let you tune it properly and integrate your wideband for closed-loop control. You're basically at the limit of what the stock ECU can handle safely. At minimum, grab a proper wideband (AEM UEGO or similar) and keep it visible while testing.

For testing, I'd go 2→3→4psi in steps, monitoring AFR religiously. If you stay on the stock ECU past 4psi you're playing with fire even on 100 octane.

Also add some Arduino safety logic - cut boost if battery drops below 74V, if IAT spikes, if AFR goes lean, etc. Also think about boost-by-gear limiting to stretch battery life.

The concept is solid and honestly pretty brilliant for avoiding parasitic loss. Just needs proper fueling/tuning infrastructure to be safe at higher boost. Keep us posted on the progress!

What's your long-term plan - street or track weapon?

 

Constant sc rpm regardless of engine rpm...what rpm is blower running at?

Maybe use compressor map, determine its most efficient rpm, run it there most of the time, so less iat increase. Then above a certain engine rpm, ramp it up as close as you can to sc redline for max boost (C30 that's what, 110k rpm?)

So like 2 stage rpm, if that is possible. Unless you're already at electric motor max rpm?

 

I'm controlling the blower rpm with an arduino to get a flat boost curve, with 6psi the goal. The blower impleller is running about 55k rpm, with the input shaft about 6k rpm. This gives about 6psi of boost at lower engine rpm's, but the blower rpm has to increase to about 60k rpm at higher engine rpms to keep the boost constant. The rpm has to increase as the C30-94 efficiency is starting to drop. The boost is progressive with throttle and the above numbers are at WOT.

 

I have all kinds of real time data to look at. I am sending boost, IAT, A/F, and Fuel Pressure to my cell phone which displays it as a gauges on a virtual dashboard. Uses the Android App "RealDash". The actual Wideband is tucked away under the dash and sends data to RealDash through an arduino. That way I don't have gauges stuck everywhere, just my cell phone in a holder. My cell phone can also bluetooth to the speed controller and the BMS, where i can access tons of geeky information including battery status. My plan is to add battery status to RealDash so it is always there, but there is some work to get it there. Also, in driving around the last couple of days, the battery voltage has barely moved. It recharges as I drive.

I've thought long and hard about replacing the ECU, but I'm being a little stubborn about keeping the car as stock as possible. Right now, I can return the car to stock in a couple of hours. As I understand it, a Haltech requires quite a bit of rewiring. Is that true? I drove the car today at 4psi and was amazed at the transformation, especially at mid range before and during VTEC. The car was actually fun below VTEC. This is close to 5psi for an engine driven supercharger. I think I'll keep it there for a while and decide what to do later.

Good idea to add some safety logic. Been thinking about that. Its coming.

This is a pure street car. I have had big HP blown cars before. I'd like to keep the character of the s2000, just enhance it some.

Thanks for you comments and suggestions!