Das Beast: My E30 track / street build

[dvallis]

A bit more bodywork before Rob gets back.

Air dam turned out Ok. I'm getting pretty good with the cup gun.



Rain gutter plastic work shot with final coat. I like this paint.



Riveted on the hood pin metal plates. Gorilla glue wasn't getting it done.



BMW emblem modified for hood pin.



Some aluminum fab work to set the pin.



Hood complete. Next time work on the car it will be startup prep.

[dvallis]

Had a bit of time today to get rolling on engine start.

Frigging radiator drain plug was interesting to figure out. There wasn't one in the rad. JEGs insisted it was 1/4-20. However, the actual hole was ~0.50 in in diameter (i.e. NOT 1/4"). We tried a 1/2-20 plug. Wouldn't thread. Tried a couple of metric sizes, but the rad is US spec so that didn't make sense. Ultimately the drain hole turned out to be threaded 1/4 --->NPT<---. Diameter is 0.54" with 18 threads per inch. Ugh. Thanks JEGs. This thing will do for now.



Got the thing back up on jack stands. And swapped the wheels. None of you noticed they were on backwards. (LoL) Pay attention to direction of rotation.



Filled the diff with Redline 75W90NS. You're supposed to use gear oil without the NS suffix for limited slip diffs, but these cars are old. Better to have a bit more lock up.



Rear sway bar hooked up and torqued. All suspension corners adjusted back down to lowest setting.



Filled the transmission with Redline MTL 75W80. Transfer pump came in handy getting to the fill plug.



Engine oil and water next ....

[ForcedFirebird]

The "NS" diff fluid causes the clutches to chatter. Don't be surprised when you pull it out and turn hard, very noticeable at low speeds.

[dvallis]

Crank It

Had a bit of time today for startup prep. We pulled the oil pump access port on the block, removed the drive gear and hooked up a drill to the oil pump.



Added 1/2 bottle of startup additive ...



Then 5 quarts of 0W30 and spun the oil pump.



Obviously the drill doesn't make a lot of oil pressure, but we could see oil slowly pumping from the head distribution line.

We just cold cranked the engine for a few cycles to check if everything looked good. Click the picture below for a video.



So far, so good.

[wworm]

o0o0o0ooo00ohHHhHhhHh mmMMMAaaAAAAAnNNNn1!11!!!!!!!!

cant wait cant wait cant wait

[dvallis]

Guys

Can someone share a Megasquirt PNP2 Gen2 M20 tune for wasted spark? I've gone through all the TunerStudio screens and think they're right, but would feel better starting from a known good tune.

Thanks in advance

[dvallis]

Moving this discussion back from a thread in the ECU forum.

I finally found the definitive answer on setting up a MSPNP Gen2 V1.3 with Quadspark for wasted spark here.

There is a typo on their web page. Says DB15 but this is actually what is looks like on a 1.3 board. It's a DB26.



With "Spark A" set to IGN1, and number of coils as "Wasted Spark" TunerStudio drives Spark A, B and C out on the DB26 option connector pins 17, 16, 15 as I have it wired.





So .... everything should work this weekend. Will verify with Test Mode first.

Fingers crossed.

[dvallis]

So, before starting car the we fired up Tunerstudio in test mode. Just in case ...



Good thing we did. It melted the coil and blew up the Quadspark.





There could not have been that much current flowing, since the coil base conductors did not melt.



At least the Megasquirt survived.





There's only one damn explanation: This page by the manufacturer is wrong. Specifically this graphic. "Spark Output Going High".



That means the external coil drivers are supposed to be open circuit with their trigger input LOW and connected to ground when HIGH. Current is supposed to flow BRIEFLY from the battery, through the coil, through the Quadspark to ground. The coil current must have been on 100% once I pressed Enable Test Mode.

I found THIS little gem buried in the question mark next to Spark Output. Notice its qualified by "typically". Thanks Tunerstudio for a totally cover your ass comment. Fits with my analysis of why it blew up though.



I've got an email into DIYPNP to resolve this: Are Quadspark trigger inputs active high or active low? When I get a new Quadspark I'll test the damn thing on a bench with scope, logic analyer and probes. Will also be testing the MS2 Spark A/B/C outputs the same way. Grrr.

To be continued ....

[Digitalwave]

This is the point where opensource stuff starts to get fun...

[dvallis]

Fortunately debugging this is easy compared to my day job.

[citizen_insane]

You may need to confirm how your PNP board is wired and where the spark outputs are coming from.

Based on the documentation for the QuadSpark, "Going High" is the correct setting for MS2 https://www.diyautotune.com/support/...on/megasquirt/. This makes sense to me for two reasons:

1. Based on the documentation I've read, the BIP373 collector to emitter switch closes when it's given 5 volts. This means that "Going High" will switch the coil to ground (which is exactly what you want)
2. The QuadSpark is basically 4 BIP373's in a convenient pre-wired package

I've found the schematic here helpful, although note that the 330ohm resistor is wrong, it should be 100ohm (it works with 330s because that's how mine's built right now, but I need to correct it at some point).
http://www.msextra.com/forums/viewto...?f=131&t=42038. My understanding is that the resistors are not necessary for the QuadSpark as it has them internal, but I couldn't find a full schematic of the device.

You should be able to tell what the ECU is going by setting the dwell to minimum and probing the spark output. "Going High" should basically just read 0V (since the dwell is such a low duty cycle). "Going Low" should read 5V (I think). Alternatively hook it up to a JimStim on the bench, bump up the dwell and wire one of the LEDs to the spark output. You should be able to tell which direction the duty cycle is going.

Found this helpful as well, just a confirmation what "going high" and "going low" actually mean: http://www.msextra.com/forums/viewto...=53734#p398118

[dvallis]

Thanks for the detailed input. Megasquirt tech support confirmed:

1. The Quadspark FETs close on logic high 5V

2. There is a MSPNP2 internal jumper J12 which has to be set to SEQ for wasted spark to work right.

3. Their comment in the yellow bubble about going low is for MS2 not MSPNP2.

Will be testing all this on a bench with my scope. Only way the coil & Quadspark could have burned is if I didn't power cycle before entering test mode.

[Nader393]

[dvallis]

LoL. I resemble that remark.

[dvallis]

Das beast electrical engineering lab

Right. Got the garage set up as an electrical engineering lab.

First testing the new Quadspark. Light bulb standing in for coil pack. The trigger is definitely "going high" i.e. 5V on the gate closes the FET.



Bulb on resistance gets a a little low, something like 24 ohms. Swapping that out with a 1.8K resistor gives more reasonable 7 mA current. We'll use this as a test coil later.



THIS is why my coil burned. You're looking at the Quadspark coil A trigger output from pin 17 of the option connector with MS in test mode on the yellow channel. Voltage scale is 5 V/div. Time scale is 50 ms.div. The signal is SUPPOSED to be a 3ms pulse every 200 ms. Clearly it's not. This would look like ON to the Quadspark, thus frying itself and the coil in the process.

There's a software error in Tunerstudio or MS2PNP.



Re-flashing the MS I got this waveform in test mode. Looks like it should. Coil B (green) and C (blue) also behave correctly.



This worries the hell out of me. When will Tunerstudio or MS choose to fry my coils again? At the very least I need a fast acting 10A breaker on each coil pack drive circuit.

Hmmm.