I (Nathan's dad) built the MS board pictured above. Nathan built the rest of the Lotus 7 starting with a load of steel tubing, a scrap Miata from the police impound lot, and a welder. The Lotus has a '94 1.8L Miata engine in it. Right now the engine is pretty much stock, except for the turbo. The engine runs fine with the stock ECU, and the turbo isn't doing anything at the moment. The next thing we will try is to get the engine running with the Megasquirt using the stock coils and injectors.
I built the board according to the megamanual, then modified it using the Megasquirt 2 assembly instructions at miataturbo.wikidot.com. CMP comes in on JS10, CKP on XG1, and OptoIn is connected to VCC. I put the boost control FET at Q16, driven by JS11, and it's output is wired to IAC1B. Q7 was replaced with a 2N2222 and is wired to IAC2A for fan control. We used Joe Perez's ignition inverters driven by the D14 and D16 drivers. The board talks to TunerStudio and has been tested with a JimStim and a scope. I am now in the process of designing the wiring harness modifications prior to installing the Megasquirt in the car. This will be a standalone installation.
I became confused when some instructions showed the ignition signals crossed and others not crossed. I decided to not cross the ignition, so D14 inverted goes to IGN which will drive coils 1&4, and D16 inverted goes to IAC2B which will drive coils 2&3. Assuming the falling edge of CKP that occurs during CMP high indicates cylinder 1 TDC, then we appear to be getting the correct firing pulses.
When I started to consider the various injector wiring diagrams I could find on the web, seeing INJ1 driving injectors 1&3 and INJ2 driving 2&4 initially appeared to be an error. I even found one harness drawing that shows INJ1 driving 2&3 and INJ2 driving 1&4, same grouping as the ignition. I spent several evenings trying to figure out what the MS is doing with the injectors, and why. It appears that the default setup (batch injection) is to fire each bank once per cycle, not necessarily synchronized to anything, which reminds me of the CIS injection on my '74 Audi Fox. In this case it may not make a lot of difference how the injectors are wired up.
I'm still experimenting with the JimStim and the scope. I tried setting the MS up to do semi-sequential injection. What it appears to be doing now is firing injectors 2&4 during the exhaust stroke of 1, and injectors 1&3 during the compression stroke of 1. Here is a drawing (using Excel) of what I'm seeing on the scope:
Here is the same data as a polar plot. Each quadrant of the circle is one stroke. The upper right quadrant is the power stroke of cylinder 1.
Here is a chart of what is happening in each cylinder on each stroke. Red is when the injector is squirting, orange is the fuel just sitting around in the manifold, yellow is when the intake value is open and the fuel is pulled into the cylinder. Red and yellow stripes is when the injector is squirting into an open intake. Asterisk indicates plug firing.
This is with the engine loafing at 1500 RPM, BTW. I realize the squirts will get longer (start earlier?) at higher loads.
if I swap the injector wires, then injectors 1&3 squirt during the exhaust stroke of 1, and injectors 2&4 during the compression stroke of 1, and I get this sequence:
This setup seems more symmetrical. Whether it is better, I can't say. From what I've read, the engine will run either way, and we can experiment and see if one arrangement works better. Maybe we should use barrel connectors in the injector wires so we can swap them more easily.
One thing we would like to find is a plain vanilla msq file for an unmodified Miata engine. We have an MS3 CPU daughterboard, and most of the example msq files we found are for earlier boards and can't be loaded into the MS3. The only example I could find that would load into our board was for a modified engine, and I'm not sure how appropriate it is for a stock engine.