Megasquirt sequential injection

[Lex]

Perhaps the Link does a single fire per 720 degrees which is a little extreme. The MS uses 2 drivers to fire 2 cylinders at a time. Most of the time, a valve is not open long enough to accept all the fuel a cylinder needs so fuel is sprayed in the back of the valve anyways - firing it when when it's not on the compression stroke is not an issue. That being said, there are better combinations than others to start injection on a batch fire and I have worked these out. If MS could always start injection on the same cylinder (while still in batch) it would be great and IMO all that is needed even for really large injectors above 550ccs in a 4 cylinder. I passed emissions without issues myself on batch fire and large injectors on MSI.

The EGT/Pro disty has a dual ignition signal - one for every cylinder fire, and one that happens once per engine cycle. It would be great if MS took the second signal and always started injection when it saw that as apposed to when it saw ANY of the other 4 signals, making it predictable and allowing me to arrange my injector firing and wiring as is optimal. This is what the stock ECU does. However (to my knowledge) I can't set it in such a way.

It is best to wire your injectors such that 1&4 are injected together and 2&3 are injected together. It is not like this stock.

[323driver]

So if we use just that one signal from the disty, how do we know when to fire a coil pack in wasted spark mode? I mean for a disty, that's not a problem but would't you need something else for coilpack control?

Also, how do we know which bank of injectors to fire at startup, using what?

[Lex]

So if we use just that one signal from the disty, how do we know when to fire a coil pack in wasted spark mode? I mean for a disty, that's not a problem but would't you need something else for coilpack control?

Also, how do we know which bank of injectors to fire at startup, using what?

The miata CAS has a 4 signal wheel and a 2 signal wheel setup for wasted spark. You can run wasted spark on a 4-1 CAS wheel as well by selecting a dual disty setup - however I was referring to injection onset not ignition.

[323driver]

Ok, I'll check all that as soon as I'm done with finals and will try to get some ideas on how to work with this. Basically, what I'm interested in is sequential injection and COP ignition. Do you use msn messenger?

[Lex]

I do .... but I'm not on there too often. Remember that MS I and II only have 2 injection drivers and I don't think you can get 4 ignition drivers on one board.

MSIII is supposed to have 8 of each. However, new developments are made often so it's good to stay on those forums msextra and msefi.

[323driver]

I wanted to buy the MSII in summer but decided to wait for future expansions, that's why I'm waiting on buying it, hoping the new one comes out. In spring, I will need one for sure though so keeping fingers crossed as designing one by myself will take me alot of time and ressearch and I'm not sure I have these. We'll see how the MS developement goes.

And yes, the MS has 2 drivers but you don't have to use them, you can get the signal out to a daughter board with 4 drivers, just like they do with the router board or the sequencer.

[Lex]

Is the I/O available on the microcontroller for it?

[323driver]

The driver output is available at the board I/O but you could just remove the driver and connect the traces with a wire so to get the injetor signal directly to the output. What they do with the router is different, they take the driver output directly.

[Lex]

I didn't mean that - does the MCU have enough IO ports and are they supported in the firmware?

[matt_fulghum]

The MCU doesn't have enough PWM outputs to handle peak and hold sequential injection (it could handle sequential fired saturated injectors though) but really... it's NOT WORTH IT

Unless you wanna drop thousands of dollars at a reallllly well equipped tuning shop you're not gonna be able to tune the car accurately enough to make a real difference with sequential injection. Trust me.

Oh, and the wasted spark bit is easy. I'm putting that in BPT323's box. You just hook up both the Home and Trigger wires on the disty. The Home wire triggers before cylinder 1 fires, so it just resets the cylinder count then. Really straightforward. All you need to do it is an optoisolator and a couple resistors. Hell, you don't even need that, but it's a good idea just so you don't somehow blow out the chip in case your disty gets zapped or something.

[matt_fulghum]

I wanted to buy the MSII in summer but decided to wait for future expansions, that's why I'm waiting on buying it, hoping the new one comes out. In spring, I will need one for sure though so keeping fingers crossed as designing one by myself will take me alot of time and ressearch and I'm not sure I have these. We'll see how the MS developement goes.

And yes, the MS has 2 drivers but you don't have to use them, you can get the signal out to a daughter board with 4 drivers, just like they do with the router board or the sequencer.

No new versions of the MS are planned for at least a couple years FYI, and the router board is NOT coming out soon (they haven't even made alpha boards yet, so beta testing is a whiiiile away).

The GPIO board is mostly ready, in fact they might sell it already. It's not terribly useful at the moment though, so hold off on it.

You're an EE though right? If you need to add anything to the MS, it's pretty easy to do. The MS-II supports the CAN bus and you can get chips that support it for cheap. Microchip also makes a translator board that converts over to SPI or I2C or one of those busses. The AT90CAN128s from Atmel have the CAN stack built right into the chip. Easy peasy.

[Lex]

Matt, how well documented and easy to follow is the MS code right now?

[matt_fulghum]

I wouldn't bother with the normal MS2 code... it requires the CodeWarrior suite to compile, which costs monies. The MS2/Extra code is based on the GCC tools, which are free

Also the MS2/Extra code is a bit more featureful. I'd start there if you're interested in hacking the MS.

[323driver]

Sorry, didn't get your question. Since we can figure out when to fire the first cylinder, wouldn't it be easy to make COP ignition? From what I understand, we are limited by the number of I/O pins, am I correct?

P.S. Canbus or SPI as it is very simple, is the way to go I think. I had bad experiency with I2C. Oh man, I still remember the microprocessors course, was soo much fun (and time consuming).

[Matt Cramer]

Coil on plug is doable now using MS1/Extra or MS2/Extra code and the dual wheel decoder mode as long as your distributor has a 4 hole ring and a single hole second trigger. The MS2/Extra can also use a Miata CAS for COP, unlike MS1/Extra.

There's a few prototype Sequencer boards out (they're kind of like a sequential version of Microsquirt) but no code for them yet. Even so, I expect the main benefit to be a smoother idle and better emissions.