What data accumulation module do you have?
As long as the circuit can be accurately initialized to sync with the leading edge of the injection open signal from the ECU (and I recommend loading the ECU injector driver circuits to at least half of the current draw of the injectors inductively to prevent unstable transistor loading), and it is accurate enough to consistently produce the same duration each time a TPS value is read, and can be made to react in a manner other than linear, as most circuits of this sort are, and ... jeez, there are a lot of things to think about!
Frankly, interfacing a PIC IC (or possibly a MIL spec Z80. Brush up on your machine code!) to an EEPROM map, NMI it to injector open in, tracking TPS and possibly the crank angle sensor, and mapping a divided clock pulse count figure to injector close may be more easily and flexibly implemented. Another advantage is many PICs have addressable error code access. As for the connectors: Molex! I'll be interested to hear how it's going, so keep us posted.
Hell you even lost me a couple times too...
No seriously I have a circuit that can modify the duration of a PDM signal based upon a reference voltage. Which I think I could fool with either an analog out or just a simple voltage divider with a digital potentiometer. That is idea one.
Idea two is what I think you are referring to, basically why try to tweak the signal inline, just replace the brain all together. I've got access to some interesting stuff through work, I've got a couple of Motorola Coldfire based units from netburner.com lying around.
I think my first step is to get the signals on a scope and maybe tee off into a data acquisition system and record some data. I know the pulse duty cycle to the FI is roughly 80%, any idea what the frequency is? I'll have to sneak one of the battery scopes out to the parking lot on the weekend
I've just gotten some really cool tools at work to do my own boards with FPGA (think digital playdough) tools integrated. FPGA's provide plenty of speed for the calculations and are easy to tinker with but I would still need a fairly fast analog front end to source the current and "keep up".
I'm not going to save time or money. But I will have fun. And I might learn something too. And if I make it work, I'll publish the plans and software.
Here's my feature wishlist
Wireless connectivity, either bluetooth or wifi
Simple programming format (csv or xml)
Intelligent data logging to non volatile media (SD card)
Simple bypass switch to take the unit out of the loop and return to stock ECU
accelerometers like these
so I can get some indication of performance. In a dream world the controller would be set to "learn mode" and you would go do some top speed / hard acceleration runs and the controller would try to produce an optimal map by learning which tweaks produced better acceleration readings.
All this crap is overkill for a 919. But if I make it work on a 919, then I can make it work anywhere there is fuel injection. And I have been lusting over that new Buell.