I have been running the last rev of Board and Software in my truck for a couple of months now.
Three issues I know of
Idle Control sucks < known the time for the idle motor to actually change the idle is too long addressed it differently in the next rev of hardware.
Random Communication error -<<< big one occasionally the I2c gets interrupted by the ignition interrupt and either the trans controller crashes or the fuel crashes. Lots of fun when under hard accel it stuck in second and I hit the rev limit built into the pulse filter. Turning ignition off and back on resets the controllers in next rev of hardware I approached the whole coms thing differently. Using a single hardwired byte to convey O2 tracking for closed loop and analog distribution amplifiers to communicate Map TPS RPM and MPH between the modules. The Hardware Rev after the next I have moved all calculations off the timing controller so that only the injector pulse timer and the ignition pulse are being dealt with. On this next rev I am still calculating the RPM from interrupt period on the fuel pulse controller. It is not necessary to do so one rev three I am using a Frequency to voltage convertor to provide a RPM signal then calculation the air fill weight from Map, CLt and overlaying a Feed back Map based on RPM, Map and O2 feedback.
Trans shifts way too conservative. <-- this is just a preference kind of thing, the trans I have in the truck had a "Shift Kit for Towing" installed I have no change in Line pressure when modulated and no change in Torque convertor feel when modulated and no change in 3-2 downshift so I am only guessing that they bypassed accumulators and line pressure regulation at the valve body plate like the old B&M shift kits plate for 700R4. However my code tends to upshift sooner than I really like but I guess that it saves gas. In next rev I intend to have a higher shift point. I also am incorporating a mode switch and flappy buttons function to the trans controller.
Biggest change on next version -- direct calculation and no table for base fuel I worked out a pretty stable calculation that takes @470us and it is broken down into 8 separate parts so the base fuel updates every 2 rotations. This is actually faster than table lookups and eeprom reads. BTW atmegas have no division in the processor so all maths have to be iterated subtractions. However the hardfloat does seem to work so multipling a floating point does work pretty quick. So divisors are calculated and stored as const float at start up and used by multiplying.
I will be posting REV 2 drawings and Code in a few days, I drive truck every day but am pretty sick and have long periods of non-function.