Hijacking OEM hardware

[Sandy]

At 11:09 PM 9/26/97 +1200, you wrote:
>
>> As far as a '552 not being powerful enough, I disagree wildly!
>
>I guess that since the '552 was designed for exactly this application,
>someone thought it was suitable.
>If you go way back, there was a lot of discussion on what to base the
>EFI332 project on.
>IIRC, GNU tools was one major advantage. The other was that everybody
>wanted to do something different and the CPU32 core plus the TPU would
>leave a whole heap of CPU left over for monitoring and experimenting. That,
>and the fact that the people that actually got off their butt and did
>something, were of the '332 camp. :-)

I hear that, again, I will not do any other work until the I/O get tested,
as I really hate wasted time. Again, we can get to where we want, but more
people need to get involved besides the few that have done so up until now.
As far as the 552 saga goes, it wouldn't matter what we used, as we would
still be discussing instead of doing. My capabilities and time are limited,
I would like to get back into the 332' mode, hopefully next month will be a
good start. I have been on the road for 3 weeks, almost solid. So not much
time to do EFI stuff. My goal, again, is to complete the SINGLE BOARD
version. I have most of the circuits done, and just need to do a bit of
cleaning and then the layout.

Maybe this will help get some things going -

Single board specs -

MC68336/376/396 CPU so you can do the CAN Interface, FLASH, built in 10
Channel A/D, 7K RAM, Potential for flash if MOTO releases the 396 - SEE
MOTO's website for 336 and varients.

Single Chip 16 Bit Flash and Ram Devices

Similar Crank/Cam Interface as existing I/O (LM1819)

Similar A/D 4 Channel Buffer as existing I/O (LMC6494)

Similar Octal Out as existing I/O (MC33298)

Interface to the CAN buss (82C250)

2-10 AMP protected High Side Drivers (IR6210/IR6216)

1 3 AMP protected High Side Tach (AUX) Driver (IRSF3021)

IAC (May Get Axed in the design, drive parts are expensive, hard to get)

The only area in question is the drivers, as I was going to use the nice
LM1951's for High impeadance injector drivers, but their avalibility is
limited. I just got the IRSF3021's (3 amp protected switch) and will build
a sample driver with the SMD parts and see if the work. If they do, and can
take the 'heat', that will be the driver, small, and short proof.

In anycase, 12 drivers will be ON THE BOARD. Use as high impedance driver
for 1-12 Cyls, or use as drivers for the subdrv board for driving coils or
low impedance injectors.

Again the circuit layout for all of the above is done, except for the 12
injector drivers. I will be doing the PCB as a 4 layer board at the
suggestion of many folks, so the PCB will hopefully be better suited to the
harsh noise we encounter. An most all components will be SMD. Don't worry,
most look like they can be soldered, except for the CPU and Memory. The
target size of the PCB will be the size of an airplane barf bag (think
about why, re-read start of post). 

That is, again the status of where I'm at. I will not go much further (or
tell anyone that if do) until the Crank/Cam interface gets some testing as
well as other areas of the I/O board. I currently have several hundred
dollars in the proto PCB's not to mention the parts.

>From the software standpoint, I have been hacking uC/OS, and am pretty
close to making that work. If I do, it will be an easy alternative to using
GCC and RTEMS if anyone wants to. I'm using SDS crosscode (Real paid for
version), but SDS has the complete 7.X system for use on their web,
limitation on the number of linked files, but that is easy to work around. 

That's the end of the soapbox stuff, now as the now cancelled talk show
host said 'Let's get busy'

Sandy