another Alfa Romeo conversion

[festy]

After having a decent read of the group 5 regs, I've given up on that idea - so making it all original looking isn't a priority at the moment.
I've got a set of ecotec injectors on the way, so will knock up a fuel rail to suit once they arrive - unless anyone can tell me the dimensions for the fuel rail's injector seats?
I'm considering making my own plenum with a VR throttle body mounting at the front, rather than the current smaller solex sucking air from across the top of the cam cover - but will probably leave it the way it is for now, and concentrate on actually getting the car running first
The plenum mounts to the inlet manifold by rubber hoses making up part of the runners, so making a new one would be pretty straight forward.

So now I'm just waiting for a few parts to arrive - injectors, eprom programmer, memcal PCBs and components.
I still haven't sorted out a wiring loom. I'm hoping I find a JE camira at the local wreckers - but failing that, how much more effort is it to strip the loom out of a VN/VP? I'm sure they'd have a few of those around.

[brenyw]

try LD Astra or whatever the equivalent Pulsar for the loom, probably easier than commodore

[festy]

Well my injectors never arrived.
After 3 weeks of excuses from the ebay seller as to why he hasn't posted them yet, I think I'd better look for another set.

Today I stripped the ecu, loom, tb and a few sensors out of an N13 Pulsar. I opened up the distributor to get the module out, but somehow got sidetracked and left without it
The tb is roughly the same size as the Alfa's original one, and looks like it might even bolt up without too much modification. It's also set up for a throttle linkage instead of being cable operated like the commodore, so I might get lucky and be able to use the factory linkage rods etc.

I've been slowly chipping away at preparing an initial $12P tune, ready for when my PCBs arrive. I thought I should burn an eprom so I can bench test my comms link to the ecu now that I have the wiring harness.
I had a bunch of eproms scavenged from old computer bits, but no way of blanking them. I tried sitting them on the glass of a 400w HID lamp for 30 minutes after looking up the specs and seeing that it should have a decent spike at the required wavelength... but no luck, just very hot chips with their memory perfectly in tact. Hmmm... I bet the light's lens is UV cut glass

So next I tried the ultimate source of sunburn, my TIG. I lined up a few eproms, and struck an arc a few inches away.
After about 30 seconds exposure, I read each one. The 27c512s were blank, one of the 256s was all blank except for one single bit, and another 256 was still more or less unscathed.
I wrote an image to some 512s and retried blanking them, all good in under 30 seconds. Do 27c256s take more exposure to blank than 512s?
The TIG is probably putting out far more UV than needed, and shortening the lives of the eproms... but it works damn fast

[VL400]

Certainly a different way of erasing EPROMS!

[Holden202T]

lol!

[festy]

Yesterday I tried to get comms working without much luck, but made some progress tonight.
I wired a _very_ cheap USB TTL bridge into an 808 (simple 3 wire conection, no echo or anything fancy), then plugged the ECU, tb, map and temp sensor into the loom, and tried to fire it up with a VR man memcal - but no data.
This morning I read the docs for the USB bridge, and found out it won't do 8192 baud in standard form so modded it's eeprom to lock any requested baud from about 7000 to 10000 at 8192 (or 8191 actually, but close enough).

Back in the shed, I tried it again but still nothing. Checking the pinouts, I found out I hadn't supplied power on the ignition line
Fixed that, and the USB bridge's data LED started flashing madly.
I fired up hyperterm and could see data on the com port, so tried TP5.
No luck, but starting it with "autodetect hardware" checked crashed the program.
I tried with the flash tool, it asked whether it was a VR auto or man, then seemed to work fine. It downloaded the bin, ram and cal without any problems (but I haven't checked the contents of those files).
I tried a few more times with TP5, but couldn't get it to find the hardware.

So then I tried a different laptop, with a freshly downloaded copy of TP5 and OSE plugin. Same sort of result though.
I did manage at one point to get some sort of connection, but the dash was jumping around like crazy so I'm not sure it really was working.
I think I have to do a bit more reading on TP5, as I was more or less guessing wildly at which buttons to push etc so it's probably user error

[antus]

Possibly not. Im on my phone now but will send you a PM next time im at my pc. I made some changes to the timing in the latest tp5 plugin, and for the most part it was an improvement but i have heard of some crashes, but have not been able to reproduce it. If you can make it happen and are willing to help we can probably sort it out. Not sure if there is another problem, but it should never crash.

[VL400]

If the data led is flashing at key on its still got VATS enabled. Disable that and should work ok with the plugin and TP Although that does not help antus and his bug solving so still try and work with that if you can.

[festy]

I seem to have no problems making it crash, on two different XP systems
I caught one of the crashes with a debugger and started looking for causes, but didn't persist as I assumed I was at fault.
Happy to help out if I can.

VL400: it would still have VATS enabled, it's a vanilla VR memcal. It never ocurred to me that might be what was happening, I'll install a socket and a $12p bin tomorrow and see what happens.

[antus]

Run dbgview available here at the same time as the plugin. You'll see what its doing, and when it crashes send me the log in a pm and describe what you were doing at the time, include what vehicle was connected, and what type of interface you were using, and the version of TP5 and the plugin. It should be enough info me to see where and why it happened in the code. Thanks!

http://technet.microsoft.com/en-us/sysi ... s/bb896647