another Alfa Romeo conversion

[festy]

I installed a socket on a 4cyl memcal today, and tried an OSE $12P bin, but couldn't get the ECU to boot.
So I spent an hour re-checking my soldering, comparing eprom pinouts, re-downloading and re-burning the bin, still nothing.
Then I started comparing the bin to other 12P ones, and think I found a problem with the one I was using, so tried burning the OSE BLCD bin.
The ECU booted, only a quick flicker on the data line so no VATS issues this time... so I start TP5 and it immediately found OSE pro v1.1.1 hardware

Now I just need to find myself a set of injectors and i can start installing....

[festy]

I'm starting to get the hang of TP5...
I've got the ECU connected to a laptop in my shed, so I can explore TP5 remotely from the comfort of my loungeroom while watching tv
I wired up an old o2 sensor I found in the shed this morning and it seems to pass the propane torch test, the spikes correlate to me waving the flame back and forth past the sensor once it was up to temperature:

o2.jpg

My PCBs arrived today, but not my component order - the DS1245Ys are on back order so they're waiting for stock
I could do up a quick adapter card and make my own nvsram chip out of a KM681000, DS1210 and CR2032s, but it was less effort to just order a couple nvsrams from another supplier, and scrounge around at work for the rest of the bits I need

I'm really happy with my USB->TTL bridge so far, I can't fault it except that the eeprom hack means that I've lost 9600bps capability on the uart unless I re-flash the device back to standard.
Not that I suppose it really matters, as it's now part of the ECU so I shouldn't ever need it to do anything besides 8192bps.
TP5 shows constant errors in between valid packets when logging with the OSE_$12PV111_1BAR - Petrol.adx, but no errors at all when using the message 6 adx, or when downloading the cal etc. I first thought that the errors were interference or crosstalk on my uart, but seeing as they go away when using a different adx I don't think that's the case.
Is this maybe because the bin beacons message 6, and the OSE_$12PV111_1BAR - Petrol.adx isn't interpreting the message correctly - but switching to the message 6 adx means the beacon messages are all understood?
I've had a quick look at the flash tool's message editor, and assume the adx has to match the specified message format?
Or I guess it could be that the ECU is sitting on the shed floor without a car connected, so some of the data would probably be considered invalid?

hmmm... a lot more reading required, I think

[antus]

It could be that your interface is still talking 9600, and the message 6 packets are short enough that they are collected before timing drifts too far, but the longer packets of mode1 arnt? PM me another dbgview log of a few seconds of bad comms and i'll be able to see if its a plugin issue or if the wrong data is coming in from the port.

[festy]

Or... maybe it might be in some way related to a flat battery
I just fired it up and logged a minute with the "normal" adx, and no errors at all in that time.

According to my earlier log, my battery voltage was about 12.2v when I was seeing the errors, now it's a healthier 13.5v after being on a charger all day.
That's the only thing that's changed between the two attempts.
If it does it again, I'll check it out with my protocol analyzer to confirm baud and levels etc.

[antus]

Ive had successful logging in cars with as low as ~ 10 volts (faulty alternator) so 12.2v should be fine.

[festy]

I went back and did some more poking around, and think the DA errors might be resource related.
The laptop is only an old 800mhz p3 cpu, and it's sitting on a steady 100% busy.
I've stopped every process and service I can, and it's a bit better - but the error rate follows CPU activity.
Viewing the dashboard seems to make it much worse, probably because it has to update all those objects on screen - whereas there's almost no errors when on the monitor graphs.
Running it through vnc is making it a lot worse too.
I'll try it with a faster machine and see what happens...

[VL400]

Making some good progress Certainly sounds like the processor could be holding it back if the dash screen causes more errors.

[festy]

I've done a bunch more testing, and now it's just strange...

With the 'standard' 1,2,3 bar adx I don't get errors on the Monitor view, but do on the dash and history.
With the message 6 adx, I get no errors on any view.
If I enable cable echo in the plugin preferences, the 1,2,3 bar adx works perfectly with no errors on the dash or any other view.
Bin downloads fail with cable echo set, as does the message 6 adx. Turn cable echo back off and it's back to everything working but DA errors on the dash view.
I repeated this half a dozen times, setting the echo on or off and re-testing.
So now I'm thinking aliens are messing with my computer

[festy]

I finally got myself a set of injectors, Bosch 0280150712.
They're 16 ohm, roughly 23lb and good for about 47hp each so not huge, but will do the job.

[festy]

I started taking measurements for the fuel rail today, and hit a little snag...
In the pic below, you can see that the thermostat housing is right where the new fuel rail needs to go - it's not a problem with hose tail injectors because the rail is a few inches further away from the manifold, but it's obviously not designed for an o-ring setup

So the 3 options I can see are:
1) make the fuel rail injector bungs very long
2) track down some porsche 928 injectors (and sell my car to pay for them, which would negate the need for them anyway)
3) make a new manifold from scratch

#3 is starting to sound attractive, and it'd solve my TB mounting issue too
I'm going to see if option 1 is doable, but with the extra length, sealing might become an issue.
I suppose option 4 could be to make individual hose tail hats for each injector, and clamp each down with the original mounting bolts, then just use my hose tail fuel rail... I assume someone has tried this sort of setup before?