Interesting Alternative via Quad4

[robertisaar]

was playing around with some stuff yesterday and found what may be useful for high revving motors that don't need -E trans control(or at least not with this ECM): A8.

there isn't a hack out there for it, but i do have a fairly hefty XDF for it...

but, on to the details.

it seems the early Q4 (possibly later ones as well) used a 7X crank sensor/ring identical to the 660, so that is good to know. it also has two seperate injector drivers and intended use is with low-impedence injectors(i believe the HO injectors are ~28-32 lb/hr?), though it seems that there are jumpers coming off of the ECM specifically for this. i believe if they were disconnected, high-impedance could be used successfully. well, i think they would work fine anyway, but it's another measure of insurance.

it also uses a waste-spark setup similar to ours, just meant for 4 cylinders. i believe modifying a memcal the way third-gen f-cars do for making a V6 memcal work for a V8 could work nicely, though it may be possible just to stick in an entire 660 MEMCAL with custom programming and be even simpler. then of course, change the number of cylinders in the BIN.

and boy could those DOHC motors breathe.... GM saw fit to give it a spark table similar to A1, but the fuel table is what makes it shine... the normal VE table goes from 800-4000RPM, which is nothing special, but the extended VE is awesome.... it goes from 4000-7200 in 200RPM increments with 10kPa increments. there's also a 100RPM/10kPa resolution idle VE table, but that's nowhere near as exciting.

it also looks like there is a seperate table for spark advance while in PE mode... and i'm not talking the normal modifier table either, i'm saying it's a full-fledged main spark looking table that goes from 800-6400RPM.

it seems to be able to switch between single-fire and double fire easily as well...

it also seems to have 22BLM cells.

as for the ECM itself, it looks to have been used in 2 configurations, the 1228707 and 1228708. the 8707 is a non-weatherproof and looks similar, if not identical to a 1227730. the 8708 is weatherproof and looks similar to if not identical to a 1227727. so, chances are good that minor repinning is all that would be necessary to swap it in. the 8708 looks only to be used in W-bodies, while other quad4 applications got the 8707.

that's all for now.

[robertisaar]

took a deeper look, slight complication now.

it seems the Q4 ICM sends 2 signals to the ECM, a 1X and 2X signal, with the 1X being triggered every time the 7th notch(the 1 that's 10* apart from the rest of the 60* notches) AND the 2X gets sent at the same time as the 1X plus again when the crankshaft has rotated 180*.

now the 2X is no problem at all since it operates on the same principle that the 60V6 ICMs do. the 1X may be a challenge though.

now, what gets me is that the Q4 ICM makes more sense than the 60V6 ICM. how? because it sends a seperate signal to the ECM for it to determine both engine speed AND crankshaft position. the 60V6 ICM just sends a 2X pulse, which can't be used to determine crank position as far as i can tell, since the ECM wouldn't know the difference between the unique slot and the opposite slot. the ECM shouldn't need to know this though, since it doesn't directly trigger the spark, it uses a signal sent to the ICM to determine spark advance, which already reads all 7 slots on the crank wheel and knows which cylinder is where.

.................actually now the 60V6 ICM does make sense now. now to figure out why the 1X signal is necessary for the Q4... alldata says it's used to determine when the injectors are fired. i wouldn't think it would matter being MPFI.... unless it's bank-fire(well, for a 4 cylinder, let's call it pair fire), not batch fire.... that would explain the two different injector drivers.

the firing order is 1-3-4-2 though, and the injectors are paired 1/4 and 2/3, neither of which fire in succesion to the other.... unless that was intentional.......

anyone else have a headache, yet?

[Guest]

Look at the paired injectors, they are 180* apart in firing. Completly intentional.

I would have to look at the ECM schematics, to see if it is batch fire or bank fire. The '7749 had a pair of injector drivers but fired in batch fire mode.

The ECM in 660 applications doesn't need nor could care about actual crankshaft position, since it's simple batch fire on the OBD1 660s, so the ECM just needs to know engine RPM, which the ICM will send via the REF HI (Purple/white) wire, this is basically the 6x signal with the home notch being "filtered out".

The fact that the ICM is batch fire or not has no effect on how or what the ECM does with the timing signal, that is purely and ICM property.

I haven't looked at the actual signals for the Q4s enough to know whether the code would be easily used in another application. I would have to look at the 1x and 2x signals a bit more in depth.

No need to modify a Q4 MEMCAL to be set for a V6, just simply use a V6 MEMCAL, this would give a better matched limp home mode too, than the Q4 MEMCAL would. The only reason the V8 crowd is modifying V6 MEMCALS to be set up as V8 MEMCALs is because the V6 MEMCALS are MUCH, MUCH easier to find than the V8 MEMCALS. I have a dozen or so V6 MEMCALs, but only one V8 MEMCAL (AUJP), and I had to trade away a 4cyl $58 MEMCAL to get it.

I have discussed the idea of using Q4 programming with SuperDave a long while ago. I think we came to the conclusion that the engine spends so little time in the upper range that running off the last cell in the tables would be sufficiant to provide good fueling and spark control, in otherwords sticking with $A1 or going to something more documented like $59 or $8D made more sense. FWIW, S-AUJP has an extended RPM table to 6400 RPM. There has been discussion and even some beta testing of Code59 with an 8000 RPM upper limit.

[robertisaar]

jeez, 8K tables.... i need to see the motor that warrants that!

at least some of my suspicions have turned out correct. i can get external ECM wiring diagrams, but those wouldn't be any help at all to try and determine if they are actually seperate let alone capable of bank-firing...

at least this seemed like an interesting idea for a while... the idea of an entire new spark table specifically for PE threw me off though, i've never seen that in another calibration.

[Superdave]

I worked on this years ago, it's a royal pain in the ass. I gave up on it.. and this was before i found the good $A8.xdf that i have now.

[robertisaar]

did it ever at least work? or would it never even attempt to fire correctly?

[Superdave]

It tried to fire but there had to be a timing issue because it would pop out the intake, the next step was to dig into the wiring but at the time i didn't think it was worth the effort. I agree though, it would be a sweet mask to use.

[robertisaar]

did you try and run it with or without a 1X signal?

[Superdave]

did you try and run it with or without a 1X signal?

I tried a few different things, it was a long time ago so the details are a bit fuzzy. I'm sure there was beer involved so who knows.. lol

[robertisaar]

i wonder what the tach out signal looks like....

[Superdave]

i wonder what the tach out signal looks like....

probably a square wave

[robertisaar]

probably a square wave

okay, would the ECM possibly work with that for a 1X, assuming the tach out was a 1X signal itself?

[Guest]

probably a square wave

Beat me to it. lol

I'd have to see what these "1x" and "2x" signals are.

Can you post the documents that you have?

[robertisaar]

these are all the important things i could find...

[Superdave]

in reality the ECM/ICM doesn't actually need all 7 notches, the engine would run fine on 4 (3 for cyl position and 1 ref for RPM and timing adjustments). The other 3 are probably just there for resolution. I think that's the 2x signal you are talking about.


I think if we could find the schematics for the Quad 4 ECM and a pinout for the connectors this could happen fairly quickly. I still have my '7730 test bench so i could test things if needed.


Attached is one of the $A8 bins i was working on, if you need my good $A8.xdf or the $A8 def for Tunercat that i paid $20 for and never used just let me know...