My MGB runs - sort of!

[pbsch]

Yeah, I thought about that - spent some time going through the scalars and couldn't find it - just found it in the tables - will have to try that tomorrow night.

I also suspect that it may be an equipment mismatch. Some time ago I asked ab out the ks and ecm pairing and then forgot about the response I got. It appears the KS from the 3.4 engine is a KS3 and the KS for this ecu should be a KS8, so it may be worth my while to pick up a correct KS for the ecu too. For short term diag, though def going to try the zero out idea.

[TGP37]

I believe the KR sensor bypass is a good idea to rule it out. But incase it doesn't......

You might have a dead spot in the TPS. But you said you checked them, just a thought.

Possible you have a broken spark cable, corroded terminals or one that isn't fully seated. The increased resistance isn't enough to stop spark under no/low load....but add load and the gap resistance increases until the easier path to ground is through the jacket and into the block somewhere.

When it stumbles, is it like a single cylinder failing or all 6? When I hit false knock it just reduces power, doesn't stumble at all.

[pbsch]

It seems like it's all 6. It gets that hollow pop pop pop sound to it. But looking at different logs, it really points to timing - off to the garage to try this new map with the knock retard tables all set to 0.

[bszopi]

The knock sensor has to match the filter on the PROM. If you know what vehicle the PROM is from, then ensure the knock sensor is also from the same vehicle, otherwise replace it.

Brad via Tapatalk

[pbsch]

Well, I think I've solved the problem. Sometime during all of this diagnostic, it appears that I may have damaged the loom for the injectors. I went back to basic diagnostics this evening and found that both the #3 and 4 cylinders were not firing - when I pulled the plug wires I got super strong spark, but no noticable change in idle. Same story when pulling the injector plugs. rigged up a little test light to see whether I was getting any signal or not and one of the plugs basically fell off the harness - the wires had been broken to the point that one last pull on them finished them off and the other one didn't exhibit signs of failure, but there was no light though I was getting +12V to one terminal. I have to assume some sort of fault in the ground circuit. because I was working with the 3.4L sfi harness I had six discrete wires running back to the ecu though the ECU really only needed three (right bank ground through ECU, left bank ground and +12v. To minimize the possibility of further issues I stripped out the connector, removed all extra wires and soldered/shrink tubed all the connections. Car now pulls hard to WOT in the garage - ran out of daylight for a test drive - cant wait to get home from work tomorrow to check her out. Then itll be time for more datalogging and possibly a trip to the dyno for wideband tuning.

[pbsch]

I also pulled the knock sensor out of the block and grounded it to eliminate any knock signal as a possibility. Maybe one evening next week, I'll get over to pull a part and pull a ks8 sensor out of one of the many cars it came in. I probably should just step up and put a new one in there. I think they're only $30 odd bucks. I put in a heated O2 sensor and that makes a lot of difference in the sensitivity of the sensor to low rpm test runs.

[robertisaar]

actually, the knock sensor will pick up misfires as knock counts, oddly enough. i think that's where i'm getting my false knock from.

so, that step may be completely unnecessary.

[pbsch]

Hey Robert - can you think of any bins of the top of your head that were 5spd manual for the $a1 definition? I was thinking there was probably one for the cavalier z24 1992ish vintage - might be a better platform to re-do this whole thing from for giggles...just to get rid of all that automatic trans stuff...If I could get my logger to talk to a $88 def I'd go back to the BAWX + 3.4 timing bin I was messing with.

[robertisaar]

there were quite a few 5 speed A1 BINs.... but none of them were meant for the iron head engine.

since i've finally gone through it and created a full XDF for $88, i can say with certainty, 88 vs A1 is a matter of preference. they are nearly identical with very few exceptions.

if you wanted to spend the time to go through it, you could grab something like a 5 speed cavalier/beretta BIN and import a lot of settings from a $88 iron head BIN and essentially have an iron-head $A1 BIN. i did something similar to import the 91-93 3.4 DOHC auto and manual calibrations to A1.... took quite some time though.

[pbsch]

That's pretty much what I've done - I got the AUSJ bin and copy/pasted the 3.4 iron head spark table you gave me with the 3.1 iron head fuel table from the BAWX bin. I then went in and found any scalars that set rpm or cutoff limits for the auto trans to 7000RPM. Got the car running pretty well and then started dialing in BLM through the data log. I don't think I have too much knock - even with the two cylinders not firing, i only got 4-6 knocks counted.



I just cant get my laptop to talk to me ECM using anything but the $a1 ads file. Even then, they will only talk when I use a vm ware instance of winxp on my win7 machine. It's a real pita and I don't have a clue why it wont talk. Funnyt thing is that with the VMware, tunerpro doesn't detect the cable, but it still works. On my win7 os, it detects the cable but doesn't very odd.

[robertisaar]

in that case, if i were you, i would start out with BFUZ. it's for a 5 speed 3.1 L or J car, which is the closest thing you'll come to with a stock A1 calibration compared to the F-body in terms of weight. it's also the latest calibration GM released for the application.

[pbsch]

sweet, I'll give it a try. thanks for all your help and patience while I got up to speed on this. I think I've hit that point where I am less frustrated and making progress to where it is becoming fun. I can see how the obsession to build a perfect bin for the car could kick in. Especially now that i have figured out how to make quick changes, burn them to another chip and drop the chip in the ecu in a matter of minutes. I'm back and forth between getting a wideband o2 and controller or just going to a dyno shop. I could see where the wideband would give me the ability to make more changes should I decide to further mod things. for instance, I've read about they hybrid process where you use the imntake manis and aluminum heads off a 3.5 FWD car and get the roller rockers and cams from the 3.5 on the 3.4 RWD block for a pretty nice bump in HP

[pbsch]

Interesting - the bfuz bin I found on gearheadefi has automatic scalars and tables in it...

[robertisaar]

the items may be present, but that doesn't mean they're enabled. a lot of the automatic calibrations have stuff in them for the shift light too.

but BFUZ is definitely manual, there is at least one scalar and one flag you can check to verify it. one actually determines if TCC or shift light logic is used, the other is mainly for IAC logic.

[TGP37]

Glad you got it running much better.

I too am hooked on tuning and the pursuit of finer results.