Thanks! I zero'd out the AIR diverted scalar; it was set to the default 100 mv value. Thought that had already been done, but it had not.

 

Backed off the timing a bit more. Only time I'm getting knock retard now is when I quickly punch it at low rpm. PE need adjusting maybe? New data log and latest BIN posted here, just in case anyone wants to jump in and offer advise.

 

Something is seriously wrong if you have to cut your timing back to 12 deg total advance at ~3000 rpm to mitigate knocking (and you're still getting knocking?)... I don't think it's at all related to fueling...

Have you ever checked your timing with a dial-back timing light? Just to verify the actual ignition timing is what is being commanded by the ECM?

This car can't possibly running well at all if it's only putting in this small amount of timing...

 

Okay, so lots to contemplate. First off, I was originally (when the conversion was first done) running a "professionally-tuned" memcal. This was done via data logging and USPS. After many back-and-forth exchanges, dialed in the fuel and was declared good to go. Even then, though, I noticed some knock retard in high-load situations. As much as 9 or 10 degrees; seemed like a lot. That's when I ventured off and purchased a Moates G1 and Burn2 so I could adjust the SA tables and attempt to limit the knock retard I was getting. Starting with the "good to go" BIN (that had established good fueling) I merely adjusted the SA down in the areas where I was seeing knock retard. But when I then ran the new tune, I was actually getting quite a bit more knock retard. To make a long story short(er) I just now checked the knock filter board on the "professionally-tuned" memcal. Ohms out at 2.02K. The memcal I have been using with the G1 and my own chips is a real, live 5.0 TPI/Auto memcal. Knock board on this one measured 14K. Viva la difference. So my question is: with the 2K knock board, am I getting accurate knock retard function? I'm guessing not.

Next thought: The 1987-89 5.0 TPI/Auto cars employed the same "peanut" cam that my 1990 5.0 TBI/Auto has installed. So, I looked up a BIN for 1989 5.0 TPI/Auto (APYM) and compared the SA and PE tables. I'm aware that the APYM was MAF, so the tables may not be fully compatible, but I was hoping someone here could chime in and tell me whether or not trying to emulate these tables would be something to try. PE table is quite a bit different; I have noticed that my knock retard seems to kick in at 2,000 rpm, which is where the AXXB PE table goes negative. I have posted a comparison:

 

Two problem areas:
1. $8D uses (evidently) 100kPa SA for cold idle SA. If I use the AXKT SA (or thereabouts), it has 0 degrees at 100kPa; as a result, my cold idle SA is 0 and it does not really like it. How should I address this parameter? Should I just set 400-800 100kPa SA at 12 degrees and leave the rest alone? I have attached the SA table I have worked up to over the last couple of weeks. With this table I get no knock retard and closed-loop idle, part throttle and full throttle are all pretty smooth and responsive.
2. When coming to a stop for the first time in closed loop, the throttle hangs a bit (100-150 rpm) for a second or two, then settles to commanded idle.

 

IIRC, you can disable the cold engine SA logic. It should be this parameter, set it low:

L8031: FCB 126 ; 54.5 Deg C, IF COOL >= Disable cold spk, (113F)

RBob.

 

Thanks! New problem, if I could impose upon you once more. Been trying to fine tune the last week or so and have this problem intermittently. I have tried three different tunes (two of which seemed to work fine previously), two different ECMs and two different memcals. I get this same problem with the different ECM, memcal, tune variations. The "Spark Advance Relative to Reference Pulse" and "Spark Advance Relative to Top Dead Center" normally follow each other. You can see here they diverge quite a bit for a split second. SA relative to TDC is at ~25 degrees (teal line) and SA relative to RP (blue line) is 0. When it does this, the engine cuts out (as you probably would expect). Is this a symptom of an ICM going bad?

 

Check the log for knock retard. Only one of the two SA values shows knock retard.

RBob.

 

That's one of the things that's strange. During the three instances of this behavior (in about 5 minutes of data log), there was no knock retard. I saw some in other parts of the log, but where these two SA tracks go in opposite directions, there were no knock counts or knock retard. I replaced the distributor last summer and had this trouble a few months after. I swapped out the generic ICM that came installed in the new distributor for the one I removed from my original distributor (ACDelco unit that was original equipment on the TBI engine). I used plenty of heat transfer paste when I swapped ICMs and the problem seemed to be fixed. It has now reappeared. I went ahead and ordered a new ACDelco ICM (D1984A) after reading that the TPI engines call for a different part number than the TBI engines.

 

FWIW, swapped out the ICM this morning. No more cutting out or misfire. Actually ran real good. Open loop idle still needs work, but I think I'm gonna run this tune for a while and let 'er learn. I posted what I consider a "final" SA table just in case anyone else is interested.

 

Set my disable cold engine SA to -40c.