Hello dear community,

at first, I am newly registered at thirdgen.org but not a new tread reader at all. Found a lot of helpful posts in here during the last years.

I searched my a** off, to solve my current problem but didn't got lucky. So I decided to open a new thread.

At first the past of my car to get into the whole thing:
It is a 1989 2.8 MPFI Firebird 5-Speed. Build it up almost from the core and did some tuning to the engine as well. Sharp cam, roller rocker arms, 24lbs/hr injectors, hell of an exhaust, etc etc... . Dialed in the upgrades in my EPROM of course. Car ran good with mods and settings. Back then I ran a 302 ECM with a $3A mask. Not much tuning options but good enough to handle it all.
Of course I did the mods to support a later G60 charger upgrade on the engine. Otherwise, there would have been no need of bigger injectors. (A G60 charger is basically an external, beltdriven small supercharger. Invented by VW in the 80's).

Well, that was a year ago. Now, the charger is completely build in an runs pretty fine. Mechanically, the new build enginge is a masterpiece of myself, haha. So here comes the great BUT:
To support the increased ammount of air, I had to swap to a bigger MAF (I am using a MAF offset, by the way). No problem, ordered the big one, which can read up to 255gr/sec instead of the small one, which can only 150 gr/sec. Now, using a bigger MAF, I swapped to a 165 ECM and a $32B mask. Had to tear it down to match a V6 instead of a V8. Did the burn-off wiring, ECM-repinning and stuff as well. Of course, it wasn't as easy as it sounds, but long story short.
Right now, I have it all back together and the engine running. So far, so good. It cranks right up at every temperatur, and the throttle response is great. There is still some finetuning left of course, but first the bigger problems. I am datalogging my engine with datacat and there is still a big spark advance problem left. That displays as following:
I start the engine cold, let it idle. In open loop, everything is fine. Steady spark, no misfiring and AFR ok. This goes on, until the car has warmed up and goes finally in closed loop. Closed loop is reached at 60°C temp. Now, the sensors are kicking in and the ECM takes over. from 60°C to 70°C it runs fine as well. BLMs are between 125 and 129. But when the temp. surpasses 70°C, the spark advance increases constantly in a stair-pattern. All this happens at idle, no reving. My set max spark advance in the chip is set to 41,8 deg. And this is maxed out at 80°C. So, with close to operating temperature, my spark advance is already full maxed out just at idle and my ECM still tries to advance even more. This goes up to even 80 deg. advance!! I flashed my #1 with a timing light, and the 41,8 deg. are real. I know, how the spark tables and bias terms affect each other and there is no logical explanation for this behavior to me. Sure, my spark tables are far from perfect in this stadium of progress. But acting strange like this, is somewhat else.
The Dizzy is set to 10 deg. BTDC. Matches chip setting also. Like I said, mechanically, everything is fine.
Added a picture of my latest log, so you can see, what stair-pattern i mean.
Hopefully, you guys have some ideas, how to solve this problem. I am 100% certain, I forgot to mention a ton of important info, so feel free to ask for it, if needed.

BTW, I am from germany, so my grammar might not be the best. Sorry for eventually hard to read sentences.

Greetings, Jack

 

The only time I've seen the stepped SA increase is when the ECM invokes the forced knock test.

RBob.

 

Hey, thanks for the quick answer.
I overthought your suggestion and this might be it! My 2.8 V6 doesn't have a knock sensor, but the $32B mask with the 165 ECM I am running, does. Although I zeroed out every knock table in my BIN (using TunerCat), set the knock constants to zero and disabled the knock diagnostic (Err. 42) in my switch table, there might still be a chance, that my ECM wants a sensor response? If so, is there even a way to remove it from the engine management? Or would it be much easier, to just screw one in and wire it to the ECM? Or, maybe i can jump the knock sensor pins on my ecm somehow, to trick it into thinking, there is a sensor... . So many suggestions. But seriously, thank you for sharing your knowledge, RBob. Based on that, I have a clue, where to dig.

Best regards, Jack!

 

The best way to disable all of the knock stuff is to set both of these values high (151°C):

Did you modify the MEMCAL for a 6-cylinder engine? Along with setting the BIN parameter for a 6-cylinder?

RBob.

 

I haven't tried to set the values to max. instead of min. Good point though...

My BIN parameters are set for a V6 engine. In tunerCat, there is a constant to tell the ECM, how many cylinders it has to deal with. The memcal itself came out of a '730 ECM. About what I found out, it is compatible to the '165er as well. I punched the knock board out and plugged the MEMCAL in my ECM. My custom chip is fitted in the MEMCAL via the qick swap adapter from moates.net.

Jack

 

Hello again,
today, I was finally able to make a testrun with the modified EPROM. Sadly, the same results. But it is definitly a knock-sensor problem. Here my further thoughts:
My ECM '165 with $32b mask gets its knock sensor response via the ESC module mounted on the firewall in the engine bay. The signal from the ESC module then goes to the PIN B7 on the ECM. That's how the knock-retard system normally operates on this offset.
But in my case, I dont have a knock sensor or a ESC module at all. And the PIN B7 on my ECM is not wired to anything. Technically, I am running the ECM with a missing essential input.
But how to trick the ECM into thinking, every value needed, is given? At least I have to "simulate" a ESC-module output signal to the PIN B7, which indicates a value, the ECM wants. In my understanding, I need to feed it a value, that indicates a knock-signal from the ESC. This would proof the functionallity of the knock-sensor to the ECM. Of course, this would also maintain a permanent knock-signal to the ECM, but errorcode and retardings are disabled in the PROM. The only effect would be a positive knock-test from the ECM.
Otherwise, it would advance the spark to the moon, waiting for a knock-response, that never comes.
At least, this is my understanding of the whole thing. Please feel free to correct me, if i got something somewhere terribly wrong, haha. But if not, what is the correct output value in Volts from a ESC-module? Means, how many volts are given during a knock-detection to the ECM? I read something about 5 volts, when there IS knock. Can anyone confirm or point the right value?

Finally, to shorten it out:
What signal do I have to feed my ECM to PIN B7, to make it work fine without a knock-sensor and ESC module?

Best regards!!