This is interesting. Some thoughts.

ideally, you would want a separate (load, rpm) ignition map for every one of the eight cylinders. This could be calibrated with very high octane fuel with otherwise pump gas like characteristics. The ignition would always be set to the mean best torque per cell by this map. Cal this the base map.

Then, there would be a separate (load, rpm) adjustment map for each cylinder. This adjustment map would only retard ignition from the base map and would learn relatively slowly. Some low load cells would never retard but would instead always be at the cylinder specific base map values, because those cells are not octane limited even with pump gas. Other cells would be very much retarded from the base map.

One could also have a very fast decaying temporary response to knocks, but let’s ignore that for now.

If that’s the ideal system, the next question is how can this system be made more “parametric”. The reason to make it more parametric would be to make it fit in small amount of memory and also to enable cross-learning between cells.

Is this on the right track?

 

I think it's covered given the hardware limitations?

The stock timed retard stops knocks with a temporary additive retard level.

The cruise map is the best torque map averaged over all the cylinders with some allowance for NOx limits?

The load input from the LH is biased toward light throttle. Max load is reached well below full throttle so few load columns are used at medium to heavy throttle.

The chip creates a map for each cylinder which is 2.5 times more detailed than the cruise map (1500-6200 range, 12 cruise / 31 chip rpm cells) with two load levels, cruise and WOT.

 

It’s an interesting problem and a clever solution. I’m just thinking out loud here.

Are you confident that the engine is knock limited everywhere on the map where the logic is turned on? This system works the best if turned on in the regions of the map that have all cylinders knock limited.

It makes sense to for the gradual advance to stop after certain number/frequency of knocks has been detected. This is what you are doing, right? Then you still allow retard with further knocks, but no more advance? One idea that I have is only allow the retard stage to learn for a fixed amount of time or cycles at the particular rpm. This way, the ignition map freezes itself in some hours and then remains constant. The stock knock retard system (which can be calibrated to be more or less conservative) will then have to deal with the occasional hot day and bad tank of gas.

 

It was tempting to cut off the learning entirely at some point but I decided against it for safety's sake. Folks will have to reset the EZK to correct after a deep correction for bad fuel.


GT cams version! (Auto option for cam swaps, too.) Only the cruise and WOT maps were different vs S4. Interestingly, the values in the idle maps are the same as S4. Only the idle rpm scaling for the map to change the rpms the values apply to was modified to match the higher warm idle rpm range the GT needs to run at. (The ignition idle map assists the idle air controller with advance/retard to keep the idle around the target idle rpm.)

 

Forgot to mention, GT also has different rpm scaling for the cruise and wot maps.


GTS maps have all sorts of changes in the knock sensing maps for the larger displacement. Only one notable programming change in the stock GTS code. A map was added for the automatic transmission shift retard. Instead of a blanket momentary 16° retard shifting up/down there is a sliding scale 18-25 degrees over 3600 rpm.

 

Thought the GTS pistons were 100mm? 5.4 was done by crank (longer stroke)?

 

Well...the pistons are different too

 

Placed an order for my 1990 GT today, cant wait to try this out IRL.
Tack Ken, you are such a great contributor to the 928 world.
Been waiting for this chip solution for a long time!

 

Chip installed.

But this is what was in my EZK?

 

Here’s another idea. How about seeding the cylinder specific advance/regard values with non-zero values that are known best values for say 91 octane California pisswater?

 

Will be an interesting comparison. (Would like to read that chip in if you'll let me!)

The AutoThority S3 chips were just a rough bump in timing and fuel at 3/4 load with no custom programming. Oddly, they had no change to the WOT maps so they could not dyno higher than stock!



Attached a stock 90 GT EZK binary. I was unable to find a good GT bin on the internets (the 8K GT bin on 928S4.com is not a valid file)

 

That bin was from my GT. Probably some super special 89GT stuff that you can't decipher LOL.... I'd be happy to let you check it out just to see if Porsche did do anything different for the 89GT or perhaps just evolve it for the 90.

 

I can mail it to you and you can mail it back after having a look at the programming.

First impressions are good. I just took my car off the lift and moved it to my driveway. Normally when cold my car was always running a tad rough and the throttle needed to be feathered taking off in 1st or reverse until it got warm. When I just moved it now it was rock steady and didn't feel like it wanted to stall, as it would in the past. I will get out for a drive later today.

My car has always run a bit rich. Looking back at my service history the first owner changed the LH for a used unit back in 2000. But in 1996 he had a shop install performance chips, so when the LH was changed in 2000 I don't know if they transferred the upgraded chip. I have since had my LH rebuilt, so my current LH is stock.

 

Well, the car runs as it has never run before. I can't compare with stock since my car had the aftermarket chip, but it has improved in many areas.
-No longer rough when cold
-Idle dip when going into neutral is totally gone, idle drops right down to 750ish.
-Top-end power is noticeably increased, especially between 5 and 6k
-Car feels like it just wants to go, even when close to 6k
Assometer says at least 30hp up top

Only dumb dumb was I currently have a tank of 94 oct but my regular fuel is shell v power 91 ethanol free. So I will have to reset the chip when this tank is done.

On a side note, it looks like my ignition switch has died, at least it happened in the garage.

 

That'd be cool. If you had your LH rebuilt they would have transferred the chip that was in it? The AutoThority fuel chip is probably not helping things if it is still in there. I can send you a stock one.

Too much to fit on the case label but technically you would only want to reset if you go to a higher octane. If you go to a lower octane the chip can adjust downward.