How do you setup your tune when tuning for spark? (my problem within)

[pauly24]

Ok I've been logging data looking at ways to increase my spark timing

Now my octane scaler for a 20 min drive was a min of 0.33 and a max of 0.79.

Its an E38 ecu so an octane scaler of 0 means it is using the high table 100%.

Now if the min it ever got to was 33%, this means that even after a 20 min drive it never commanded the spark that was in my high table.

So if I wanted to increase the spark timing at any point it would be pointless because it never commands the value in the high table anyways.

I hope someone understands what I'm getting at and can explain how its done.
Thanks.

[swingtan]

Rule #1: Don't tune spark until you have tuned the AFR's
Rule #2: Remove timing and work back up, don't work down.

Check other threads for descriptions on how the octane scaler works. They will help you understand what is going on.

Simon.

[pauly24]

I know AFRs are first (im installing my wideband tomorrow)
I have the AFR tuning part down pat in my mind and playing with the tune, just tomorrow will do the real deal.

But until then I've been trying to understand how the timing works.

I believe I understand how the scalar works, when knock occurs the scalar increments which in turn moves the timing towards the low table. After time it slowly goes back when the specific conditions are met which moves it closer to the high spark table.

I dont think I have the right PIDs for logging spark, I use base_spark, I believe this is suppose to be the actual value that is commanded after the scalar is applied. Im also logging ignition timing advance for cyl #1. But this value is always lower then base spark so obviously there is more being pulled out.

I have other things logged like spark being pulled from intake temp etc... I just cant find where the timing is going.

Im trying to find some threads that explain all this but cant seem to find any. Will keep reading but any help is much appreciated!

[swingtan]

For the E38, log {SPARKADV} Base spark is a reference table and not actually used for commanding any timing.

Final timing is a combination of the hi / low tables, octane scaler, IAT table, Torque Control tables, ECT tables........ The SPARKADV PID shows the final commanded advance value.

[pauly24]

Arh yes thank you.

Base spark is just the value after the scalar is applied.

Then from the basespark you subtract the values from the IAT tables, ECT tables etc as you mentioned and then that value is the commanded spark SPARKADV PID.

I can now figure out exactly where my spark is coming from and see what is affecting the timing :)

Next question though (well same question from before) the octane scalar, I know how it works and have seen it in action, but it just doesn't go down fast enough.
When people are tuning on the dyno, do they paste the high table into the low so they can commanded the timing 100%? Or do they play with the octane scalar settings to make it gain quicker?

Once the car is tuned, do people leave the high map pasted in the low? and this is how its driven on the street?

[swingtan]

I've never actually logged Base Spark.

Here's what I log ...


SAE.SPARKADV: It gives the actual ECM commanded spark advance after all the calculations have been made.
GM.KR: This shows the amount of knock retard.
GM.ASPARK: The Octane scalar.
GM.SEMRET: If you haven't disabled torque reduction, this shows if it is pulling timing.
WB AFR: You really need to know what the fueling is like to get close on spark.


Then you just log, lots of data.....

Go back through the logs and look at the relationships between the PIDs and what occurs when.


If you were on a dyno and in a controlled environment, Id say to make the low spark map the same as the hi. This helps remove false knock events. But unless you have knock phones, or you are very sure of what you are doing, it's not a good idea.
Log the data ( did I say get lots of data? ) and then sit down an go through the logs.
If at any time the WB indicates a loss of fueling control ( sustained +/- 5% difference from commanded ) then fix this before playing with the timing. You can get false knock from running too rich and too lean.
Look for knock , start with the biggest hits and remove timing from those cells in the HI map. Use a "knock MAP" to average all the results and make applying the changes easy.
copy the hi map over the low map and then subtract 8' off the entire low map. This is to keep a buffer that will allow the scalar to move and pull timing if things go a bit crazy. Do this after all tuning if using a dyno.
Save the tune and load it, then go logging again. What you are aiming for it to get the Octane scalar back to 0 and have no KR appear in the log.
once complete, you should see very little KR ever on your logs.


Now for some notes....


Is this the "best" method? No, the best method is on a dyno, with knock phones, where you can measure and prove every change you make. It is a pretty good way though and gives good results.
Try and do all the logging when the temps are "cool". Currently in Melb it's a great time to log as the IAT's will remain pretty constant. As the IAT's go up over 35', you will get additional knock and need to tune the IAT spark table.
Ensure the engine is up to temp before logging. Give it a good 10 to 15 min drive before logging. Before that the data is tainted and could give false results.
Tune the VE and/or MAF before playing with the spark.

[Boost]

Its an E38 ecu so an octane scaler of 0 means it is using the high table 100%.

Now if the min it ever got to was 33%, this means that even after a 20 min drive it never commanded the spark that was in my high table.


I have a GenIII LS1 ('99+) PCM in my '03 Silverado. I logged my octane scaler and it is ALWAYS 100%. Does this mean that I am not using the high octane spark table? Or is is backwards on this PCM. I run premium 93 octane all the time and am seeing some pretty high timing, 30+ degrees and 2' being pulled for knock. Someone please explain. Thanks!

[5.7ute]

100% means that the high octane table is being used with the ls1 PCM.

[Boost]

Thanks, and I just answered my own question also by reading something that Joecar posted :

"The octane scaler is viewable in the pid E38.OCTSCL_DMA;
The scaler is used to compute a weighted average between the HO and LO tables.
E38.OCTSCL_DMA = 0% indicates full HO weight;
E38.OCTSCL_DMA = 100% indicates full LO weight;
(this is in the opposite direction compared to the LS1 PCM)."

[The Alchemist]

if I was to cautiously look at on road timing tuning (which I wouldn't) I would reduce timing in the higher load points say over 0.4g/s on the high octane table by about 0.5deg per row and blend it to look smooth. You want to aim for around 18deg at 4400rpm at around 0.76 to 0.8g/s. Copy this table into the low octane table so they are the same. This removes the scaler effect from the timing equation. Then follow Swingtans advice and move upwards in timing slowly (0.5 to 1 deg increments) and carefully doing lots of logging. The best timing depends a lot on the fuel you are using too (as you probably know already). For example over here in NZ with our crap 95octane I run 18deg at peak torque at 4400rpm in a VE V8 but over in Aussie you guys can run around 21 to 22 deg at peak torque on your 98 octane.
Hope this helps.
Mike

[swingtan]

Hey Mike, do you get E10 in NZ? There was a discussion over at ls1.com.au and most tuners found they could run a timing map close on that of 98 octane fuel. Don't discount E10, I know if I couldn't run 98, I'd certainly run E10 before running 95.

Simon.

[Tordne]

Gull Force 10 is an E10 blend here. I only run either BP Ultimate 98 or Mobil Synergy 8000 98, but that is not available at all service stations.