RomRaider

I think it might depend on what is meant by a 'reset.' I think learning might be preserved on some ECUs when trouble codes are cleared using an OBD-II scan tool.

Not on my ECU. A8DH200Z or on A8DH200Y - not sure whether this is a 32bit thing though. On the older 16 bit ecu's it does.

Most of the 32-bit ECUs do, some don't though.

I have a 05 STI running A2ZJB11J and after flashing I have to perform a reset. It remembers all the learning including the IAM.

Question for Merchgod or somebody who knows?

I want to know more about the timing compensations.....

1. When does the Alternate coolant compensation come into play. I think I read in a thread by Airboy that it was under 3800rpm. Is it also under a certain load?

2. When does "Timing Compensation A - Intake Temp" come into play

3. What relevance does the "Timing Compensation A - Intake Temp" Activation have. I assume this does nothing currently as my table is at -100. I also assume the furtherest most left column is "Load" and I enter the % I want timing to go up or down at certain RPMs. Or do I have it totally wrong?

4. What effect does the Activation table have on the "Timing Compensation A - Intake Temp", Does this mean it isn't working?

5. When does "Timing Compensation B - Intake Temp" come into play

And finally
6. When does "Timing Compensation (Coolant Temp) come into play

I have attached my factory 06 WRX Rom to show what tables I am talking about.

Thanks

Michael

Base Timing Alternate (Coolant Temp) comes into play when the target for this table is greater than the normal base timing target, as explained in the help text.


The factory values for your ROM for the activation table are -100% up to 0.40 load. This means there will be no Timing Comp A up to 0.40 load (and interpolated decrease between 0.40 and 0.60 load. At 0.60 and above, in your map, the timing comp a iat is not modified. That is, this activation table directly determines the portion of timing comp a iat that is applied based on load and rpm.


I explained this here:
http://forums.nasioc.com/forums/showpos ... stcount=36


IIRC, it is active all the time, but I'm not 100% positive on this.

Thanks for explaining the comp B table thing.

Could you give an example of the bold part for the 32 bit ECUs?

Thanks

There are some variations on this point with the 32-bit ECU, but one example would be (both conditions would have to be met):
The last FLKC is greater than 3.9 degrees or less than -3.9 degrees.
The last FLKC is greater than last timing advance (maximum) map value * 0.39. or less than the last timing advance (maximum) map value * -0.39

IIRC, some ECUs omit the last condition.

How can the last FLKC be >3.9 degrees? I must be missing something as I don't get it. What do you mean by absolute correction?

From what I see FLK is usually a small value, like 1.4 or maybe 2.8. At least that is what I see in the learning view.

Absolute correction = *absolute value* of the correction. The magnitude of the number is another way of saying it.

stock 2L WRX can get some large values for FLKC (-5.x*), especially around 5000 rpm. The log file in this thread has large FLKC numbers.
topic4317.html

The conditions above is for exiting Fine Learning mode and into Rough Correction mode. There must be other ways to enter Rough Correction mode?

If you look at the conditions to exit fine corr mode, such as the timing advance map value threshold, conditions to exit that mode and enter rough correction mode are similar and therefore it is likely that once FLKC in the current cell is changed to an extreme value that a drop in IAM (if negative) or an increase an IAM (if positive and IAM already low to allow that type of value) would occur immediately, making it difficult to log extreme FLKC values before the change in IAM takes place and the entire FLKC table is cleared.

So say my ECU is in fine correction mode, my IAM is at .80.

Now the most my FLKC can be positive is about 1.4 degrees. How then could it ever change back to rough correction mode to raise the IAM?

That is an interesting point. I just triple checked the reference ROM in question (16-bit ECU) to see if I was missing something, but there's no other means of entering rough correction mode (if already in fine corr mode) unless at least that threshold of -3.9/3.9 degrees is met (along with all other conditions). With the factory map, with the highest max advance being 10.9 degrees within the RPM/load thresholds for rough correction, that an IAM of 11-15 would never meet the positive threshold. The IAM could be in that range and effectively change and end up higher, however, due to the following:
1. hard or soft reset - you default to rough correction mode after a reset (and a flash).
2. Enter rough correction mode due to hitting the negative FLKC threshold, but when the IAM is re-evaluated, end up higher than the current IAM due to the lack of knock in a different load/rpm area of your maps as each step change is completed (which are not necessarily consecutive).

It might be that this was not noticed before because typically, if your tuning and see an IAM less than max (most tuners), you probably are going to make some changes to the tune, resulting in a reset and default to rough corr mode. Or just general tuning if IAM < max is acceptable to you (and end up with the same reset due to the flash).

I'm not positive that this also applies to the 32-bit ECU as there are some oddities that I haven't looked into yet.

Thanks for the the reply. At least I know my brain still works. I thought I didn't understand something.