Hi Mrfred

Can you tell me the subroutines that point to the WGDC update correction delay, as you can see I have found some values in a similar to those you have found, for those not familiar with the program these came from I added the name my self.

ROM:0000164F .data.b h'5A ; Z
ROM:00001650 .data.b 0
ROM:00001651 .data.b 3
ROM:00001652 .data.b 0
ROM:00001653 WGDC_update_delay_1:.data.b h'A
ROM:00001654 .data.b 0


Car is a JDM VII. Thanks if you can help.

MB

 

Quick update. Boost control is finally working as I expect it should. A correction interval of "3" along with proper TBEC values and good BWGDC values allows true load-based control of boost without any boost overshoot issues yet. I'll post up the settings after my next run.

 

The interval settings are looked up in the subroutine where the TBEC lookup is done. The two interval settings are used near the top of the subroutine. The one with the lower address should be #1. What prog are you using?

In the two ROMs I've studied ('06 and '05), the values are stored at 0x1654 (#1) and 0x1656 (#2).

 

I'm still working on my TBEC table ... I'm adding some more aggressive upward correction. I should have data to report this evening as well.

Mrfred, any thoughts on my WGDCC equation? Further logs show it to be fairly accurate.

 

My thought was also an offset of 64 (128/2), but I've seen data from some runs where this offset doesn't scale the data to the expected values. I need to study that section of the code further.

 

Hi, this is from IDA pro. I'm just taking a pot shot in the area you have identified to see if I have anything that ties in with your work. I've not got very far in the disassembly of my ROM.

MB

 

Here's the easy to find it.

1) Find the address from ECUFlash for either the load error or the WGDCC values for the TBEC table and go there in IDAPA.

2) Double click the reference address that calls the table. This will land you in the data table for the WGDCC subroutine. You can click on the reference the subroutine data table (not the TBEC table) to take you to the actual code where the TBEC lookup occurs.

3) Then scroll up a page or two, and you'll see two constants called out. The addresses for the constants should be pretty close to 0x1654 and 0x1656.

 

Hi

Many thanks, how does the attached look, it shows 1a54 which gives me a value of 10 in ECUflash.

MB

 

That's the correct subroutine, but not all of it is there. There's some stuff above it that is part of the subroutine. Based on what you have, it looks to me like 0x1A52 should be #1, and 0x1A54 should be #2.

 

Sorry missed a bit off. Many thanks for looking at this for me, I'll have a go at porting your work onto my ROM. Not many of us using the JDM VII ROM, so I'll just have to get my butt in gear some more. Great work on this.

MB

 

I attached logs for 2nd, 3rd, and 4th gear pulls. The boost control system is definitely working now. I still have it set a little too conservatively though, as its pulling to much boost away after the peak in my 2nd gear pulls, and its not hitting 260 in my 4th gear pulls. I think my next iteration will be pretty well dialed.

 

Man,
That explains exactly what I was seeing this weekend while tuning my green with ecuflash and gm 3 port. I zero'd out the error correction to dial in the boost curve via wgdc alone, then logged 2 byte load to create the BDL tables, then added back the error correction and boost shot up by 3 or 4 psi, logged wgdc was a good 20% higher than it "should" have been (60.5 rather than 50.5).

I was trying to figure out if the internal load used for the boost error correction was off by that much or if there was some adder value applied to the WGDC maps.

So I'm reading correctly, your saying the ecu uses a difference internal calculated load from the 2-byte load correct?

And, we can change the load calc used to point to the 2-byte load correct?

 

Its a 2-byte load, but its different than the one that everyone is logging. Yep, it can be changed to the one that everyone is logging. Doing this will make the boost more consistent across the temperature range.

 

Hi

Have you looked at flattening the Air temperature correction map rather than change the boost load.

I'm still trying to get my WGDC logged as MUT 86 returns a constant value, rather than one that varies, and work out how what you've found fits into my JDM ROM.

MB

 

Not sure which air temp correction map you mean. If you are thinking of the main one, I would be leary of changing that table as I think it affects many things. And at the moment, the mods I've done are working perfectly for me. I think its just a matter of getting them translated over to your ROM. I can help you further with that. Can you upload your current .idb file to AktiveMatrix?