I <3 Mrfred.

These are some amazing findings man. Thank you for taking the time to do this for the community.

 

This is good stuff mrfred!

I just wish boost control was closed loop, ie use the map sensor to feedback and control the solenoid.. maybe in the future hehe

 

So this is why we rarely see negative correction to the WGDC even when the load values are above desired ... changing the reference load will make a huge difference then ...

 

I've got another tidbit for you guys: load error logging. This is the difference between the target load and the actual load. It is the value on the left hand side of the TBEC table, and is actually more enlightening than the WGDC Correction because you can see exactly what's happening with the load. Its a bit more complicated to setup, but its well worth the effort if you want to have a go at it. Here's what to add to your 88590015.xml file:

----- (don't add the "-----")
<table name="Load Error RAM Address" category="Turbo" address="41E06" type="1D" level="1" scaling="Hex16"/>

<table name="Load Error RAM Address in Load Error Table" category="Turbo" address="7254" type="1D" level="1" scaling="Hex16"/>

<table name="MUT8B" category="MUT" address="3F0B2" type="1D" level="1" scaling="Hex16"/>
-----

Instructions

The "Load Error RAM Address" and "Load Error RAM Address in Load Error Table" will both read: 69CA. Change both to 6EE2 (again, you'll need to enter it as 0x6EE2)

Change MUT8B will read 6B41. Change it to 6EE3 (again enter it as 0x6EE3)

Load Error logging: In EvoScan, insert the following entry:

-----
<DataListItem DataLog="Y" Color="" Display="Load Error" LogReference="LoadError" RequestID="8B" Eval="5/8*(x/4-32)" Unit="%" MetricEval="" MetricUnit="" ResponseBytes="1" GaugeMin="-100" GaugeMax="100" ChartMin="-100" ChartMax="100" ScalingFactor="1" Notes=""/>
-----

A few notes: The units are load, so what you see will not be a % difference but rather "actual load - target load". When the value is negative, the actual load is below the target load. Also, the load error value will be wacky until you get on the gas and activate the boost control system.

 

Exactly. You'll mostly likely only see a downward correction in cool/cold weather.

 

So couldn't we counteract this by lowering the target a little bit? Is the difference constant?

 

Mrfred, I tried out the new values this morning on the way to work. I've got the load variable for boost control set to "6B42" but the WGDCC is trying to hold load ~10 higher than desired. I assumed (as seen in my post above) that changing the load variable would rid me of this phenomenon. My target load is around 265 and I'm holding steady around 275.

Also, my WGDCC log seems to just be an offset of the WGDC%. Is this correct? For instance, my WGDC% is ~69.5 and the WGDCC% is 64. I was expecting something like +/- 3.5 ... If this is correct, how exactly is the data to be interpreted?

 

The difference is not constant. That's the problem. FFFF6B42 does not vary with temperature which is good, but as the temperature goes up, FFFF6B48 goes down, in essence trying to push the boost up as the temperature goes up. Not a good situation. I don't think this was Mitsubishi's intention though. The stock boost system is designed to be limited on the high end by the restrictor pills. When we start changing restrictor pills and other things, this problem appears.

 

I hadn't tried logging WGDCC until this morning. Seems I need to spend a bit more time on understanding that variable. In the mean time, I suggest logging load error. That one I have tested. And if you know the load error, then you should have the WGDCC.

 

Let me just toss this out ...

For the WGDCC value, would "(x/2)-[WGDC]" give a more logical output value? So, in my case it would end up as:

(64)-69.5 = -5.5 WGDCC

Now that you've logged it Mrfred, maybe you could take a look and see if this could work.

EDIT :: Actually, that formula might be backward ... I'll have to double check the logs.

 

There is a bunch of math towards the end of the WGDCC routine that I breezed through too quickly. I need to go back and look at that stuff again to see exactly what's happening. The WGDCC logging I did this morning wasn't too helpful for me because I set the TBEC tables incorrectly. I'm going to try for a few more logs this evening.

 

Alrighty. I look forward to what you find.

 

Hey 'MrFred'... kinda OT a little but, you don't happen to have an updated version of your "EvoScanPlotDataHPTQ.xls" spreadsheet that would also graph load error would you? I was going to try and modify the spreadsheet that you posted on here previously, but then my head started spinning so I gave up.

Thanks!

 

I'll see if I can get something together. It will probably be a day or two.

 

Definitely no hurry, I was just curious if you already had something you were using. But if you get bored of being a Evo ECU guru. Of course I could always get off my lazy a** and figure it out myself too.