Sir, I might have a MBC (hallman) configuration that is capable of less than 21 psi (no spike) but not less than 19.5 at redline as the picture displays. I don't run this low of boost so I will have to double check.

The problem with MBC's is it is more responsive to psi as air flow increases which also increases with rpm. So at low rpms where static psi is higher in respect to velocity than at redline, the mbc opens and signals a WG which is already being loaded with more exhaust.

I basically tee'd a 20psi pressure relief before the MBC to soften it's response. The relieve contraption has a larger surface to be more sensitive also.

I'll turn the boost down to 20 and see what happens, if so I'll take a video and post

 

Based on what you are seeing, it sounds like instead of calling the table "Max WGDC", it should be called "Baseline WGDC".

 

I think it is a maximum, but also the starting position at a given RPM.

 

Yah, from my observations, the logged WGDC is what happens after the error correction kicks in. I suppose you can set the MWGDC to be the same as the logged WGDC post error correction. Or zero out the error correction, set the DC until you have your desired boost curve, then set the BDEL to your load at each rpm point, then enable the error correction.

 

<Request LogReference="WGDC" RequestID="86" Eval="x/2" Unit="%" Logged="y" Response="2" />


I just realized that people are frequently leaving out the Logged="y" (or Logged="n") item.. In versions 1.5alpha and earlier, it didn't matter as the Resonse="" item wasn't used as there was a 2 byte response checkbox that does the job.

Mitsulogger v1.6 beta 3 and newer honor that directive and its now required. You will get an error if the logged="y" is missing. Its actually a bug but I'm working on fixing that for subsequent versions.

 

My corrected logs definitely show MWGDC as being the starting point, but the logged WGDC can definitely surpass the values in the MWGDC table. My corrected logs are showing that the highest logged WGDC is exactly the "MWGDC" + maximum upward correction.

It seems to me that the ECU only knows how to deviate away from the "MWGDC" table based on the load error.

 

I've been analyzing my WGDC data this evening. At least for WOT runs, the "Max WGDC" table defines the baseline WGDC curve vs rpm. As the boost error correction algorithm makes changes, it just moves the baseline curve up or down. Also, the correction algorithm does not appear to be continuous in time or rpm. The algorithm seems to make adjustments about every 500 rpm or so.

 

Interesting. I really think it would be better if continuous, and also if there was an integral correction where it ramps up and down rather than simply adds the correction. Otherwise we have offsets.

 

Here's an example showing where the changes are made, and towards the end of the run where no further changes are made, the WGDC curve just follows along the BWGDC curve at an offset of 10.

 

I noticed you had 2 byte load on that graph, does that represent actual load?

How close is your actual load vs. your BDEL? I would assume from your graph that your BDEL is a bit lower than your actual load, and the TBEC is adjusting the WGDC to get closer towards your actual load.

 

Yes, the 2 byte load that I'm logging is one of the two load values used by the ECU. My BDEL with offset is 240. With the exception of the waviness, the ECU is doing a pretty good job of hitting the BDEL value.

Since the recorded WGDC values are always above my baseline WGDC values (I'm now using the term baseline WGDC instead of Max WGDC), what's actually happening is that my BWGDC curve is too low to meet the desired load, so the ECU is correcting my BWGDC curve upwards. Cool, eh.

Last night, I moved my BWGDC curve upwards to match the values that the ECU changing to during driving. I did a few runs this morning, and found that the ECU was not quick enough in detecting when the motor had reached a load of 240 during boost build up at low rpm, and the ECU was still correcting my BWGDC upwards which was allowing the load and boost to shoot way past target values. I hit a load of 270 at 3500 rpm and 25.5 psi in 2nd gear. I was kinda worried until I saw that the knock sum only went to 3. So now I've reduced the upwards WGDC correction limit in the boost error table.

 

Jack, could you post a modified XML including this? I dont have any program that will edit it to my knowledge. I used explorer to look at it, but I cant edit it I don't think.

 

XML is just text. Notepad can open/edit a .xml file. So can Word. If you can't get it with one of those, then let us know.

 

lol, just goes to show how much I suck at this code stuff, I'll get it guys, don't worry I'll give it a shot in notepad.

Edit: Did it in dreamweaver... the new Word doesn't like it.

 

I've been doing more WGDC logging, and it looks to me like the ECU does integrate over time or rpm, but there is no derivative which is causing late BCS response and load overshoot in 2nd gear. (Target load is 240.) However, I've got pretty good control over the load in 4th-6th gear where the rpms build more slowly. I'm trying to figure out how to fix this without having to slow the rate of spool-up. Anyone have any suggestions?

2nd gear pull:


4th gear pull: