shouldnt error correction help with taper by adjusting WGDC all by itself? lets say you increased top numbers in the correction table to like 20% (from 8%stock) wouldnt that help? That also probably means that you shouldnt exceed 70-80%wgdc or there will be not enough room for correction...

or does it just not work for anybody and that's why it's beeng adjusted by wgdc instead of desired (plus offset) boost tables?

 

It seems that the issue with error correction and the modded boost pill is that when you make the pill smaller, you also increase the response time for the wastegate to activate. This results in an erratic boost curve while the WG tries to catch up to the load tables and it basically gets stuck in a loop of over-correction for a while.

The approach I'm going to try is to only allow for correction if the load value is below target. This should allow the WG to correct and raise boost but not freak out when the boost is a little high.

 

Alrighty ... I got bored a little while ago and decided to try out my new boost settings.

Here are the settings I logged ..



(Don't ask about my boost pill because I couldn't tell you what it is. I couldn't find the correct bits so I ghetto'd one up that is tapered from about 1.4mm to 1mm.)

You'll notice that I'm using the technique I mentioned in an earlier post. I've increased the offset value to allow lower desired boost and free up headroom in that table. I've also ramped up the desired boost cells slightly toward redline. The boost error correction table still allows for correction if the load is low, but not much correction for overshooting the desired load.

Now, for the logs ...

3rd Gear Pull

4th Gear Pull

(I know I'm running rich)

If you look at the WGDC on both logs and compare it to my ROM, you'll see that I never hit my set max WGDC cap. Therefore, the slight ramp up I've left on the max WGDC table does not affect the boost curve at all. I mentioned this in an earlier post as well.

Here are the loads for the 2 logs above compared to my fuel map ... (green is 4th and black is 3rd)



Here is a comparison to a previous 4th gear pull and it's correcponding fuel map. You'll notice that I richened the spool up area in the map above and spooled a little later. The curve is also smoother, likely due to the tweaking of the boost error correction table. (faded green is old 4th log)



Comparison Pull

So, now the plan is to make the ramp up on the desired boost table a little more drastic and try to eliminate a little more taper on the load curve.

Titanium, is this along the same lines as what you are doing to tune your boost?

 

Similar yes, But mine is at this moment A little simpler...
Here's today's current results

Log #1
Baseline run. A/F didn't log for some odd reason. a reminder: #58 pill, WGDC at 90 flat. Max load in third 258 @ 3400 to 230 @ 5500 to 205 @ 7000.

Log #2
I richened up the fuel mixtures all over the 260-240 load value to account for some knock at 4.5-5.5k, as well as adjusted WGDC to be 90 from 1k to 3.5k then 100 from 4k to 7k. A/F again did not log, but knock was gone. Boost curve about the same, as Bubble has before mentioned.

Log #3
This time, I only adjusted the "Boost Limit" table, which was at 255 flat to 3k-3.5k @255 and 270 the rest of the way down. the result? WHOA!
The car seemed to have a second life on this run. Log showed that the boost actually peaked TWICE, once at 3.4 @ 258 as before, dropped to 4k @ 240 and again at 4.5 @ 269, and tapered only to 215 @ 7k. Knocked twice at this new 4.5 peak.

Log #4
Richened the HO fuel table at all 240-260 cells again to prepare for less taper. I now smoothed the WGDC table at the 4k cell to 95 and smoothed the Boost Limit table to look like so:
255
255
260
265
270
275
275
275
275
However, my "tune strip" was comprimised by traffic so the only logging I did was a quick 3rd pull on the way home, which yielded similar results as Log #3, but with less dip at 4k and less taper to 6k, which dropped sharply to 7k @ 220.
A/F on this run was beautiful though, 11.5 throughout to 11.2 at 7k.

More to follow.

 

Hmmm ... interesting. If you are getting higher loads by adjusting only the boost limit tables then it would seem that you are hitting the boost cut in the higher RPMs. If the limit was at 270, increasing it shouldn't matter unless your load is trying to hit above that number.

Post up some logs so we can see what is going on.

 

Now that I look at it, my numbers were slightly off in my post. I actually hit 271 load at my second peak. Anyway... Here is what I beleive is log #3 from my post. I'm not going to post my baseline or the last one because they are fairly erroneous.


Err... It would appear my computer is throwing a fit about attaching the zip...
leme see what i can do.

EDIT: Yeah, apparently my computer sucks. Or I have a firewall problem or something, so I can't post the log. Might be able to e-mail it, bubble p.m. me if you want.
----------------------------------------------------------------------------------------------
I can't go back out yet, but I'm bored and have been thinking about everything that has been said and looking at the tables in Ecuflash and I've come up with a theory.

The mitsubishi engineers designed the ECU to rapidly come to peak boost (USDM 21 psi), taper slowly through the climbing RPM range and settle low, where engine loads are high to improve durability. That said, the method they apparently chose was to let the Waste gate duty cycle be completely free to adjust as the ECU see fit, this is why stock cars all ship with 100% Max WGDC tables, but set up a system of "well at this PSI we see this load, the ECU speaks in terms of load, so if it see above this "X" load the ECU with increase the wgdc to slow the turbine." This allowed for tapering because the "Boost Limit" Table is ramped similarily to the Load curve we commonly see.

By modifying the waste gate oriface, we are modifying the "signal" that the waste gate receives from the BCS, allowing the waste gate to operate at a lower duty cycle while the BCS is sending the SAME signals. With no ECU changes, this creates a much higher boost profile, because where the "Boost Limit" table is offset.

However, to control our boost, we though Oppositely of the mitsubishi engineers. we decided that the wastegate should vent as much as the Max WGDC will let it, effectively ignoring the "Boost Limit" table, as it is now a more relative value because the Wastegate sees less of what the BCS is telling it. This means because my Max WGDC is at a 10% free play, adjusting the "Boost Limit" table had a drastic effect on the 100% max wgdc area of my curve.

In conclusion, 2 routes could be taken under my theory to control boost and taper.
1. Set Max WGDC flat 100%, find resulting desired load corresponding to "Boost Limit" values and set them accordingly.
2. Max out "Boost Limit" tables and curve WGDC to reach desired boost and taper.

How does BDEL fit in to my theory? It doesn't. I have no Idea what correlation it has.

 

I'm not sure that your thinking is 100% correct. As far as I understand it, the boost limit settings work with the boost limit timer as a safety feature. Basically, the boost limit table determines the maximum acceptable load. If the values hit above that limit for longer than the timer allows, then the wastegate opens to redirect exhaust and drop the boost to protect the engine.

The boost desired load tables affect this because they are what control the load values. Jack up the boost in the desired load tables too far and you will exceed the limit and hit the boost cut.

This can be shown by adjusting single tables on a stock car. Modify only the boost limit settings and you will see no change in the boost curve. You could also lower the boost limit to really low values just to see the effect they have when you exceed the values.

You are correct in how the max WGDC fits in. It is set to 100 and allows the BCS the freedom to control the boost. The stock boost pills control the maximum limits of the system.

Don't quote me on this ... it's just how I understand it and I'm no expert.

 

Actually, the WGDC is capable of up to 127% or so. Basically, the limit is equal to the stock settings of 100% WGDC + the lowest boost correction value. there is debate of how much this +xx% can really be effective though.

I'm not sure why that tables doesn't affect taper. You can take a look at the ROM values and logs I posted above and see that it doesn't work. I've retained the +% values in the correction table, but the WGDC doesn't increase to correct for the taper even though it's below target load.

Maybe an expert can weigh in on this?

 

Would be nice, even though at this stage of this newbie "boost control blog" with no other expert input, i don't expect it to start now. It's a good learning experience.

Even if what you say about the boost limit table is true, the fact that it did create a late spike on my log shows that it could possibly be used to tune boost and taper...

I will also play with the BDEL when I go out to see it's impact, although I have heard that the 159.4 value is a maximum, which would eliminate increasing it at higher RPM for de-tapering.

 

I FINALLY got out and got a lot of logging done. I'm not going to break in down to a log by log basis because there were just too many, but I'll give you a basis of what I learned.

I was able to lengthen the time at peak boost from 3.5k spike to a 3.5-4k plane by playing with the Boost limit tables, specifically by raising the boost limit of these 2 values while holding the same MWGDC. The boost would then spike at 4.5k and plummet at 5k, then tapering as normal.

I tried to compensate this spike to lengthen the peak boost plane by adjusting the MWGDC, but it had little effect. At this point it seems the boost limit table is the only table having an effect on my boost control. After several of the same logs with both minor and drastic changes, I give up on Boost Limit and play with BDEL, which also had no effect but lowering the 3.5k boost plane.

BUT!
I DID SUCCEED!
At current, my Load curve "peaks" at 3.5k, and holds at 260 +-5 load to 5k RPM, in long and in short ELIMINATING BOOST TAPER.
How? I made one single change that caused this dramatic improvement, Boost Delay Timer. I began thinking about what bubble said about the boost limit table only coming into play when the load was over the table for the amount of time in the delay timer table. By setting the delay timer to 10000 ms (10 seconds) The spike to 270 and the resulting "canyon" dissappeared in a single flash. From there, I ramped the MWGDC table to the nice, flat load curve my car currently runs. From this point I tested it in gears 1-4, where it solidly and consistantly loads at 260 from 3.5-5k before tapering to redline. I then logged a trip around the area, occasionally flooring it in strange conditions to test for negative side effects of no delay timer...but there was none.

Why does it still taper from 5k? I'm almost 100% sure this correlates to the fact that my MWGDC 100%'s at this point to redline. With a smaller oriface, I'm sure I could gain another 1k-1.5k of peak boost before taper. I'll have to try this with #59 tomorrow to see.

EDIT: After some more thinking, I've come to the conclusion that it may be possible to leave this safety feature in play for odd scenarios by raising the entire Boost Limit table to a much higher value such as 290, but I'll have to double check before I say for certain.

 

Do you have logs of your WGDC? If so, then does the WGDC you log match your Max WGDC? If yes, then you are using the technique I mentioned earlier.

This is what I assume is happening based on your description ...

You are requesting more boost than you really want and limiting it with the Max WGDC table. You've eliminated the taper by ramping up the Max WGDC (which allows boost to rise toward the higher, dialed-in load) and increasing the boost limit so it simply doesn't intefere. Increasing the delay timer finished off the job by getting the boost limit completely out of the way.

I still stand by the boost limit table being only a safety cap.

In related news, I've found that ramping up the boost desired load tables does not help reduce the taper. I ended up with the EXACT same curve in my latest log, which had increased BDEL cells toward redline. This makes NO sense to me, but whatever ...

I guess tomorrow I'll play with the boost limit table and test Titanium's current findings on my own tune.

 

No I don't, but I will start logging it today. I'm certain this is the case. I'll also try raising the boost limit to a much higher value than I'm hitting and re-applying the timer to see if that works.

Also going to try the #59 pill though.

As I was slowly dialing in my AFR at the same time, I noticed at previous peak load (3500) that my HO Map is at 7.4, which i think is a maximum rich value. This worries me because this might indicate that my injectors might be near their 100% duty cycle and become an issue, but again i'll have to log for it to see. I know lean spool IS counter-acting this effect, but I need to know if i'm running into a minefield.
----------------------------------------------------------------------------------------------------------
Logged WGDC and set my Boost limit table to 270 flat with 1000 delay timer, no effect on curve. Yay! As I suspected, WGDC followed my table to the T and is the reason my boost drops after 5.5, however after looking at the curve I noticed it actually had very little taper after this, only down to about 220-225 at redline, so I think I'll skip the hastle of fabbing a #59 oriface for now and call this chapter of my tuning done.

Oh and I found my injector pulse width is 16.6 at 3500rpm, I don't know what this translates to in terms of DC.

Next Chapter: Timing Advance.

 

By not enlarging or removing the orifice at the BCS, you are limiting the boost correction that can be controlled with the BCS. I've taken a somewhat different path than you, in that I left the pill at the turbo unchanged and replaced the orifice at the BCS with a bleeder (a hole that bleeds to atmosphere). This this setup, by ramping up the WGDC, I can maintain a load of 240 from roughly 3100 rpm to 6000 rpm. I could maintain it a bit further out, but the engine starts getting a bit knock prone as the turbo gets further out of its efficiency range.

The general algorithm for the boost routine is that the ECU tries to maintain the combined value of the BDEL and boost offset load. The ECU will try to maintain the desired load by altering the WGDC within the correction range in the WGDC correction table. The correction applied to the WGDC is not continuous though. The ECU seems to make adjustments about once every 0.5 sec or something in that range. The boost limit load should only come into play when the engine load exceeds the boost limit load.

 

Ok this post is a little late, but I've made a few more changes.

I noticed my load curves still had a slight peak at 3.5k and dropped to 250 then never got above 255 again... so I changed the smoothing of my MWGDC table to 97% at 4k and 4.5k, and it still dips at about 3.7k but returns to 260 and holds.


TIMING

This was really tedious and intimadating. ESPECIALLY with 3 timing maps on the IX.

Log #1
I logged Load, RPM, Knock Sum, and Ignition advance. At stock timing, I had one or two counts of knock...one during spooling the other during midrange boost...about 5250 or somthing like that, which did slightly pull timing. I lowered the timing cells in these 2 spots by 1 degree, then raised the timing by one degree on all other load points.

Log #2
Flatline Knock sum at 0, yay. I raised timing at all loadpoints by 1.

Log #3
Same result...car is actually feeling faster, and in-boost but not peak boost response feels better.

Log #4
Again, raised all loadpoints by 1...I'm running 11 degrees of advance at 260 load at peak...

Log #5
Knock EVERYWHERE! 9 counts at 4k tapering to 1 at redline. What happened here? Didn't raise timing...re-test

Log #6
Knock gone again. Raised by 1 everywhere, but had to go home. Logged the trip home to find it was knocking in gears 4-6 under load, dependable 4-7 counts. Shoot. I think i'm going to lower the 240 and 260 load cells across the board down 2 degrees and re-test the high gear result...results to follow.

 

Are you logging 2-byte load or calculated load?