NefMoto

I am trying to figure out what is causing severe boost overshoot on the 2-3 gear change and beyond.  I am trying to build a low boost tune (22psi) based on the antilag no lift shift tune I posted.

As a background, this was an issue with my car before I switched to the hybrid turbos, and I always ran an MBC as a result.  I removed it and installed a new N75 and wanted to make this work the way it should.

My problem is, for example, a second gear pull, you can go WOT from say, 3k to 7k, and the actual boost matches requested boost perfectly, around 22psi.  The same thing occurs with a third gear pull - boost matches requested.

Do a WOT run in second, upshift to third, and I see a massive spike to 30+ psi which doesn't drop into the MAP range until 5500 or so (probably because the turbos can't move that much air).

My graphing spreadsheet is broken, so I will post logs in a bit.

Any ideas/advice is appreciated.

what does your kfldrl look like?  what is n75 duty doing.  just from what you've written i would guess your kfldrl table needs some tweaking. 

Isn't that only for open loop boost control?  Or does that work for closed loop as well?

(http://i.imgur.com/OPIjx.png)

once you get above the map sensor the n75 duty will mirror this map... start by scaling back everything above 3krpms... for ko4s the far right column past 3krpms should be between 80 and 85 to start.  bear in mind that if you have changed wg pressures they will react differently... i would scale a few other parts as well, but you can feel it out with logs and driving impressions.

Thanks, I will give it a shot...  Boost control is one thing that has always irritated me, so I've just been lazy and ran the MBC.

You need to tinker with KFLDIMX and KFLDRL  ;)

Trust me, it works.

KFLDRL seems to have reeled in the overshoot.  Good call on KFLDIMX - it is overshooting slightly - I take it this has to be scaled DOWN as well?

Now I get to figure out why my fuel curve looks like something from the redneck back yard truck axle dyno. :D

Yes. i would have started with KFLDIMX first to be honest though.

I have a customer with a GT28RS running @25psi with N75 ;)
Guy wanted aggressive tune haha.

I'm still overshooting and then oscillating.  I am worried if I start tweaking the PID that it's going to affect boost control everywhere - something which currently works just fine?  It's worth a try I guess.  Going to reduce P a tad to see what happens.

(http://i.imgur.com/b1dG1.png)

After lowering P slightly.

(http://i.imgur.com/FKJKb.png)

Not sure if I should try to correct the overshoot with KFLDIMX/KFLDRL or lower P a bit more.

adding 10% to I:

(http://i.imgur.com/GgJzo.png)

wow that is fantastic.  do you feel its done?  please share the maps after its all cleaned up!

I definitely want to get it cleaned up a bit more.  Today I was a little too aggressive with I and things became unstable.

(http://i.imgur.com/JQIoo.png)

Jason,
Have you tried logging the PID variables to see exactly whats going on?

No, not yet.  I was thinking about that today, and how I could probably use the output with some pid tuning software.

I usually leave a little spike in when I tune the boost. Its just more fun that way and you get faster spool than trying to get rid of the spike  ;)

Well, normally I just use an MBC and set my boost to the "O" in "VDO" on my gauge :D

That works  ;D

lol, this is the best method i've seen yet.  ;D

You may want to look into both the dynamic and stationary "I" values.  I find the dynamic is best for tuning the initial ramp and settle to target and then you have a separate method for when you're maintaining a near constant level.