I set my duty cycles today this evening and ended up with Min: 17%, Max: 90%.

I went out to make some pulls, results below. I'm having massive issues with integral windup, this was with 40 for the delta, 4th gear (6sp). It's bad enough the sensitivity slider is useless since it overboost regardless of the setting. I've already got a conservative boost duty table set up, so I skipped "setup mode" and went straight into "basic mode".

**Actually now that I look at the screenshots, maybe it's getting a little better with increased sensitivity. Seems like it's staying out of boost cut longer each pull.**









I also made pulls in Open Loop and at Wastegate pressure to demonstrate what the boost duty table is commanding, as well as that the boost controller is working.





I noticed 18psi's settings have the solenoid frequency at 26Hz where mine is at 39Hz, is my setting wrong and possibly causing an issue? It's a DIY Autotune EBC.

Any suggestions are welcome. I guess I can go ahead and try the lower delta, but I've been down that road and know what issues it opens up.

 

Aidan, right, I understand that. I'm just trying to figure out why you're overboosting like you are. Is it a function of things just not playing nice with the EFR wastegate?

Pdexta, see if my EBC frequency works better for you. Aidan has written that the lowest value makes it EBC easier to tune.

 

He doesn't have an EFR. Like I said before its "windup" the control loop is trying to hit something that is physically impossible, so it keeps trying harder and harder.

pdexta. In the latest 1.5.1 betas they added logging for the actual P, I, and D calculated values. If you have the ability to SDCard log you should give it a try.

http://www.msextra.com/forums/viewto...6c352&start=20

Also. I have found that even when the solenoid is making noise at 90% DC, it isn't actually affecting spool until down at 60% for me.

Until I see multiple setups where the "noise test" and the "spool tests" are identical I won't trust just listening to the valve.

 

He has a DIY ebc. You've written elsewhere that the lowest value that works for the freq is easiest to tune, no? That's why I'm using 19 AFAIK.

 

From the DIY ebc page, "It operates on a frequency of 19.5-40 Hz and is practically a wire-in replacement for the now-discontinued GM EBC Solenoid."

Also I wonder what keeps my setup being so easy to tune.

 

Can you do 4th gear pulls to 22psi starting at 1500rpm and starting at 5k rpm? If not your setup is irrelevant. 10-15psi is easy to tune. Things happen slower, and you aren't that far above wastegate pressure.

 

I can post a log of 50psi pressures just from thsi week

not a miata tho

 

 

On 1.4.1+? With an internal gate?

 

Fair enough

 

Don't know what log you're looking at but its not the one I can see. There's no wind up. You can tell by the way the boost duty is getting lower not higher.

 

Compare his open loop log vs closed loop log. It is not dropping as fast as it should be. I bet if you data logged the PID parameters you would see the integral term winding up.

 

I'm more interested in comparing his wastegate and open loop logs. They look near identical.
Its likely that the overboost in closed loop is caused by the initial duty cycle value being set too high.

 

Um one is like an extra 4-5psi?

163 vs 190

 

This was perhaps an enlightening moment for me. What if "max" wasn't "max duty cycle that you can hear the solenoid clicking" but instead "max useful duty cycle for your given setup"?

The open loop duty cycle for my target boost is 27. Setting "max" to 30 creates an artificial ceiling that prevents windup. I gave it a try and after struggling with this for over a year the result was quite a beauty.



I'm sure it gives up a little spool, but nothing I can't live with. From the log it looks like I can increase my max a little higher to help with that and bring it up to target quicker. Now I'm going to start bumping up the sensitivity slider to be more aggressive. This definitely seems like a great solution for those of us struggling with windup issues.

 

Glad to see it. Setting min and max for me was an enlightening experience. It makes things a lot better.

 

This is exactly why my max is set to 60 instead of 86. Think of all the extra work the PID algorithm has to do if it's oscillating/calculating a value 20-30 higher than it needs to be.

I feel like the reality is that the 'max' value is something that you can find a happy medium with. I.e. faster spool, without overboost. The sound test just gives you a starting basis for what the operational (read:not functional) min and max are. I started with a pretty narrow range so as to get a better hold of things.

I think my max could be a bit higher and my table could be set up to be a tiny bit higher since i undershoot at target. But it works pretty well so....
I do wonder how things would be affected if I was running 2x as much boost or something. Unfortunately my ***** aren't big enough to try that haha.

 

In theory, it defeats the purpose of EBC, since it doesn't work much differently than an MBC. In reality, the difference is small and you can't argue with the results.

 

I think EBC has a lot of advantages over MBC, other than holding atmospheric pressure in the wastegate actuator:

1) Ability to have boost rise with RPM
2) Ability to have boost target move with throttle position
3) Ability to change the settings with a computer, and know what you changed them to (once dialed in) without even consulting a boost gauge
4) Ability to have key-switch actuated high and low boost range

 

^ I think Savington was specifically referring to my low max setting, in that respect he's correct. My low max setting solved a major issue I was having, but it makes my ebc work a lot more like a manual controller.

But like rtc said, it's all about finding a happy medium. I can certainly gain back a little of the spool I lost. For now I'm just ecstatic to feel like I'm really making some progress after a whole lot of failures.