Erat

iTrader: (5)

Thread jacking here.

I have that wacky unit that you built, and got hacked up. My EBC is hooked to the stock purge valve that is no longer in use for ease of wiring.

wolfram

I am only questioning these other numbers based on what i have seen. My boost duty when it starts pulsing up and down one or two psi in the 9-14 range shows a direct relation to the boost duty going from 0 to 100 very quickly back and forth. Wondering if there is a better way to shrink the upper and lower bounds of the PID equation to get any "smoother" boost duty. First time tuning a EBC so feel free to tell me i am full of crap. I am open to all quality input.

Erat

iTrader: (5)

How do i set my boost control pin to PA0 when it's not an option?

wolfram

So i think i have made some progress. I now have 8 as a P value (I guess i never bothered going low enough). This now spools quickly to 12psi, then stays pretty steady for a while and begins to creep further in the rpm range. Does this value make any sense? It actually runs pretty well with P at 8, I at 0, and D at 15. Are these values reasonable, or am i still lost?

Braineack

iTrader: (62)

add I and see if it helps combat creep.

wolfram

Hate to beat the issue here, but i cannot for whatever reasons find a happy place with any of the PID values. The 8 to maybe 10 values for P were consistent in 2nd and 3rd gear to not over shoot, but when i make subtle corrections to I and D to get it right 3rd, 4th and 5th gear pulls go shooting way past. So then i start over with a higher P that is consistent for 4th and 5th gear and my 2nd and 3rd gear boost is jacked. What am i missing, or what else can be done? I am about to go back to my ol' ball and spring...

henksmiata

It drove me nuts too, so yeah thats what I did back to MBC

wolfram

Hope someone is still following this thread. I made it back to closed loop tuning after i got open loop working well. My problem is now combating the boost oscilation. Works pretty well, but my boost duty curve needs to be smoothed. Been a while since i have done any "real" math, but what PID value will smooth this out? I assume its the derivative as it should smooth the slope of the curve. PID values are as follows:

P - 17
I - 18
D - 98

Attached is my log. I had a lower D value around 20 and kept increasing to see if it helped. It did not. Am I working with the wrong value for this? Should i be increasing I instead? I would guess if i increase I i'll need to bump P along with it. How much I and D is too much for the P value of around 15-20? These P values are what is necessary to get it to get up to the desired target, so I imagine my I and D values would be smaller in proportion to someone with a much higher initial P value. Thoughts?

wolfram

Here are my BC settings. I have the DIY EBC. Their documentation says to use 20ms and the 19.5Hz settings. Should i be working with something else? This seems to work fairly well in open loop, so i was hesitant to change them for closed loop.

wolfram

For what its worth, i may as well attach my boost table... LOL.

Should i have gradual build up for this, instead of just pegging it at a decent TP? I figured for testing i would at least have it smooth across the board.

Any input would be great, Thanks!!

Canyonfive

I love how these threads always end with no resolution. So what did you end up doing?

Mr.Pibb

Hello.
I had a similar complication tuning CL.

This may, or may not help, but just start from post #14
https://www.miataturbo.net/megasquir...ng-tips-71182/
It explains how to get it where you want it in a very understandable way.

You will most likely have overshoot no matter what.
But by how much? I'm overshooting by 1-2psi. Can u live with that?
Slow climb it's dead on, but you will see it when downshifting mostly.

Using KPA it looks huge, but it's almost nothing.
Perhaps you can compensate by lowering target slightly and raise it back
to sorta smooth over the spike.

In my experience it's difficult to completely tune out spike.

If you completely tune out the spike it will oscillation, then tune out the
oscillation and it will spike.....

Both of those settings will be fighting each other it's a never ending cycle,
so just start from scratch and get it smooth and if it's 1-2psi spike I think it's OK.


My settings:
P-21
I-30
D-140

Some of the numbers are allot high only because I was chasing my tail on
the unbreakable oscillation/spiking cycle for a bit.

Then you can have fun like this after you get it down solid.

wolfram

I have left it on Open Loop for now. The open loop seems to be running really well. No overshoot and rock solid. Maybe later this summer once it warms up, I'll give it another shot.

Mr.Pibb

Just take the little time you need to test it out.
CL should be the only option.

You tune it in 4th gear and it'll boost overshoot in 5th
and get very low boost in 1-2-3rd.

Or if you tune it in 2nd gear and decide to give it a nice 150mph run it'll
over shoot like crazy.

Braineack

iTrader: (62)

yeah, youll want to when youre down 3-4psi when it's warmer due to your laziness.

Mr.Pibb

Yes, that too.

Another thought is tuning on the freeway with all that resistance and finally getting it
to a stable 20psi, then on the dyno, boost drops down to 17 perhaps.

It did on mine anyway. Not very consistent is the point.
CL is completely bad-***, holding target all the time, even while smoking the tires.
The car will be faster just by getting constant boost. Hope this is convincing.

wolfram

I guess i was just waiting for some decent (consistant temps) here in Denver. So if try to tune it again for CL, the ambient temps are almost irrelevant right? I am putting in a new IC soon, so i think i'll wait after that anyway...

Braineack

iTrader: (62)

ambient temps are irrelevant.