this

 

Oh. OK.

 

EDIT: This was not on 1.4.1+ Still good EBC, but different tuning process.

This is the best I could ever get out of EBC, PID closed loop. It moves around up and down about 0.8kPa from target. No filtering at all on the MAP signal being plotted. It is set to hold valve closed until 20kPa from target for best spoolup. No overshoot, no undershoot.

EDIT: Top plot, red line is MAP, white is RPM. Sorry, pic didn't capture trace labels.

 

Still messing with it, but it's getting better. I don't even think I'll mess with advanced mode at this point.

5th gear pull now overshoots just a bit but nothing crazy.


I'm on my stock wheels with all season tires, so 1st and 2nd are useless, but 3-4 and a touch of 5th gear look great.

 

 

A little overshoot is normal. I think that's just the algorithm working. In fact, from what I've seen it's actually what you want. I'd only worry about it if it causes overboost or oscillations.

Those logs look so much better!

 

Overboosting is not what you want. The idea of boost control is to control the level of boost. A boost spike right at spoolup in the midrange is exactly how you break stuff.

 

Was this on firmware version 1.4.1 or higher?

 

Oh, not it wasn't. I guess this thread is specific for that version?

EDIT, yeah, sorry, it's in the title. Regardless, spikes and overshoot/undershoot are not normal or good.

 

Yeah, the algorithm and setup process changed for that version. PID tuning is still similar but it uses an ideal algorithm and bases off of a bias table.

 

This is going to be a picture dump of (3) settings on my EBC. First is the settings themselves, with Max % at 75%, and Derivative at 20. Note that I have a large LLDelta as I want the controller controlling as much as possible.

Then snapshots of logs at:
65% & D=10
75% & D=10
75% & D=20

I'm not sure I saw much difference, but look and see if you think the performance is what should be expected. This is with a baby Mitsubishi turbo.




General Settings



Bias Settings



Targets

Then throttle punches at 4K RPM with the (3) settings. It may be the D=20 spools a little slower. Conditions were not identical between runs. I have overlaid the target and kPa curves.



4K punch at 65% and D=10



4K punch at 75% and D=10



4K punch at 75% and D=20, starting at low kPa



4K punch at 75% and D=20, starting at 100 kPa

Largest overshoot is 10kPa. Should I try for less? I'm thinking I'm OK here.





65%, D=10, 2nd gear run. No overshoot.



75%, D=20, throttle play. Note target and MAP track pretty well.



A 1st and 2nd run-up at 75%, D=20




And last, A long pull. you can see the Boost Duty moving to try to keep SP and PV (Target and kPa) together. To me, that is what we are after.

EDIT: I had intended to only keep the top graph from each run as that is where the info is.

 

When I said overshoot, I meant like things going over by 5-7 kpa max. Not overboost.

Then again, that's what overboost cutoff is for!

 

I'm going to try and contribute my specific setup to this thread, hopefully we can build out a solid technique.
Car is a 2004 Mazdaspeed with Begi downpipe, FM Little Enchilada and a fullwidth ebay FMIC.
Ecu and boost controller setup is a recently built MS3Pro PNP and a Mac valve for boost control.


I'll start posting map shots and logs when I can run more tonight, last night got it working but needs lots of massaging.

 

LOVING all the input/contributions, thanks guys.

 

This is a little different direction, but I seem to remember aidandj mentioning a while back that he runs multiple boost settings. Is that something that's relatively easy to set up? Right now I use both wastegate 160kpa pressure and EBC @ 210kpa, but I'd really like to set up a KILL MODE for when I want to break something.

 

Boost control settings->Boost table switching

You can have it switch between tables or blend. I have it blend between tables. I believe I have a thread about it already.

Edit: found it. https://www.miataturbo.net/megasquir...control-86672/

 

I really appreciate you feeding me like you were my mama-bird.

 

With the help of our intrepid MT.net folks, I battled my way to a successful set of pulls at just under 12psi on my MSM using the MAC valve.

I'm still in "Setup mode", with my tables setup as follows for reference (test your way into the right numbers... mine aren't even done yet... but they work for ME).

Boost Target Table is setup for 170 across the board.
Boos Bias table is setup with 70 in all cells except for the last 3 higher RPM ones setup at 80. That is giving me a solid boost hit and hold on my setup currently.

One of my initial issues was not moving the Bias table up enough, I started doing it in 10 jumps across the entire table until I got to the boost level I wanted (it then tapered down through higher RPMs).

edit: Forgot to mention using a delta of 30 at this point

 

Yeah once you get the hang of it, it really works well.
The latest Miata I finished up, Eric (darkwood1) also runs a mac valve with a 2560 and has rock solid boost in both hot and cold set up with the process we outlined here and then with mild PID settings.

 

18psi, Curious. What would you describe as mild PID setting/slider settings?

I'm trying to get a reference on when I have this working as well as it should be.
My settings seem to be working ok but I still have an 8-12 kpa overshoot depending on how I apply the load.
Thanks.