MS2 Boost control - Am I missing something? - Page 2 - Miata Turbo Forum - Boost cars, acquire cats.

Welcome to Miataturbo.net   Members
 


MEGAsquirt A place to collectively sort out this megasquirt gizmo

Reply
 
 
 
LinkBack Thread Tools
Old 05-17-2015, 05:32 PM   #21
SADFab Destructive Testing Engineer
iTrader: (5)
 
aidandj's Avatar
 
Join Date: Apr 2014
Location: Beaverton, USA
Posts: 17,992
Total Cats: 1,457
Default

Fair warning though, in the next release the boost control will be different, it uses ideal PID instead and will have a slightly different tuning method. No initial value table, instead it has a bias table.
aidandj is offline   Reply With Quote
Old 05-17-2015, 05:35 PM   #22
Sadfab Union President
Thread Starter
iTrader: (3)
 
Join Date: May 2014
Location: Kansas
Posts: 2,690
Total Cats: 113
Default

I understand the premise of the initial value table, I just do not like the application of it in relation to closed loop idle entry, it is never as close as I expected it to be. I've been using last value forever now and it seems alright.

I anticipate initial value on boost control to be similar, so I am happy they are changing it. Only reason I was messing with it is because I wanted the stupid valve closed at idle, heh.
deezums is offline   Reply With Quote
Old 05-17-2015, 05:39 PM   #23
SADFab Destructive Testing Engineer
iTrader: (5)
 
aidandj's Avatar
 
Join Date: Apr 2014
Location: Beaverton, USA
Posts: 17,992
Total Cats: 1,457
Default

It really is helpful though, it makes a pretty big difference in being able to lower your lower limit delta, which means better spool. In the beta the initial value table is now a bias table, so it is part of the closed loop calculations all the way through the pull.

Valve closed at idle is useless, really the valve can be wide open until the time that your wastegate would start to open.

Make sure that you tune the max duty cycle. It made a world of difference. It really helps almost eliminate integral windup.
aidandj is offline   Reply With Quote
Old 05-17-2015, 05:47 PM   #24
Sadfab Union President
Thread Starter
iTrader: (3)
 
Join Date: May 2014
Location: Kansas
Posts: 2,690
Total Cats: 113
Default

Are you certain that using the initial table is giving you best spool in all circumstances though? Maybe it does so for a easy 2-7K 4th pull, but I don't see the resolution required to properly stage the valve like I need too in all cases.

When I was trying that with the idle valve I always seemed to overshoot the target and have a unnecessary high idle. If I try to push the envelope, aggressively tune straight at idle speed I want (or spool I want) I'd end up with garbage on both sides, I'd still be idling high occasionally, and I'd start dipping the idle as well, a new problem.

I don't see it being too different for EBC, except with worse results than a high/low idle, so I'd rather just turn the dashpot up, or the limit delta, and let the maths handle the odd cases for me.

Max duty cycle? You leave yours at less than 100% DC?
deezums is offline   Reply With Quote
Old 05-17-2015, 05:50 PM   #25
SADFab Destructive Testing Engineer
iTrader: (5)
 
aidandj's Avatar
 
Join Date: Apr 2014
Location: Beaverton, USA
Posts: 17,992
Total Cats: 1,457
Default

Quote:
Originally Posted by deezums View Post
Are you certain that using the initial table is giving you best spool in all circumstances though? Maybe it does so for a easy 2-7K 4th pull, but I don't see the resolution required to properly stage the valve like I need too in all cases.

When I was trying that with the idle valve I always seemed to overshoot the target and have a unnecessary high idle. If I try to push the envelope, aggressively tune straight at idle speed I want (or spool I want) I'd end up with garbage on both sides, I'd still be idling high occasionally, and I'd start dipping the throttle too.

I don't see it being too different for EBC, except with worse results than a high/low idle, so I'd rather just turn the dashpot up, or the limit delta, and let the maths handle the odd cases for me.

Max duty cycle? You leave yours at less than 100% DC?
With the current code your using, the initial value is only used when entering closed loop. So each initial value can be tuned to get max spool when starting at a higher rpm.

Yeah. Most of the time you can lower the max duty cycle a lot without effecting spool. I was able to lower mine all the way to 65% with no change in spool. This makes a huge difference because it lets the PID equation work inside a smaller window. The equation wont be trying to go all the way to 100% when it only needs to go to 65%. So you will end up with the closed loop winding up and then you get overshoot, or over correction. By limiting the min and max duties you can give it a smaller window to work in.

Its just like how we set the min and max idle valve duties.
aidandj is offline   Reply With Quote
Old 05-17-2015, 05:57 PM   #26
Sadfab Union President
Thread Starter
iTrader: (3)
 
Join Date: May 2014
Location: Kansas
Posts: 2,690
Total Cats: 113
Default

I see...

You are right though, I need to find best ever possible spool and log that then just work my way back to it. I don't really think I have much of an overshoot problem though, I'm not sure how and it might just be my exceptionally terrible wastegate and downpipe setup but I don't seem to have overshoot issues too often.

I've been all over the overboost protection though hehe
deezums is offline   Reply With Quote
Old 05-17-2015, 05:57 PM   #27
SADFab Destructive Testing Engineer
iTrader: (5)
 
aidandj's Avatar
 
Join Date: Apr 2014
Location: Beaverton, USA
Posts: 17,992
Total Cats: 1,457
Default

You will just get overall better results if you limit the window closed loop is working in.
aidandj is offline   Reply With Quote
Old 05-17-2015, 06:01 PM   #28
Sadfab Union President
Thread Starter
iTrader: (3)
 
Join Date: May 2014
Location: Kansas
Posts: 2,690
Total Cats: 113
Default

I'm gonna go reset my preload and go log some stuff, see if I can't get this dialed in.
deezums is offline   Reply With Quote
Old 05-17-2015, 06:02 PM   #29
SADFab Destructive Testing Engineer
iTrader: (5)
 
aidandj's Avatar
 
Join Date: Apr 2014
Location: Beaverton, USA
Posts: 17,992
Total Cats: 1,457
Default

Take lots of logs, and welding gloves for changing preload on the fly
aidandj is offline   Reply With Quote
Old 05-17-2015, 08:33 PM   #30
Elite Member
 
Join Date: Apr 2010
Location: Newcastle, Australia
Posts: 2,638
Total Cats: 25
Default

Open loop works so well that you would be silly not to give the closed loop the same kind of initial values to work with.

I mean, it works so well to only use open loop (which is basically an initial value table with no correction) that I have NEVER used closed loop.

Dann
nitrodann is offline   Reply With Quote
Old 05-17-2015, 08:34 PM   #31
SADFab Destructive Testing Engineer
iTrader: (5)
 
aidandj's Avatar
 
Join Date: Apr 2014
Location: Beaverton, USA
Posts: 17,992
Total Cats: 1,457
Default

They you don't drive in enough conditions. I can tune open loop in the afternoon and have a 10-30kpa difference the next morning.
aidandj is offline   Reply With Quote
Old 05-17-2015, 09:03 PM   #32
Elite Member
 
Join Date: Apr 2010
Location: Newcastle, Australia
Posts: 2,638
Total Cats: 25
Default

Power output is always very very similar. Pressure isnt the only factor in air density, as you have pointed out.
nitrodann is offline   Reply With Quote
Old 05-17-2015, 10:26 PM   #33
SADFab Destructive Testing Engineer
iTrader: (5)
 
aidandj's Avatar
 
Join Date: Apr 2014
Location: Beaverton, USA
Posts: 17,992
Total Cats: 1,457
Default

Small quote about control interval:

Quote:
OK, if it's working that's fine, but with your current control interval, PID is making changes and expecting responses more often than it's capable of actually causing a change to the valve position, which will also cause the gains to be over-sensitive.
aidandj is offline   Reply With Quote
Old 05-17-2015, 11:19 PM   #34
Sadfab Union President
Thread Starter
iTrader: (3)
 
Join Date: May 2014
Location: Kansas
Posts: 2,690
Total Cats: 113
Default

Yeah, I think I need to dumb my valve down a lot...

So I've found what I imagine is best spool, now I just need to reign in the valve range, looks like 0-60%DC and fine tune the PID a bit more.

I was spooling ~7psi at 3800-4000, so this is much improved. Never overboost, never overshot...



This here is bare wastegate, my go-fast switch just interrupts the valve signal. I can't find a fancy boost table switch or anything in MS2. In other words, ignore the boost DC in this. ~6psi at 3600 or so.



Edit: 19.5Hz = 51.2ms, so I'm thinking a 50ms interval is best at a minimum.

Last edited by deezums; 05-18-2015 at 04:54 AM.
deezums is offline   Reply With Quote
Old 05-18-2015, 08:03 AM   #35
Boost Czar
iTrader: (61)
 
Braineack's Avatar
 
Join Date: May 2005
Location: Chantilly, VA
Posts: 72,838
Total Cats: 1,787
Default

Next person who says to run open loop in an closed loop tuning thread gets banned.
Braineack is offline   Reply With Quote
Old 05-18-2015, 08:11 AM   #36
Elite Member
 
Join Date: Apr 2010
Location: Newcastle, Australia
Posts: 2,638
Total Cats: 25
Default

No one said to do that. Cry harder.
nitrodann is offline   Reply With Quote
Old 05-18-2015, 09:35 AM   #37
Boost Czar
iTrader: (61)
 
Braineack's Avatar
 
Join Date: May 2005
Location: Chantilly, VA
Posts: 72,838
Total Cats: 1,787
Default

I realize that, but others have been recently.

it's just a warning; im tired of seeing ******** who don't know how to tune give bad advice (like tuning for 12.8:1 in boost).
Braineack is offline   Reply With Quote
Old 05-18-2015, 02:54 PM   #38
Junior Member
 
Join Date: Jan 2014
Location: Port Saint Lucie,FL
Posts: 418
Total Cats: 3
Default

Quote:
Originally Posted by deezums View Post
Yeah, I think I need to dumb my valve down a lot...

So I've found what I imagine is best spool, now I just need to reign in the valve range, looks like 0-60%DC and fine tune the PID a bit more.

I was spooling ~7psi at 3800-4000, so this is much improved. Never overboost, never overshot...



This here is bare wastegate, my go-fast switch just interrupts the valve signal. I can't find a fancy boost table switch or anything in MS2. In other words, ignore the boost DC in this. ~6psi at 3600 or so.



Edit: 19.5Hz = 51.2ms, so I'm thinking a 50ms interval is best at a minimum.
51ms control interval help stabilize the ebc quite a bit for me, but I still needed a crazy amount of "I" when adjusting the PID to get rid of the rapid oscillations... I think my "I" is now at 175.

Last edited by Mech5700; 05-18-2015 at 03:46 PM.
Mech5700 is offline   Reply With Quote
Old 05-18-2015, 03:14 PM   #39
SADFab Destructive Testing Engineer
iTrader: (5)
 
aidandj's Avatar
 
Join Date: Apr 2014
Location: Beaverton, USA
Posts: 17,992
Total Cats: 1,457
Default

That boost duty is fucked. It should be relatively constant. Like this:

aidandj is offline   Reply With Quote
Old 05-18-2015, 03:43 PM   #40
Sadfab Union President
Thread Starter
iTrader: (3)
 
Join Date: May 2014
Location: Kansas
Posts: 2,690
Total Cats: 113
Default

Quote:
Originally Posted by Mech5700 View Post
51ms control interval help stabilize the ebc quite a bit for me, but I still needed a crazy amount of "I" when adjusting the PID to get rid of the crazy oscillations... I think my "I" is 175.
I think this run was at I125 or I130 or thereabouts, I was starting to subtract a bit form each run to get the overspool I want before I went home for the night. Do you have an initial overshoot, or at least one tuned out?

Need to go back out again, more tuning

Do you remember where you were getting this T3 to spool Scott? I'm pretty impressed for a <$200 5K+mi turbo.
deezums is offline   Reply With Quote
 
 
Reply

Related Topics
Thread Thread Starter Forum Replies Last Post
Project Gemini - Turbo Civic on the Cheap Full_Tilt_Boogie Build Threads 57 07-19-2017 05:11 PM
Back to Stock Part Out!! Turbo Parts, MS2 Enhanced 01-05, Suspension, and MOAR! StratoBlue1109 Miata parts for sale/trade 16 10-02-2015 10:39 AM
Going back to stock. Need some 1.6 parts. Trent WTB 2 10-01-2015 01:15 PM
Bad head gasket or ? shooterschmidty Engine Performance 8 09-30-2015 11:28 PM
Missing/knock at idle, can someone point me in the right direction? Mikel MEGAsquirt 4 09-28-2015 05:46 PM


Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off



All times are GMT -4. The time now is 12:26 PM.