More advanced closed-loop boost entry

Testing and development of Megasquirt 3

Moderators: jsmcortina, muythaibxr

Post Reply
cj.surr
Helpful MS/Extra'er
Posts: 68
Joined: Fri Jul 29, 2016 4:28 am

More advanced closed-loop boost entry

Post by cj.surr »

Fast spooling turbos with limited wastegating (such as an IWG EFR) are extremely inconsistent when entering closed-loop boost control. Having one lower limit delta value is not enough, as different RPM and boost targets make huge differences in the rate of boost increase. When the turbo is spooling quickly (high rpm), the delta will be far too tight and a spike will occur no matter what. When the turbo is spooling more slowly, the delta will engage closed loop too soon, and boost duty will be increased while it’s still spooling, which again causes a spike. High-sensitivity closed-loop boost control is really needed with these turbos because of large variances in required boost duty for a target boost.

Two options I believe could help control boost entry much more consistently:

1) Table of lower limit delta values based on rpm and boost target
2) Function to calculate estimated time to reach target boost based on MAP-dot and difference between MAP and target boost. Disengage lower limit 100% duty when estimated time to boost target is less than a specified length of time.

Number 2 would not require any additional tables but possibly another variable:

-A lower MAP value to reach before running the function (to avoid high MAP-dot values near vacuum-boost transition)



I have already made a thread about PID delay before entering closed loop boost. I think the combination of these two improvements would really help myself and others with very fast spooling turbos.

At this point, I think the MS3pro is really far behind in boost control when compared to another big ECU in the price range that has a handful tables of modifiers for boost duty, as well as more control over boost control entry. For myself, boost control is so inconsistent that I am faced with programming my own boost controller or switching ECUs to keep the car driveable without having to change settings frequently.
franksidebike
Experienced MS/Extra'er
Posts: 226
Joined: Fri Apr 08, 2005 2:08 pm
Location: Wathlingen, Germany

Re: More advanced closed-loop boost entry

Post by franksidebike »

+1
I would prefer:

"2)Function to calculate estimated time to reach target boost based on MAP-dot and difference between MAP and target boost. Disengage lower limit 100% duty when estimated time to boost target is less than a specified length of time."

Frank
aidandj
Experienced MS/Extra'er
Posts: 255
Joined: Fri Sep 26, 2014 1:38 am

Re: More advanced closed-loop boost entry

Post by aidandj »

+1 for all of the above.
93supercoupe
Master MS/Extra'er
Posts: 581
Joined: Sat May 10, 2014 4:57 am
Location: New Haven, Ct, USA

Re: More advanced closed-loop boost entry

Post by 93supercoupe »

Id also like to point out a scenario that i face with Launch boost target that this idea could help.

In my car, i run a 6 psi target launch boost. The problem with that is my delta is 40kpa for closed loop entry, so the boost controller basically opens at 100kpa.
BootlegTuned
cj.surr
Helpful MS/Extra'er
Posts: 68
Joined: Fri Jul 29, 2016 4:28 am

Re: More advanced closed-loop boost entry

Post by cj.surr »

Any plans to look at more effective boost control entry? MS3PRO is really far behind the game on this one, and some simple additions could really help.
KaPower
Master MS/Extra'er
Posts: 501
Joined: Wed Dec 22, 2004 9:43 pm
Location: Bonney Lake, Wa

Re: More advanced closed-loop boost entry

Post by KaPower »

+1... Definitely need better closed loop boost control. So many different factors seem to affect pw% vs actual boost. The single value map threshold before entering closed loop isn't enough control to optimize spool at lower rpm, and not over boost at higher rpm on my super responsive GTX3071 twin scroll setup. We really need an rpm based threshold value to optimize spool-up.
cj.surr
Helpful MS/Extra'er
Posts: 68
Joined: Fri Jul 29, 2016 4:28 am

Re: More advanced closed-loop boost entry

Post by cj.surr »

Bump
DisgruntledMotorist
MS/Extra Newbie
Posts: 17
Joined: Sat Sep 02, 2017 11:18 pm

Re: More advanced closed-loop boost entry

Post by DisgruntledMotorist »

KaPower wrote:..So many different factors seem to affect pw% vs actual boost. The single value map threshold before entering closed loop isn't enough control to optimize spool at lower rpm, and not over boost at higher rpm
That would be quite difficult to achieve. You'd have to have a well defined model of enthalpy at the turbine which I don't believe is really achievable as you'd need a load of pressure and temperature inputs which the ECU is not capable of supporting. Even if it is possible with the current hardware it would take a lot of dyno time to calibrate accurately.
cj.surr wrote:Fast spooling turbos with limited wastegating (such as an IWG EFR) are extremely inconsistent when entering closed-loop boost control. Having one lower limit delta value is not enough, as different RPM and boost targets make huge differences in the rate of boost increase
As for the delta P value, I believe a single delta P value is not the optimum way to enter closed loop. As has already been pointed out, the rate of vehicle acceleration has an affect on the spool characteristic of the turbo. With a high rate of acceleration (i.e in lower gear / down hill) a low delta P is optimum, with a lower rate of acceleration (i.e high gear / uphill) a high delta P is optimum.

The way I would like to calibrate the closed loop entry delta P is with 2d table. Maybe 2 or 3 DeltaP values and have MAPdot or RPMdot on the X axis.
TurboBob
Master MS/Extra'er
Posts: 503
Joined: Sat Dec 02, 2017 10:24 am

Re: More advanced closed-loop boost entry

Post by TurboBob »

to get even more picky: I believe the asynchronous nature of the wastegate PWM relative to the switchpoint has an effect.

if the dutycycle drops from 100% to 50%, and the PWM output has just started its ON phase, the solenoid will be on for an additional cycle as compared to those times
when the opposite is true. Just adds an element of inconsistency to a system we struggle to get consistent....

I'm pretty sure we can't change this, but it is likely a factor limiting how well optimized a control system like this can be.

Bob
Nightstalker1993
Helpful MS/Extra'er
Posts: 75
Joined: Tue Jun 03, 2014 11:09 am

Re: More advanced closed-loop boost entry

Post by Nightstalker1993 »

Highly agree on this. i have trouble with the PID increasing boost duty once i pass the Delta P when entering boost slowly from a low rpm(low mapdot) then overboosting as the boost creeps up past its target, then when the boost climbs up quickly due to acceleration from a higher RPM, It just overshoots and overboosts before the PID can bring the boost down.

Looks like the development for MS3x had been slowing down seeing the lackluster response from the team on this forum?
93supercoupe
Master MS/Extra'er
Posts: 581
Joined: Sat May 10, 2014 4:57 am
Location: New Haven, Ct, USA

Re: More advanced closed-loop boost entry

Post by 93supercoupe »

Nightstalker1993 wrote:Highly agree on this. i have trouble with the PID increasing boost duty once i pass the Delta P when entering boost slowly from a low rpm(low mapdot) then overboosting as the boost creeps up past its target, then when the boost climbs up quickly due to acceleration from a higher RPM, It just overshoots and overboosts before the PID can bring the boost down.

Looks like the development for MS3x had been slowing down seeing the lackluster response from the team on this forum?
Ive noticed that also. Maybe they are working on ms4 hopefully?
BootlegTuned
David_G
Helpful MS/Extra'er
Posts: 98
Joined: Wed Apr 02, 2014 1:14 am

Re: More advanced closed-loop boost entry

Post by David_G »

+1

plus a 2D table to incorporate higher altitudes in the boost initial duties table.
Higher altitudes need much higher boost duties and therefore would require much more aggressive PID values which cause overshooting and are also far too aggressive in lower altitudes.
So a multiplication factor of the initial values table over barometric pressure as a 2D table oder curve would be extremely beneficial!

Too low initial boost duties in high altitudes cause CL boost control need to work really hard to not let MAP fall under the "lower limit boost delta" which then sets boost duty to a 100% which again causes really strong boost fluctuations.
Setting the "lower limit boost delta" to a higher value is no option since that unnecessarily sacrifices boost control in lower altitudes.
DisgruntledMotorist
MS/Extra Newbie
Posts: 17
Joined: Sat Sep 02, 2017 11:18 pm

Re: More advanced closed-loop boost entry

Post by DisgruntledMotorist »

I think an easier way to calibrare around the altitude issue would be to somehow include the pressure ratio into the boost control algorithm.
David_G
Helpful MS/Extra'er
Posts: 98
Joined: Wed Apr 02, 2014 1:14 am

Re: More advanced closed-loop boost entry

Post by David_G »

Easy but it won't work very well I'm affraid.
Each engine, turbo and wastegate setup responds differently to increased boost duties. This makes me think this needs to be tuned individually for each engine/setup.
However EVERYTHING would be better than NO correction for altitude at all. The current code really is lacking that opportunity!
Post Reply