Proper boost control config

[oskar]

Hi,

I am kinda lost in benchtesting boost control.

Got it to work in open loop mode. Output at IAC1. Scope shows nice waveform and PWM signal is changing duty as supposed.
Also looking at boost duty in MegaTune works, duty changes according to rpm

However, going to closed loop, the boost duty gauge in MT just jumps back and forth between 5% and 95% ( which is my value for closed duty and open duty in boost control settings ). No signal at IAC1 with scope ..

Any hints what I might be missing ?

Thx
Oskar

[muythaibxr]

Your PID settings are most likely too aggressive, or your control interval is too short.

Ken

[oskar]

Well, looks like I figured it.

the value for closed duty has to be less that that one for openduty
Then 50 for P and 10 for I works ...

Oskar

BTW: I have modified the code a bit to allow closed duty > openduty. Gives a reverse output. Do you want me to post that in MS2 dev, or PM

[muythaibxr]

There should never be a closed duty > than open duty.

Instead, there is a setting to invert the output.

Ken

[oskar]

Ken,

fine. Then its a bug in the docu to be fixed

in
file:///C:/Program%20Files/MegaSquirt/MS2_Manuals/MS2-Extra_Hardware.htm#boost

closed shows 50, open 15 %

HTH
Oskar

[muythaibxr]

Can you post a link to it somewhere other than your hard drive?

I can probably find it myself as well.

Ken

[oskar]

sure,

http://www.msextra.com/ms2extra/MS2-Ext ... .htm#boost

the link on my hd points to where the MSIIextra release setup installs its documentation to

HTH
Oskar

[muythaibxr]

Thanks, I'll have to see why that doc was done that way and get it fixed.

[95_2.4T]

So is the "boost target" working well now? Last I heard alot of ppl were calling it a lost cause, I have really been looking forward to it, if it is

[oskar]

a little bit too early to call it all working. Would do that after some test driving.

However, I have it working on the bench, controlling a Mercedes bypass valve. The valve requires a PWM signal of 367/8 Hz and a PWM duty between 10 and 90 %.
Again, works very nice on the bench. Will take me probably another couple of weeks till the car is runnig and tested .
( Have to finish intercooler plumbing and then get the electronis in )

I will let ya know

Oskar

[transaxel]

a little bit too early to call it all working. Would do that after some test driving.

However, I have it working on the bench, controlling a Mercedes bypass valve. The valve requires a PWM signal of 367/8 Hz and a PWM duty between 10 and 90 %.
Again, works very nice on the bench. Will take me probably another couple of weeks till the car is runnig and tested .
( Have to finish intercooler plumbing and then get the electronis in )

I will let ya know

Oskar

Hi!

I´ve tested a long time the closed Loop Boost control. But it is very difficult to find the right terms in PID. My greatest problem is that pressure always overshoots before PID kicks in. At high rpm´s boost is build up very fast and a second of waiting by PID is much too much pressure.

Greetz Gerald

[muythaibxr]

you need a bigger D term, and/or a faster control loop.. the D term will start opening the waste gate based on how quickly boost is going up.

I wouldn't call closed-loop boost a "lost-cause," but more like "PID seems to be difficult for people to tune."

It's not difficult if you play on the bench a bit and see what the various changes do.

Ken

[transaxel]

you need a bigger D term, and/or a faster control loop.. the D term will start opening the waste gate based on how quickly boost is going up.

I wouldn't call closed-loop boost a "lost-cause," but more like "PID seems to be difficult for people to tune."

It's not difficult if you play on the bench a bit and see what the various changes do.

Ken

Control Loop is the control intervall, right? This one is set to 10 ms standard and i´ve not changed. But how to find out the soleonid frequency - Is there a way to check out - and how critical is a wrong setting? I let this at 78 Hz because it works in open loop.
About PID: If PID swaps from high to low (like a sinus) P seems to be to much, and if it reacts generally to slow to changes D has to be increased. Is this right?
But what is I for?

Best regards from Austria
Gerald

[GintsK]

Long time ago in MS-1 was good idea coming from subaru boost control strategy using two tables - one table for reference (duty cycle) position and another for closed loop (desired boost). Something similar have some BOSCH systems too (i knew about diesel only). At boost buildup some time after throttle moved only open loop working, and later PID take a control.

I think control boost with a single PID is very difficult because of unpredictable response after throttle moved and asymmetric nature.
May be it is possible for WOT discipline such as drag race, but for disciplines such as rallying or autocross, where throttle used for control over trajectory - it is not for single PID.

Gints

[oskar]

Control Loop is the control intervall, right? This one is set to 10 ms standard and i´ve not changed. But how to find out the soleonid frequency - Is there a way to check out - and how critical is a wrong setting? I let this at 78 Hz because it works in open loop.
About PID: If PID swaps from high to low (like a sinus) P seems to be to much, and if it reacts generally to slow to changes D has to be increased. Is this right?
But what is I for?

Best regards from Austria
Gerald

Gerald,

you are Austrian, so reading German is not a problem. Check this out here, a German website explaining Control loops http://www.roboternetz.de/wissen/index. ... PID-Regler

HTH
Oskar

[transaxel]

Control Loop is the control intervall, right? This one is set to 10 ms standard and i´ve not changed. But how to find out the soleonid frequency - Is there a way to check out - and how critical is a wrong setting? I let this at 78 Hz because it works in open loop.
About PID: If PID swaps from high to low (like a sinus) P seems to be to much, and if it reacts generally to slow to changes D has to be increased. Is this right?
But what is I for?

Best regards from Austria
Gerald

Gerald,

you are Austrian, so reading German is not a problem. Check this out here, a German website explaining Control loops http://www.roboternetz.de/wissen/index. ... PID-Regler

HTH
Oskar

The Problem is to get a feeling for this different Control Loops and what Changes to the terms make.
But now i think that i´ve checked the meaning of the parameters and will try it another time. As i said - open loop works fine but i have different pressures. Cold weather more boost, warm weather low boost about 20-30 kpa. So closed loop would be very nice.

Gerald

[muythaibxr]

you need a bigger D term, and/or a faster control loop.. the D term will start opening the waste gate based on how quickly boost is going up.

I wouldn't call closed-loop boost a "lost-cause," but more like "PID seems to be difficult for people to tune."

It's not difficult if you play on the bench a bit and see what the various changes do.

Ken

Control Loop is the control intervall, right? This one is set to 10 ms standard and i´ve not changed. But how to find out the soleonid frequency - Is there a way to check out - and how critical is a wrong setting? I let this at 78 Hz because it works in open loop.
About PID: If PID swaps from high to low (like a sinus) P seems to be to much, and if it reacts generally to slow to changes D has to be increased. Is this right?
But what is I for?

Best regards from Austria
Gerald

No, this isn't really how PID works. P is "proportional" gain. Generally you want to adjust this as high as you can without getting into an oscillation condition.
I should be adjusted the same way (but may not be necessary for boost). I is basically to correct for accumulated error, so the only place I could see it being useful in PID is to close the wastegate a bit as RPMs climb to compensate for boost dropping.

D is used to slow down the correction as it gets closer to the target based on how quickly you're approaching.

For the frequency setting you should use the lowest one you can that still works, as this will give more resolution on controlling the valve.

Ken

[muythaibxr]

I will consider a 2-table strategy for closed loop boost control for a later release.

I can see this being easier to set up and get working.

Ken

[intense98rt]

I've started to play around with boost control today. I can't seem to get it working properly.
I've tried closed loop with the settings in the manual. For target I set everything to 150kpa. The valve will just open/close and it goes to wg spring pressure(115-120kpa).
Since I've heard that more people had open loop working better, I decided to try that. If I set duty to 0, the valve stays open. If I set duty to 100 the valve stays closed. If I try any duty in between the valve will just open/close and go to wg pressure.
What kind of settings should I try to get this working in either open or closed loop?
BTW, I'm using the gm solenoid with frequency set to 19.5Hz.

[tristanlee85]

So should closed loop be 85% and open loop be 15% (how it is in the documentation) or should closed loop be 15% and open loop be 85%?