Boost control change testing (MS2 and MS3)

[nathaninwa]

I just want to chime in and say for me its working great. After tuning the Bias table to flatline boost, I turnrf pid on and seems in the weather conditions im seeing and even wheel slip boost stays within 5kpa or so. Ive even turned on traction control and in the logs can see boost duty drop a percentage and boost follow nicely to stop the wheel slip.

[duxthe1]

A big part of my problem is the discrepancy between the initial duty numbers I need to get to my target v/s the duty needed to hold the target. In my screenshot an initial duty of 27 got me to 211.4 KPA with a 210 KPA target. The problem is that it needs a duty around 55 to 65 to hold the target. If I put numbers like that in my initial duty table I get huge overboost spikes (20 30 KPA). I had PID real low in that datalog but higher numbers dont help. I end up getting huge swings in duty but marginal if any boost control. Despite my hot side being very restrictive, the cold side is the weak point in my setup and likely wont allow me to tune closed loop. Im going to try open loop and see how it does.

[elaw]

Wow... that's weird! Does it do the same thing with "old style" open-loop control that doesn't hold the boost duty at max until you're near the target?

Is the boost chamber on your wastegate very large? I'm wondering if all or part of the problem might be the restriction in your boost control valve combined with a large volume inside the wastegate causing pressure inside the WG to build slowly. That could delay the valve opening and causing the spikes.

[zmanco]

Is the boost chamber on your wastegate very large? I'm wondering if all or part of the problem might be the restriction in your boost control valve combined with a large volume inside the wastegate causing pressure inside the WG to build slowly. That could delay the valve opening and causing the spikes.

Another factor could be if you have a restrictor in the boost signal hose. Too small could be part of it, although in my case changing to larger didn't solve the fundamental issue, just required new initial duty values.

[nathaninwa]

One thin i did to help overshoot, was move from a delta of 20kpa to 40kpa. I lost spool in the logs, but my back to back testing on the street i didnt feel much of anything.

[Jdmfreak]

OK, I was able to see in the code why the second table wasn't working for bias.

Some of the logic in that area of the code could use a longer look than I have time to look at it tonight (interaction between the various things that can set target and which bias table should be associated). I should have time to fix it this weekend, so hopefully we'll have test code out by the end of the weekend.

Do you happen to have the test code yet?

[muythaibxr]

I do. We will probably release it as a new alpha/beta soon, but I can send it to you if you want it sooner (when I get home from work
later)

[benc23mustang]

Hello I'm looking for some help with closed loop boost control. I have a 2.3 turbo ford mustang with a holset hy35, internal wastegate, running beta 3, and DIY bcs. I've had some success but can't quite get it dialed in as well as i would like. My current settings are Delta=30, open duty 80, closed 0, slider at 259, p=8, i=40, d=0, freq=19.5, control interval=51ms.

I've been able to dial 3rd and 4th gear in "ok" but 2nd gear is much more volatile. As you will see in logs it can take a long time for the boost to "settle in". All logs attached are 2nd and 3rd gear pulls

thanks
Ben

[benc23mustang]

And here is my tune

[muythaibxr]

Are you using the test firmware? MS2 or MS3?

[benc23mustang]

Yes the test firmware. Ms2 extra

[pit_celica]

For the 290kPa run, it looks like the open duty at 80% is limiting the boost a little too much. I would try to set the open duty at 90%.

Sam

[benc23mustang]

ok ill give it a try. What would most make the code increase or decrease the boost duty the fastest? raising P? lowering I?

[pit_celica]

For a standard closed loop PID, the P is usually the variable that have the biggest impact on the response of the loop. I would suggest you to increase P a little bit so that you have a little more overshoot so that when the boost drop right after the overshoot, you will still be near the target. Then, the I term will take the relay to stick the boost to the target. If you feel you have too much overshoot, add some D.

Please note that these suggestions are base on the 1.2.x MS3 code, not the new actual way with the bias table. Maybe Ken has better suggestions for you.

Sam

[benc23mustang]

appreciate the help. More P definitely helped with reaction time and so did D. Im up to P=16 I=40 D=8. Increasing D seems to help substantially with faster reaction to boost/boost duty. I will post more logs hopefully tomorrow.

[Chiburbian]

I have a question regarding the initial duty table...

Say you are in 3rd gear and you hit your target at say 3200rpm. Under what circumstances does the code check the initial duty table again? Does it lock out initial duty table once it hits target or does it keep referring back to the table?

I will be spending some money at a dyno tomorrow to work out boost control and work on getting my HP and torque to where I want it to be, but I am not sure how to best tune the initial duty table. I am thinking my best bet will be to start at about 600rpm lower than my target in 3rd gear and floor it until I hit my RPM target and adjust my initial duty until I hit my duty kpa at the RPM I am at. Rinse, repeat.

Am I close?

I am going to re-read the thread tonight in it's entirity so I don't waste any money tomorrow but some more guidence would be appreciated. Thanks!

[elaw]

I think with the old-style code, it would check the initial duty table when you first go into boost, then not again until you go out of boost (load < 100) and back into boost again.

The modified code that's being discussed in this thread does it differently and the table is named differently - it's called a "bias table". In the case of the new code, the value from the bias table is used continuously. Every time the code calculates the boost duty, it looks up the appropriate data in the bias table based on the current RPM and requested boost. It then takes that value, adds to it the output from the PID calculation, and the result is the boost duty sent to the valve.

Think of it this way: with the old code, the initial duty was a "starting point", and from that point forward for any boost "pull", the valve duty would be controlled by PID only. With the new code, the bias table is always in the equation - so its value is used to get close to the boost target, and the PID is used for "fine tuning" to increase accuracy.

[nathaninwa]

On your way to the dyno do your self a favor. Make a 3rd gear pull from 2000 to say 5000rpm and datalog it. Then retest on the dyno making sure the boost is built the same. If you dont all the bias table values will be off and fueling will be off. The dyno can adjust the load to match what you see on the street. I have to do this with the Dynojet I visit.

I have a question regarding the initial duty table...

Say you are in 3rd gear and you hit your target at say 3200rpm. Under what circumstances does the code check the initial duty table again? Does it lock out initial duty table once it hits target or does it keep referring back to the table?

I will be spending some money at a dyno tomorrow to work out boost control and work on getting my HP and torque to where I want it to be, but I am not sure how to best tune the initial duty table. I am thinking my best bet will be to start at about 600rpm lower than my target in 3rd gear and floor it until I hit my RPM target and adjust my initial duty until I hit my duty kpa at the RPM I am at. Rinse, repeat.

Am I close?

I am going to re-read the thread tonight in it's entirity so I don't waste any money tomorrow but some more guidence would be appreciated. Thanks!

[sheek]

im confused

the manual says set the polarity more duty = less boost but the tooltip says the opposite. which is it that i have to use to maximise the use of closed loop boost?

this is ms3x btw

[elaw]

Heh... you're confused with good reason!

In older firmwares, and this is probably what's in the manual, higher duty meant less boost. I think the philosophy was that duty would relate to wastegate flow, and of course more wastegate flow would equal less boost.

At some point, I seem to remember it was around a year ago, the developers decided people were getting confused by that and changed the "sense" of how boost control works. So in more recent firmwares, higher duty should equal more boost, not less.