Tuning Boost Control With Bias Table

aidandj · Post by **aidandj** » Thu Apr 23, 2015 7:42 pm

What did you learn. I'm still having issues getting it to work right. I get some long term oscillations, but more I just hits overboost.

E30_Driv3r · Post by **E30_Driv3r** » Fri Apr 24, 2015 6:13 am

I firstly had a slightly loose hose going to my solenoid that I fixed - So i required less duty to hit my targets. I tried setup mode, played with the duty % and it uses my bias table hitting my target perfectly, so to start I would suggest doing setup mode, enter duty number and see what %'s get you close to your target, like open loop pretty much (I did this last year setting all the gains to 0's before setup mode was created.)

Where I get lost is in basic mode the code is using duty numbers much, much higher than the ones on my table - why I'm not sure. In the picture attached the code kicks in at 43% duty - there isn't a 43% duty number on my bias table?

E30_Driv3r · Post by **E30_Driv3r** » Fri Apr 24, 2015 6:41 am

Ok, so looking at nathaninwa's thread with the same "issues" I think I've found my cause for mine. I'm going to test and report back.

edit: no dice, I made my target table all one value and still hit overboost. going to try adjusting frequency next.

nathaninwa · Post by **nathaninwa** » Fri Apr 24, 2015 10:08 am

How are you doing boost. Just one table or blended? I found with gear based blending the code is goofy along with gear switching. I'm able to run single boost level or the boost by gear on the main boost page with great success.

At the track I was able to nail my targets in setup mode with only about 5kpa oscillation. I then turned basic mode on and with the slider at 1/4 only my targets were met.

aidandj · Post by **aidandj** » Fri Apr 24, 2015 10:56 am

This is as close as I could get it today.

What would you suggest changing?

graph.png

E30_Driv3r · Post by **E30_Driv3r** » Fri Apr 24, 2015 7:31 pm

nathaninwa wrote:How are you doing boost. Just one table or blended? I found with gear based blending the code is goofy along with gear switching. I'm able to run single boost level or the boost by gear on the main boost page with great success.

At the track I was able to nail my targets in setup mode with only about 5kpa oscillation. I then turned basic mode on and with the slider at 1/4 only my targets were met.

I have two target tables, based on ethanol %. At the moment I'm using only one as I have yet to get some e85 in the tank and tune it.

Do you know if the slider is active in setup mode? At the moment it's the only thing that works for me even when I set the slider to zero in basic mode it's automatic overboost. If Ken or someone could explain the differences between the release code boost control and this so I can understand it would be a good start, this new system has potential to be fast reacting when dialled in.

nathaninwa · Post by **nathaninwa** » Fri Apr 24, 2015 10:15 pm

You must have a current msq posted. I'll check it out tomorrow

In a nut shell, release code uses and initial value, then bases the rest of the duty cycle on pid alogorythms and not really having a reset point

The current beta stuff I like to think of it as a ve table of sorts. We tune for an afr (boost duty) open loop style then once were close we turn on O2 correction ( slider and pid) to maintain and correct for day to day changes. Since the code is using the bias table it always has a new reference to base boost duty on so the code doesn't have to work so hard

Setup should be just that. Setup mode ran open loop only

As with my case, turn any boost blending off and use a single target table only to not have the ethanol table a possible problem like I found.

Post by **muythaibxr** » Sat Apr 25, 2015 6:14 am

Nathan did you ever post logs/msq of the blending problem?

nathaninwa · Post by **nathaninwa** » Sat Apr 25, 2015 7:42 am

I did Ken, in the 1.4.10 boost thread we were posting in

Boost by gear on the main boost page is working great and allows me to tune each gear in a single pass

E30_Driv3r · Post by **E30_Driv3r** » Sat Apr 25, 2015 9:02 am

I understand, the bias table is to get close your target so then pid can do its thing. Does it function the exact same as the old named "initial duty table"?

Here is my msq:

nathaninwa · Post by **nathaninwa** » Sat Apr 25, 2015 9:58 am

I look at your msq when I get home. The bias table is load specific and rpm dependent. I'm finding for most installs after peak tq the dc needs increased. And the code will use what you put in the table, so if it's over boosting in setup mode the the dc is to high.

nathaninwa · Post by **nathaninwa** » Sat Apr 25, 2015 11:00 am

If I was testing your car. I would disable the blend table and tune boost table 1 only right now. I would also add some resolution to the bias table by rescaling where boost actually starts, meaning rather than a 1500 rpm bin it might start off with a 2500 or 3000 bin

I know your in setup mode, but move the slider to zero and data log that

Is the turbo small for the application?

E30_Driv3r · Post by **E30_Driv3r** » Sun Apr 26, 2015 6:36 am

Good call on fixing the axis, I now have it from 2500rpm up.

I have a Comp turbo 6465 turbo on a 3.0L inline 6, T3 manifold.

On the older code I had no issues w/boost control, even on the dyno it was great, change the desired boost target and MS would hit it.

Going to try basic mode with slider low, I turned off blending for now so there are no "distractions" haha

nathaninwa · Post by **nathaninwa** » Sun Apr 26, 2015 8:57 am

Your setup just about the same as mine. 3.0L with a precision 6466

Post a log with blending undone. Your bias table seems backwards to me then. I roll into boost say 30dc and come out the top at 35dc to flatline boost. Yours is high and tapers low which might be causing a false overboost

E30_Driv3r · Post by **E30_Driv3r** » Mon Apr 27, 2015 1:12 pm

My reason for tapering is I have a pretty free-flowing exhaust, cams etc and when i keep DC the same or higher uptop boost rises. Say if I punch it at 5-6k i would tend to hit overboost protection with higher DC% since the turbo lights up really quickly.

I have a log with basic mode, slider to 0 and again this makes no sense. see first photo. will also attach log.

why does it go to 49% and totally ignore my bias table? and then ridiculously ramp up DC%?? see record 6177 in log.

I then turned switched back to setup mode with no other changes, and you can see the difference in behavior. Second pic. see record 10240 in log

here is the log:

https://drive.google.com/file/d/0B3GWeG ... sp=sharing

E30_Driv3r · Post by **E30_Driv3r** » Sat May 02, 2015 5:59 pm

anybody?? I dont want to go back to the release code but I just might to regain my boost control!

nathaninwa · Post by **nathaninwa** » Sat May 02, 2015 6:07 pm

Let me charge my laptop up and i take a look

nathaninwa · Post by **nathaninwa** » Sat May 02, 2015 7:32 pm

That last tune you posted you said you made the boost blending changes so its just a single target? And in that last msq what is up with the boost control pin setup to off?

E30_Driv3r · Post by **E30_Driv3r** » Sun May 03, 2015 6:57 am

Yeah, just one target with blending off. That's strange it's off it was surely on previously. I've turned back on boost blending as I filled up on some winter blend e85 from NY so I'm setting that up. I can post my latest msq shortly. Appreciate the help nathaninwa!

edit: here is my lastest msq

E30_Driv3r · Post by **E30_Driv3r** » Sun May 03, 2015 6:31 pm

going back to the last release when my boost control was working to see if something is up with my setup or a firmware difference.

nvm, gonna try changing the frequency for 26 to 39 (using a 3port mac valve)

Megasquirt Support Forum (MSEXTRA)

Tuning Boost Control With Bias Table

Re: Tuning Boost Control With Bias Table

Re: Tuning Boost Control With Bias Table

Re: Tuning Boost Control With Bias Table

Re: Tuning Boost Control With Bias Table

Re: Tuning Boost Control With Bias Table

Re: Tuning Boost Control With Bias Table

Re: Tuning Boost Control With Bias Table

Re: Tuning Boost Control With Bias Table

Re: Tuning Boost Control With Bias Table

Re: Tuning Boost Control With Bias Table

Re: Tuning Boost Control With Bias Table

Re: Tuning Boost Control With Bias Table

Re: Tuning Boost Control With Bias Table

Re: Tuning Boost Control With Bias Table

Re: Tuning Boost Control With Bias Table

Re: Tuning Boost Control With Bias Table

Re: Tuning Boost Control With Bias Table

Re: Tuning Boost Control With Bias Table

Re: Tuning Boost Control With Bias Table

Re: Tuning Boost Control With Bias Table