I've been working on trying to dial in my closed loop idle settings, and I've run into a bit of a wall and feel like I'm chasing my tail. The car idles well with the AC off, and tends to reach the target RPM without much fanfare. However, I'm getting some RPM oscillations both when the compressor comes on and goes off. When the compressor comes on, the idle hunts a bit but tends to settle in within a second or three.... oddly, sometimes its good and others it takes a bit longer. When the compressor goes off is the real trouble, the RPMs crater to the point of almost stalling then bounce around a bit. I know my idle VE table isn't fully dialed yet, but with the way the RPMs move around it's hard for me to make a determination as to what to change. I also thought I understood PID controllers enough to sort this out, but adjusting those parameters hasn't helped much. I attached my current tune along with a log of the car idling with the AC compressor cycling a few times. I'm hoping someone can take a look and give me some guidance as to where to focus my attention. I've only had the car running on MS for about three weeks, so I'm sure there are things that I've done with the tune that aren't 100% correct, and I know that my WUE is a bit wonky. I'll get back to that once I get the VE stuff sorted.

 

I am not very familiar with tuning closed loop idle as I use open loop exclusively, but I do recall that there was a "dashpot adder" in the closed loop settings which helped minimize rpm drop after a load was applied/released. I could be wrong about this though as I'm not even on my laptop to verify it in the software. Another possibly good thing to look at too is to make sure your local VE cells around the main cell being used are smoothed out to even increments and likewise values. Having them not coordinated can lead to rev hunting.

Hope this helps!

 

I do have quite a bit of difference between the AC off cells and the AC on cells... I didn't even consider smoothing them out. The only dashpot adder that I've found so far says that it's for when the throttle is lifted. I'll keep looking, though.

 

That's the dashpot I was referring to but it seems like it has no effect from your description

 

The AC off vs AC on cell difference shouldn't matter as much. You'd just be tuning based on different loads so give it whatever it wants. But maybe make the creepdown from AC on to off smoother to allow for a better transition.

 

That's the goal, just have to find the right setting to tweak.

 

I'm gonna be lazy and ask you what FW version you are running on which MS?

 

Damn, I knew I would forget something and look dumb.

It's an MS3PNP with firmware 1.5.1

 

Try working with idle advance. It goes to a greater resolution of idle cells and gives you better insight of the transition of AC on/off.

 

Thanks for pointing me in that direction. Honestly I thought that was turned off, but apparently it isn't. I haven't modified any of those settings from the base map, and the timing correction curve seems a little weird to me.

I'll make some tweaks to that when I get home from work today and see if it helps.

 

I would reduce the timing correction, but first try it with it off. It is adding timing and the result is an over shoot. You might also try reducing your timing at idle and neighboring cells. I run 14 degrees. The rest of the timing table also needs work. Your running too much timing in the lower Kpa cells for your AFR. If you were leaner you might need more timing down there.

 

That timing correction is way too tame. But you cannot make it much more aggressive, because your idle Advance table is so Advanced.

I have attached my tune. EDIT: I am only wanting the 12-21 one, but I cannot get rid of the 8-21 one.

Items of importance are AC settings, including fan. The other fan (note this enters in for me, but may not for you, as I have some fan idle-up delay and RPM raise. The Engine States. Other idle parameters, such as the idle advance table, the idle advance correction. Hmmm, I just noticed that I can enter Idle Advance based on engine states, or not. I may play with that.

The idea of the advance correction is to have the idle advance low, so that you can add a lot of timing with change in RPM. But that timing has to be able to add torque. If Idle Advance is already at peak torque, you have nothing to add.

 

I appreciate the info, and it makes sense to some extent. I'm having trouble loading the tune, though. TunerStudio says that can't find the controller definition file for the serial signature MS3 Format 0567.07. I saw that your tune is for a beta version of the ECU firmware, so I'm guessing that has something to do with it.

EDIT
I have downloaded one of Brain's maps and there's quite a bit of difference between his spark map and mine at idle KPAs and RPMs.

 

I should have specified that I meant reducing the range from -400 & 300 to -200 & 200 to see if it can react quicker. This might reduce the large increase in duty. Currently the idle duty is going from 25 to 41.6 before there is any increase in the timing. Then the timing goes from 16.8 to 25.7 at 1381rpm. Change the timing table in the area from 700-1500rpm X 25-45kpa to 14 degrees. Then alter the timing table or add in the idle timing correction. I could not make both idle advance and idle timing work well concurrently. The load is less consistent so I recommend working with the idle rpm correction only. Also the initial drop is minimal it would not be bad if the drop increases a bit if you reduce the overshoot substantially.

 

Also the idle is too lean to be hardy. It should be 13.8-14.2.

 

Since pump gas is E10 virtually everywhere now, I change the stoich setting to 14.1 and run 14.1 in my commanded AFR table everywhere I would run stoich values.

Therefore, you will want to run even richer than 14 at idle, more like 13.3.

This is why it's better to tune off Lambda than AFR, except I don't think the MS uses Lambda that I have seen.

 

Y'all don't just run 14.7 when at stoich/idle? Even running that just about everywhere only yields me like 22-23mpg, can't imagine what I would be getting running at 14.1 or richer

 

The balancing of an equation of reactants is a theoretical calculation. It is used as a metric by manufactures of lambda sensors. That is the only connection with an IC engine. 14.7:1 (gasoline) or 14.4:1 (E10) is too lean for maximum thermal efficiency (output), it is too rich to maximize a vehicles overall efficiency (mpg), is too lean for starting and for a hardy idle. Lets not high jack this thread too much.

 

You definitely can use lambda on the MS, I don't have my laptop handy to say where it is, but I know you can. It may need to be configured when a tune is first created though, can't confirm if you can switch between the two after the fact.

 

That would be great, I'll poke around and see.