Learn me MS idle tuning (and other puzzlements)

As is documented elsewhere on here, I'm in the process of converting a '85 MB 190E to running a fuel-only Microsquirt setup.  After about a month of flailing on getting a working RPM input signal (hint - despite the instructions, do NOT use the coil terminals as input to the OPTOIN pins on the Microsquirt), I seem to have finally gotten over that hump, and now have a reliable tach signal, and actually can reliably start and run the car. BTW, to those of you who offered feedback on the tach signal issue, mucho thanks!  And to those reading now, be aware, this is my first toe dipped into the engine tuning waters, so if it's not already clear, I have no idea what I'm doing.  Yet. 

Now, onto my current state and associated questions.  As it sits, it starts without issue.  I have to give it some gas to keep it running as it warms up and it tends to surge/die over about 1000-2000 rpm from a totally cold start (which right now is ~50*F).  However, once warmed up to 80*C, it idles no problem, and it will start and settle immediately into idle without touching the gas at that temp.  Great.  I have not futzed with ASE and WUE at all to dial that in, so I'm not worried about that. 

What puzzles me, however, is that it's idling around 1250rpm.  The Haynes manual says idle speed should be 800 +/-50 rpm.   When I first got it going, it was idling super-rich (WBO2 said ~10.3 AFR), so I progressively reduced the VE table values to where I am now.  As I did that and it leaned out at idle, the idle speed also increased.  

Today I got it up to temp, shut it off, unplugged the ICV, and then restarted it and datalogged.  See below (sorry about the yellow...).  As usual, it started right up, and it idled no problem without the ICV.   But I'm confused by the MAP reading.  It shows 102kPa before I start the engine, so that seems right, but at idle it's at about 63kPa.  Trying to read up on this, it seems like it should be more like ~30-35 at idle.  

So is there something I'm missing, either in the tune, or the engine itself, that's causing this?  

I'm also curious about the swing in the AFR when the TPS position changes.  Having never dealt with a WBO2 before, is it normal for them to swing wildly on quick throttle changes?  BTW, I don't turn the WBO2 on until just after 20s, and then it goes through its init procedure (the 13.3 and 16.7 readings).  Then I turn it off just before the end of the run. 

The first thing I'd check is the IAC valve. If it's a Bosch 2 wire unit, many of these are partially open if they lose power, as a limp-home strategy. Try increasing the duty cycle to the IAC valve to around 20% duty and see if the RPM drops.

High rpm and low engine vac might be a large leak like the IAC thing mentioned. If it isn't i'd confirm your ignition timing under these conditions. Ignition timing at idle can cause huge swings in manifold vacuum and rpm. 

For the record, once up to operating temp, there was no noticeable difference in idle behavior with the IAC valve plugged in or unplugged.  A previous datalog was showing a 2% duty cycle while the engine happily idled away (at 1250rpm).  It occurs to me now that maybe I have the pins reversed and what I think is closed is actually open...  

I haven't touched timing, still using the original MB stuff, and it idled at more like 900 before I started this, so that's not it, or at least that's not a problem that's came with the MS conversion.

I don't remember a screen / menu to increase duty cycle, I'll have to poke around tomorrow.

I am not familiar with the MB setup but I can tell you how I got my LT-1 to work.

1. used the test feature to find out the numbers that your IAC is fully open and fully closed, this can vary but you need to know where those two numbers are for your setup. Put an IAC gauge on your dashboard so you can see its position.

2. with the engine fully warmed up set the IAC to fully closed and then adjust the throttle blade on the TB to give you an idle about 50 to 100 rpm lower than target. I had to remove a cover to find the adjustment screw for the TB.

3. Now you can use the IAC to adjust the cold idle curve so that you can maintain the rpm you want compared to the engine temp. The thing I had to remember is the IAC can only add air so if the throttle blade is too far open (or there is a vacuum leak) then the IAC can do nothing to bring the idle down. 

I had to do a combination of IAC tuning and timing adjustments to get my engine to idle the way I wanted under all conditions, on and out of gear (auto) and with or without a/c on.

Your question about the AFR changes with throttle have to do with the acceleration enrichment settings and yes you can get wild swings very quickly. I fouled out several sets of spark plugs before I got mine in line, but it was way off to start and it took me some time to figure out that problem

The thing I had to remember is the IAC can only add air so if the throttle blade is too far open (or there is a vacuum leak) then the IAC can do nothing to bring the idle down. 

Before i had anything resembling the diagnostic ease of simply putting an idle motor position gauge on my standalone 'dashboard' I got tired enough of dealing with sticking/slow IACs that i set the throttle high enough that the IAC could never stall the engine regardless of conditions. Along with this i accepted having to do maybe one seasonal adjustment per year if it bothered me. Over time I've decided that the best idle speed is the one that bothers me the least, and that might come down to the exhaust setup or the engine mounts or some random doodad resonating on the body as much as it comes down to a number.  

Thus ends my unsolicited and objectively useless opinion. 

Fancy systems will use timing control to do a lot of their idle control because it reacts much more quickly and consistently than an idle motor.  

Use test mode to verify the function of the idle valve. As Matt stated many two wire Bosch valves are pretty open unplugged or at low duty cycle. If your idle duty was 2% that is not enough to close the valve enough to see a difference. Generally they are "fully closed" around 30% give or take and depending on frequency (most are good at 100-120Hz). 

Generally on a warm start I will set idle valve to be slightly open (maybe 35%) and use an idle screw to set base idle to be where I want it. That way the idle valve can easily add air without being able to stall the engine if it crosses the closed state. 

Once you have this configured you can start tuning fuel and ignition. 65kpa seems high for idle map but taking out fuel and adding timing back in will help it. Depends a lot on the camshaft and induction though. If it's still high, and idle is high, look for vacuum leaks. 

Can you guys explain what you mean by testing the valve for fully open and fully closed?  I mean, I understand the concept, I'm just not sure how I test this without a flow meter of some sort to find where fully open and fully closed are.  

I did try switching the valve mode to "inverted" and that shot the idle up to 3000 rpm, so that's clearly not the problem. 

The idle screw is completely backed off, so nothing is holding the TB open.  I pulled the intake elbow off the throttle body and noticed the throttle plate (off a Mercury Sable) has two tiny pinhead-size holes in it, labeled A and C as I recall.  I don't know what those were supposed to do in the Mercury setup.  As a test, I put a piece of tape over them, restarted, and the idle dropped to ~1050 rpm, so that's one source of air.  Tomorrow I'm going to plop a lacrosse ball on top of the throttle body, which is about the right size to completely seal off the intake air tract, at least in theory.  If it keeps running then I know I have a vacuum leak somewhere. 

I'm pretty sure your idle valve is hooked up to some rubber hoses, so all you have to do to figure out if it is passing air is squeeze one of those hoses flat with a pair of pliers and see if the idle changes. You have the easiest possible diagnostic scenario in that respect since newer idle motors usually have both inlets and outlets cast internally to whatever they mount to and are slightly harder to definitively prove they aren't passing air.

Every throttle body is a calibrated air leak when 'closed'. None of them actually 'seal up' from the factory. The majority of idle air still goes through/around the plate.  Some will do this by holding the plate cracked open, some will let it close but drill holes in it etc. Only a percentage of idle air is affected by the idle motor, otherwise the idle motor would cause huge swings in idle speed and be difficult to control. At least back then, when idle motors sucked. Now everything is drive by wire, idle motors are extinct, and the throttle body motor just opens the throttle plate a little more when necessary. I think the airflow of a butterfly valve follows a bell curve so it's not super touchy at the near-closed positions.

Vigo said:

I'm pretty sure your idle valve is hooked up to some rubber hoses, so all you have to do to figure out if it is passing air is squeeze one of those hoses flat with a pair of pliers and see if the idle changes.

Sometimes I get so far down in the weeds I forget to take a step back.  Duh.  

I tried two separate tests today. 

Anyway, first test,  I disconnected the intake elbow so the throttle body was directly accessible, and I taped off the holes in it so it couldn't get air through them.  The only air connection before the throttle body that I took out of the loop by doing this is the PCV/breather inlet, which goes from the head to the intake elbow.  I got the engine warm and while it was happily idling away, I covered the TB with a hockey puck.  The engine died immediately, no sputtering, gasping for air, or trying to idle at 300 rpm.  Just immediately died.  That suggests to me the intake tract is pretty well sealed, no significant vacuum leaks.  

Second test, I put the intake tract back on (not sure why, but I did), and pinched off the idle valve hose using a fuel line clamp.  As the line got more closed - the clamp gradually restricts the line as you twist the knob - I could first hear the air getting sucked through the air filter, and then when it was fully closed, the engine sputtered a bit and then died.  Which means the idle valve is in fact providing air to the engine at idle.  

At this point, I'm inclined to say that for whatever reason, this engine naturally wants to idle a bit high.  Does that sound feasible?  I briefly messed around with CL idle and it kinda-sorta got it down to more like 800 rpm, but it wasn't particularly stable.  This is with no tuning of the CL idle, just literally turning CL idle on and setting the CL idle target to more like 1000.  Seems like with some tuning I should be able to get a decent idle speed with CL (and hopefully match the idle target RPM).

Thoughts?  Really appreciate the thoughts you guys are passing along...

There is a test mode in MS2/microsquirt, right most menu, find the Test section near the bottom, it will allow you to change the valve frequency and duty cycle manually and see how the engine reacts. You can also do this in the idle menu when warm by just changing the rightmost PWM duty bin up and down and see how it reacts. Leave idle open loop until you figure out how the valve reacts to different duty cycles. Enabling closed loop will just confuse things without this. 

If it "naturally" idles high you have a vacuum leak of some kind, intentional or not. 

Manuals are at msextra.com go into the ms2 section and find what you need in the tunerstudio reference. OR drive the FB groups nuts, it's good for quick questions.

NorseDave said:

Can you guys explain what you mean by testing the valve for fully open and fully closed?  I mean, I understand the concept, I'm just not sure how I test this without a flow meter of some sort to find where fully open and fully closed are.

You have a flow tester - it's the engine!

Disable closed loop idle, and start stepping through different duty cycles.  Increase by perhaps 5% every step, and see what the resulting idle speed is.  Record duty cycle vs. idle speed, and see what the shape of the curve is.

If your Bosch idle valve is like the one that was on my Volvo, then it may have a failsafe position at 0% duty.  0% duty is actually maybe 75% open, but 30% duty was fully closed (or open, my motor may have been inverted as well...).

If you can post the results of that test (duty vs. resulting idle rpm), we can do a better job helping with how to configure it properly.

Wanted to circle back around to this now that I've had a little time to do some proper investigation, reading, testing, etc.  

First, I did discover one vacuum leak, apparently I forgot to fully tighten the hose that goes from the manifold to the brake booster (has a flared fitting on it), so when I was systematically undoing and capping each non-essential vacuum connection, that one was suspiciously easy to undo.  With that capped, and then with it properly connected, warm idle settled down to the ~800rpm range that MB says its supposed to be.  

I was still getting confusing results when I messed with the idle valve, though.  Then I remembered that I actually had a true flow meter - a combined MAF/IAT.  I'm not using the MAF currently, so I had disabled that input as well.  I enabled the input so I could use the MAF readings as a true measure of airflow. 

Once I confirmed I was getting MAF data, I got the engine warm, used a large fuel line clamp to gradually clamp off the line to the idle valve, and confirmed that both the RPMs and the MAF reading went down as I restricted it. 

Then I did some testing with the idle valve test mode, and I do indeed have the idle valve polarity reversed.  As far as I can see there is no polarity marking on it, so I just guessed when wiring.  Now with the mode control setting set to "Inverted" I get the response I'd been expecting - as I increase the idle valve PWM duty cycle, both the air flow and the RPMs increase.

So now that I've got it behaving like I expect, I can determine the idle valve DC limits like Fladriver64 mentioned, and decide where I want to set the idle screw, and then recal the TPS.  

Gracias for the suggestions from everyone...

By the way the actual polarity of the wires going in to the IAC probably doesn't matter - it's a magnetic device, so it only really cares about how much current is flowing through the coil, not the direction of the current.  More current -> closes valve.

In reply to NorseDave :

Good to hear you are making progress, you dont want to know the number of mechanical problems (like vacumm leaks) that I fixed that I thought were computer problems. 

Matthew Kennedy said:

By the way the actual polarity of the wires going in to the IAC probably doesn't matter - it's a magnetic device, so it only really cares about how much current is flowing through the coil, not the direction of the current.  More current -> closes valve.

Yeah, that's what I thought, which is also why I just went ahead and wired it up without paying much attention to where the 12V went.  But I can say with absolute certainty now that if I configure the valve in "Normal" mode in Tuner Studio, increasing the PWM DC results in a decrease in air flow.  And if I configure it in "Inverted" mode, then increasing the PWM DC results in an increase in air flow.  So maybe it's not the polarity of the wiring, it's just this is one of those "weird" valves that operates the opposite of most.  

Actually, I'm 99% certain this valve came out of a late-80s Volvo, so reading your previous post about "0% duty is 75% open" it all starts to make a lot more sense. 

As usual, figuring all this out on your own tends to take longer and results in more interim confusion, but you end up understanding stuff a lot better in the end!