MS1 Pwm Idle does not work.

[Keithg]

Yes, the upper/lower limits and slow/fast recovery form a line. As you get closer to idle speed, you should slow down (higher recovery number in seconds) the correction interval (more time between rechecks of idle speed). You say it counts down from 100? It should crank over and the idleDC should be 35, then it shoudl decay to an idle speed. The idleDC should be no less than 20unless you are driving then it will close to 15. If it does not do this, let me know. Crank your dashpot settle time to 255. This will give it 2.55 seconds to decay to a decent idle speed.
 
Since your ford valve is closed at 0dc, you could set the closedc to 0 to get better engine braking.

Do you have ego correction turnerd off at idle? Is your idle AFR pretty constant? Can it maintind a decent idle without PWMidleClosed loop?

KeithG

Posted by email.

[superchargingmachine]

When the code is working it never exceeds the cranking value of 35DC but from memory it was going to at least 114DC and counting down slowly until it hit about the lower RPM set point then jumping to 114DC or so and repeating. I had it doing this a couple times today and don't really think it was anything I did as far as settings that got it to stop and act normally. Once the RPM it is within range of the fast recovery I can make it count down more quickly or slowly depending on the time I set.

Good idea about the 0DC for closed I will try that. The car idles at about 23-24DC

The car is a turbo MR2 and the IAT becomes heat soaked at idle. I am using EGO at idle but I can't give it enough authority to keep a steady AFR. I was hoping PWM could help compensate for my fluctuating IAT's and AFR, which cause idle issues.

[Keithg]

Interesting... wierd that it thought it needed to open up the AIC to get to a steady idle. If idle speed is below the set point, it will continue to open the AIC until it reaches the deadband.

There is a safeguard of a lower setting (not to go below X and a few tests to make sure it did not get a value below min at certain times. As for greater, I do not know. Please make sure it is not underflowing for me. (If it goes down and through zero).

KeithG

Posted by email.

[superchargingmachine]

It gets stranger. This morning without making any changes everything is working great again. I am convinced that when making adjustments at the wrong time I get the code "confused". It isn't "thinking" it needs to add that much throttle it acts like it is stuck in a loop.

Anyway, it seems that the basic code is working. If this happens again I will cycle the power to the ECU and see if it corrects the situation.

I set 0DC for closed and that is as low as it ever goes.

Do you have any idea if this valve can be blown open by boost? I just started a new calibration so it may have nothing to do with the valve but the car feels flat up in the revs.

[superchargingmachine]

Update, I took the car out to lunch and PWM got itself boogered again. This time it was a little different. I didn't make any setting changes since it was working decently on the way to work. The DC counts would count down from about 53DC to maybe 33-35DC and never get near the idle set point RPM and then it would repeat this behavior. After getting lunch it acted normally on the drive back.

I hope to catch this in a datalog at some point when I do I will post the datalog.

[Keithg]

Did you set your dashpot delay to 255? I think I should hard code this... This is the time it waits for it to decay to idle. If it does not decay fast enough (controlled by that 4 point cline on page 2), it will dashpot again and teh idle will pogo along.

It sounds like you are very close to having it working. Sorry it is so finicky,

KeithG

Posted by email.

[superchargingmachine]

No I forgot to set dashpot to 255 but I will before I go home tonight. I'm not sure that's it tho because the dashpot should just be an adder right? When this occurs the DC makes a large jump, at least 30DC maybe more, I have the dashpot adder set to 3.

I'm just happy to be able to give you some feedback. I have had to tune PID loops before and this seems very similar. It is never as easy as it seems like it should be.

When it's working it seems to work well. I think this valve is a little large for my engine as the DC adjustments are very small relative to the RPM I get. If 0DC to 100DC were 500-800RPM I would have more DC range to work with. I am trying to work in the 20-30DC range of this valve.

[Keithg]

No, not the dashpot adder, the dashpot settle time. The settle time is a time that it waits for idle to settle out. An adder of 255 will race the engine to the moon!


KeithG
Posted by email.

[superchargingmachine]

I found a different cal problem that seems to have helped PWM idle a lot. I was using the wrong timing table. My timing was retarded about 8 deg.

Then I set the dashpot settle time to 255.

Between the 2 my idle was absolutely perfect on the way home. I bet it is not moving 25RPM.

In my previous post I was thinking that the dashpot would only add the 3 or whatever is in the dashpot adder box. It seemed to be adding a lot larger value, which is why I didn't think that any dashpot setting would help it. What can I say, I was wrong:-)

NEW QUESTION:
Do the settings on the second page for rate of close control the closing speed when you go out of PWM range? The valve closes really slowly and I'm thinking if I were to jump in the gas I would have boost wistling out the valve until it shuts. It would be nice to have it close quickly if the throttle goes to a 50% value or something.

[Keithg]

funny you should ask. That is what closure speed is for. The value is teh time it waits before stepping it down to teh next step. Low numbers close quickly high numbers close slowly.

KeithG

Posted by email.

[superchargingmachine]

Keith,

When PWM has control of idle I need a relatively slow close speed to get a stable idle. The problem is that outside PWM idle control(IE normal driving)this causes the valve to close very slowly. It would be nice to have it shut quickly when the throttle is opened. Is this possible?

A little more feedback on dashpot settle time. If I set it to anything other than 255 I get stuck in a dashpot adder loop and idle never settles.

Anyone trying to cal PWM I recommend watching your vacuum gauge very carfully. It has been my bast indicator of what the valve is actually doing. Slow valve time movements seem to work best under the widest range of conditions. The other thing that helped was setting my cranking and minimum settings to a narrow range. I am only using a min of 20dc and a cranking of 26dc this seems to limit the movement making calibration much easier.

I will post my settings to my success story after I get a couple more days under my belt and feel more confident in them.

[Keithg]

There is really no reason to make the dashpot settle time less than 255. That is why I think I will eventually hard code it at 255. This it the time that it will bypass the rpm test which closes the AIC. See, the theory is that teh AIC should be closed when the TPS is above the threshold, always. If the revs are over the rpm threshold, AND the tps is closed (engine braking), the TPS should also be closed. It is this second test teat that 255 bypasses to allow the RPMs to drop to below the rpm threshold. If you want it to close faster when regulating idle, decrease the upper RPM or decrease the fast recovery delay value.

The closusre speed is the one that affects how fast it closes once above the thresholds. 0 means no delay as it steps closed and it will step closed in .1s * (currentdc-closedc). if you use a value in this box, it will stop at each step as it closes for that many 0.1s

The dialog boxes show 0.01s, but it is actually, 0.1s for each time the loop runs.

HTH

KeithG

[superchargingmachine]

I've been driving the car almost 2 weeks now with PWM enabled and it does sort of work. I have yet to find a set of setting that keeps a steady idle under all conditions. It seems that if any other parameter, in my case IAT, changes too rapidly during idle, PWM can't keep up.

I just want to give some feedback to the guys who write the code. So here are my observations.

1) What happens on my car is that I cruise at 3200 on the high with IAT's in the 80-130 range, depending on the day and how hot it is, then I stop and heat soak starts. My IAT's gradually rise on some days to 160. This drives PWM crazy.

2)The other thing that continues to happen is that PWM occasionally gets "stuck" where it counts down then jumps up and repeats. The car was doing this last night but it is gone this morning. Yes, dashpot settle time is set to 255. I don't think that this is a dashpot issue. I am guessing the Mapdot accel and PWM are interacting since earlier in the day I raised my cranking DC for better starts. I think that my PWM now may occasionally trigger Mapdot accel.

3)One last minor issue or observation is that I never seems to get to high idle RPM??? It could be due to the fact it is pretty warm outside and the car warms up very quickly. If the car sits overnight it is still 80F(27C) deg in the morning. I have high idle shut off at 140F(60C). If I set -20C for high idle and 1500RPM for high idle on and 60C and 900RPM for high idle off does PWM do a linear transition between the 2 points?

[Keithg]

Yes, the transition between high to slow idle is a linear interpolation through the temp range.

I have also seen the heat soak issue. What happens is it is lean, in my case, and it causes idle fluctuation which PWMidle code cannot keep up with.

The 'stuck' in an oscillation issue may be due to mapdot and other things interracting to cause an idle to be difficult to control. All I can suggest is that a stable idle without pwmidle is kind of required to have isle speed regulation. If your idle AFRs are lumpy around idle without PWMidle, it will be prone to hunting around. The heat soak issue can be helped with the coolant modified airdensity function. Beware, it is not intuitive, but you can set its driver to IAT and flatten out the airdensity curve above certain IAT values which would help the lean idle issue. Is EGO correct turned on at idle? In the lores code, this can also cause a hunt.

As a side note, I am running the hires code with this idle code in it. I have EGO turned on at idle, but the hires code is so much smoother around idle (especially with large injectors and their low PWs). It may be that your selection of components (injectors, etc) may be causing the code to be too lumpy at low pulsewidths. If you are running high impedance injectors, you may want to try the hires code.

The first thing I would do, though, would be to really focus on tuning your VE bins near idle. I know the megamanual says they should be flat. That is not my experience. They need to be adjusted such that AFR remains constant.
Second, a timing valley at idle with 5 deg or more rise below idle can help as well.
Third,try adjusting the airdensity correction to compat lean idles at high IAT.
after this, maybe try the hires code and see how it helps if you can use it.

KeithG


Posted by email.

[superchargingmachine]

Thanks for the feedback.

Most of my idle work has been done in the fast and slow PWM idle settings not in the spark or fuel table. I will try your suggestions.

Can you suggest a starting point for coolant modified air density?

I have been considering hires but the car is running soooo good I am reluctant to change much.

Is Hires stable yet? I know it is relatively new and don't want to switch until it seems to be working. If yes, where can I get a copy?

[superchargingmachine]

Is anyone else having PWM idle act strangely form time to time. I have had it work perfectly then it seems that the sun goes behind a cloud or something then it starts bouncing.

Is there a way to completely disable dashpot just to see if this is the cultprit?

I also suspected it was mapdot accel interacting with PWMidle so I went back to TPS based accel but this didn't fix the problem.

I will say that when PWMidle is working I love it.

[fred_bx16s]

hello

i use you setting but my pwm no work correctly

the valve is open for starting engine and rpm up at 4000 but no down

is verry strange ( lol )

thank if you are a idea

fred ( france , limoges , citroen bx 1900 16 valves , ms1 v3 code 29q2 )

++

[superchargingmachine]

I never updated my settings in this thread.

Dashpot settle =255(Seems this is the only setting, shich will work)
Fast recovery = 40-60
Slow recovery = 140-190

I had it working with these settings. But today all it does is cycle up and down!

[Challengr]

I have struggled to get a smooth idle too. I run a Ford valve on a V8 with a mildly lumpy cam.

What I did that seems to work well is to keep the deadband range narrow, set the fast and slow recovery RPM ranges low with quick recovery times.

I think this makes the valve react much quicker but in smaller increments to smaller rpm changes. The result is that the valve doesn't seem to over-shoot as badly, which would then cause the dashpot to kick in.

For additional idle support, I added 6 degrees of timing to the MAP at a range 100 RPM below the target idle speed.

This works well once the warm-up is complete. During warm-up, I think the dashpot gets activated which then causes a surge-loop.

Here are the values that work for me:
Deadband: 40
Upper Limit: 100
Lower Limit: 20
Fast Recovery: 3
Slow Recovery: 10
Closure Speed: 6
Startup Delay: 80

Dashport adder: 300
Dashpot settle: 255
Dashpot adder: 5

Idle Valve Frequency: 200

-Jeff

[superchargingmachine]

I am also running the Ford valve but on a stock cam 1.6L Toyota.

I will try your settings tonight. You upper and lower limit and recovery values are a lot different from mine. So maybe there is hope.

I think your valve frequency is a different from mine too.

The only things I have changed lately was the high idle speed and the cranking dc to get to that RPM.

The dashpot surge loop is where I think I get stuck. It doesn't appear to be a controller hunt but a slow countdown then jump back to the same point.