Dashpot decay issue MS2 3.3.alpha4 gslender2.6

[sd1nl]

Hi,

Running 3.3.alpha 4 gslender 2.6

Having trouble with adjusting the dashpot adder and decay function. Can't get this right on my Rover 3.5 engine.

This is what happens: car is in neutral or clutch depressed. revving up to 1500 rpm or so, lift the throttle, as soon as throttle is lifted idle-adder steps are added. So far so good.

Problem is that the adder steps have already dissapeared before the rpm has dropped to say 1300 rpm, so by the time the rpm drops to idle the adder is no longer there, and I need it to keep it from falling below idle rpm.

I have tried the dashpot decay settings from 2 to 5, but nothing seems to work quite well.

It also seems like sometimes the dashpot adder is not added at all. What are the conditions that have to be met for the dashpot adder to activate?

Is there anyone out there who knows how this works?

Thnks, Rene

[sd1nl]

Let me clarify with this megalogviewer graph.

I blip the throttle, after throttle release the adder adds 10 steps to the IAC, but as soon as the RPMdot goes below zero (meaning the RPM goes down again) the adder is substracted again. This makes it useless, because, I need it when the rpm's are near to idle.

Also there seems to be no decay at all, the adder just seems to switch off

Dunno why this is happening. Does the dashpot decay system only work with PWM idle valves, and not steppers maybe?

Rene

dashpot decay

Dashpot decay.png (117.06 KiB) Viewed 753 times

[Greg G]

Are you using the initial value table for idle valve duty? Perhaps the value should be higher. The purpose of the dashpot decay is to slow the RPMdot down just above idle, allowing a smoother transition into the target idle RPM.

[sd1nl]

Hi Greg,

I've been meaning to try this this week, but I don't think it will change anything in the decay function. I've been thinking about what is happening: because the Rover V8 is a big slow responding engine, after I have lifted the throttle the RPM is still climbing and before it's top point. The RPMdot near the RPM peak is very low or even zero so the adder-decay will very high, as you can see in attached graph.

Is it not better to have the possibility to set the rpm-adder-decay in ms (or seconds). That way I can make it reliable and consistent every time.

Rene

[gslender]

Increase the decay factor

G

[sd1nl]

By increase it you mean decrease the decay factor? That is what makes the decay longer (slower), or am I wrong? Decreasing the factor is difficult, because I am already on decay factor 2, and decay factor 1 has the effect that the RPm won't always come back to idle anymore. There are no settings inbetween 1 and 2 unfortunately.....

Is still think the current method wont work on all engines, because in my case it's measuring the rpmdot at the wrong moment. An adjustable decay in seconds will work better I guess. It's only a suggestion

[pigga]

Problem is that the adder steps have already dissapeared before the rpm has dropped to say 1300 rpm, so by the time the rpm drops to idle the adder is no longer there, and I need it to keep it from falling below idle rpm.

Then the Adder itself is probably too small?

I have tried the dashpot decay settings from 2 to 5, but nothing seems to work quite well.

Try to increase the daspot adder instead. The decay factor only determines how fast the daspot will be decreased. Both together work like a damper when TPS snaps into closed position. It is depending on RPMdot. The higher RPMdot (i.e. the jittery RPM is), the slower dashpot adder will be decreased.
If the adder itself is too small it will be decreased to zero very quick anyhow (no matter what decay factor you choose)

It also seems like sometimes the dashpot adder is not added at all. What are the conditions that have to be met for the dashpot adder to activate?
Is there anyone out there who knows how this works?

Yes, adder is only applied when the PID delay timer has expired. This is necessary because otherwise you could ratchet idle RPM up and up by tipping slightly on the gas several times. becaue the adder would be applied every time....

As one of the two G's pointed out:
Use "Intial value" instead of "use last good known value" (if you havent done yet). Make sure you fill the table with sensible values. Try it with no (or small) daspot adder in the beginning. You are on a good way when initial value of RPM settles a bit above the desired target at all circumstances. That means initial RPM matches at different Manifold Air Temperatures and different target RPMs (when CLT temp is low) your engine will see during daily driving.
Then screw up your PID Settings. Make sure that "Adaptive PID reduction" is 100% all across the board. A rough estimation for initial PIDs: Set I very low, D=0 and increase P until IAC starts oscillating. Take 50% of that critical "D" Value. Then increase "I" until you get good response of IAC at CL idle when external load to engine is applied (switching to "D" or whatever). Add some "D" for dampening to get rid of oscillations.
Then tune up "Adaptive PID reduction". If your PID settings above give stable idle, you for instance may double P, I and D and reduce PID at target to 50% (see gselnders documentation for how the PID adaptive reduction curve may look like). This gives a more agressive loop the further you get away from target.
Finally make sure that you set dashpot adder large enough that your IAC "catches" RPM above target under all circumstances. And choose decay factor large enough so that it will ramp idle to target reliably, i.e. not too fast and not too slow.

I used gslender code on a car with automatic trasmission and Idle is totally perfect, even without an "idle up-switch" wired to MS. You have no noticeable RPM drop when you switch from "N" to "D" and vice versa.
Only heat soak at hot engine is a minor problem at the moment, but it's not the last version at the moment that's on this car.

[sd1nl]

Ok, I'll give it a go with a bigger adder and play around with the decay factor and the initial value table.

Don't wanna mess around with my PID values though, because my idle (once settled) is fine and stable.

Rene

[Perkele]

I have had also trouble getting the Dashpot decay to work at all. It seems to jump traight to the current "initial value". And I have tried to add the decay factor 10-->60 with no success, but this was the first time I read that bigger value makes it decay faster so maybe I need to test alot lower values.

Other thing if the adder is added after PID delay how will it ever catch the idle? Because minimum PID delay is 2sek and when the adder RPM is for example 400rpm. It will be far below target RPM after 2sek or am I misuderstanding something?

Thank you.

[Greg G]

Don't wanna mess around with my PID values though, because my idle (once settled) is fine and stable.

Rene

Have a look at the PID reduction window option- allows you to have your PID less aggressive near target, yet aggressive when farther away. Like having your cake and eating it too

[Perkele]

Here's my setting and log concerning my previous post.
Could someone please help me to make the dashpot work as supposed and have a look what's wrong with my settings?

[sd1nl]

Your decay factor is way too high. The adder will not be doing anything like this. My suggestion is to make the "idle activation adder" a littlebit lower (250-300rpm), and the adder itself bigger. I did some experimenting today and a decay factor of 4 seems to work for me now so I think yours should be like 3-7. Also I opted for a PID delay of 2s instead of 4s now. Seems to work better.

The trick with the adder decay factor system is to make the adder a lot bigger than with the "old" adder system, and then tune the decay factor for optimum RPM drop. It does work, but it is a fine balance between adder and decay settings.

adaptive PID is inbcluded in my settings, and it makes the idle nice and stable. But it does not help for a smooth and consistent rpm drop, as it will only be active after the pid delay.

Rene

[pigga]

Other thing if the adder is added after PID delay how will it ever catch the idle? Because minimum PID delay is 2sek and when the adder RPM is for example 400rpm. It will be far below target RPM after 2sek or am I misuderstanding something?

Hi. I must admit that's easy to mix it up:
Daspot-Adder is the Amount of Steps/Idle Dutycycle that at the beginning will be added to the initial value. And it's the amount that will be decayed.
The "RPM Activation adder" is part of the condition (s) that have to be met to activate CL idle.
once upon a long ago (with the standard code) it has been something like that (please correct if you find errors in the diagram). You just have to replace "Got to last good known value" by "Go to initial value from table" and replace "Wait timer" with "Decay dashpot adder by dynamic dashpot decay"

[sd1nl]

Altough I have my "reurn to idle" reasonably tuned, it's still not nearly as good as factory ecu's. How do they manage such a consistent rpm drop every time?

[pigga]

Altough I have my "reurn to idle" reasonably tuned, it's still not nearly as good as factory ecu's. How do they manage such a consistent rpm drop every time?

Well that's quite easy: Take hundred of Engineers and some million $$$ budget and off you go

[sd1nl]

True! But I'm still trying to get the best result as possible

[John151]

Altough I have my "reurn to idle" reasonably tuned, it's still not nearly as good as factory ecu's. How do they manage such a consistent rpm drop every time?

Well that's quite easy: Take hundred of Engineers and some million $$$ budget and off you go

OEM engineers are also targeting specific vehicles for specific needs - MS is all things to all people, which is never an easy task.

[Perkele]

Thanks Rene and pigga for the help I now got the dashpot to work guite well most of the time. I'm still seeing that sometimes it doesn't add the adder but I'll try to catch it to a log.
But like I said much better now.
Thank you.

[pigga]

Hi all.
I think today I found (at least for me) a solution of striking simplicity at least for CL idle, Adder, Daspot decay and so on.
As sd1nl stated above, it is always a bit tricky to get
a) a reliable RPM entry point for CL idle (slightly above target) so that NOT
b) RPMs enter too much above idle target and take too long until it goes down.

I got a bit inspired by the way OEM ECUs do it.
Today I switched Idle ignition advance from "CL Idle" triggerd to "Manual". I set TPS rather high (30%), MAP threshold very low (30kpa) and max RPM quite high (2000RPM).
My Idle timing curve is "adaptive" (i.e. depending on cl idle RPM target error) and corrects +/- 20° within 100RPM.
As a result, Timing is retarded as soon you lift throttle. So you make sure RPM gets down instantly. On the other hand it catches reliably 100RPM above target.
The advantage is, that engine always revs down consistently. Further more transient between overrun and low TPS is much smoother.
Try it out?

[sd1nl]

Sounds like a good idea, but I run fuel only, so no option for me