MS3 Dashpot Decay Rate

[iti_uk]

Hi,

I'm running one of Reverant's MS3 ecus on my 2003 MX5 and I'm having trouble with the dashpot setting.

From behind the wheel, the revs drop okay to idle on throttle lift-off, but then bounce back up a few hundred revs and only settle after a few more seconds.

Examining the logs shows that the dashpot value is being added to the target CLDuty value and being held there for about 1.4 seconds, whereas it takes only 0.7s for the engine to drop from the point of the dashpot being triggered to the target idle. The result is that the revs begin to rise again, starting a brief oscillation.

I've seen that the decay rate and on-time of the dashpot can be configured on MS2 controllers, but the MS3 firmware seems to be missing this option. Is there something hidden I am missing or will this be available in a future firmware release? (currently using 1.3.4).

Chris

[piledriver]

bump to subscribe as I have been fighting similar issues, figured it was just me, but as the smart guys always say:



I don't have any short datalogs, but I'll try to make one rsn

[muythaibxr]

We removed the dashpot decay because it didn't work.

Usually the problem you describe is because of fuel and ignition advance and/or overrun fuel cut. You should tune with open loop idle and get the engine to come back to a stable rpm without dipping by tuning fuel and ignition first, then enable closed loop idle.

[piledriver]

My setup works fine when warm but despite jacking with the idle air presets forever it still wants to occasionally "take off" while warming up.
Its almost like some other correction factor is needed, but I'll keep at it.

I had it nailed down pretty well until I installed proper sized hoses on my IAC valve so it could flow enough air to allow cold starts unassisted.
When restricted it worked OK otherwise, now it seems to want to flow too much air at times.

I'll see if it will do it while logging now.

[dontz125]

If it's fine once warm but wants to race while cold (especially now you've added more air!), I think you need to bump you WUE values.

[piledriver]

If it's fine once warm but wants to race while cold (especially now you've added more air!), I think you need to bump you WUE values.

Done, and the added WUE did help, but am surprised the MAP sensor>ve loop isn't compensating as fast as the IAC does.
I do have a large oil catch bottle in the evap/manifold vac/case evac setup, may be causing lag at lower MAPs.
(MAP sample is taken from the bottle, ~10mm ID x 400mm line to plenum, I suppose I could change it but it has worked well for years)

My idle air setup is more complex than most, I run the crankcase at ~85 KPA at idle & cruise through a variety of means, my "base idle air" is routed through the crankcase via a restriction on the #3/4 side rocker box breather, via some check valves...
(Think ~1.8T case evac setup, sans the boost fed vacuum generator, although one is available and planned for use)
... need to add an exhaust vac source to provide more evac under load when the hair dryer/s eventually go on.

[arran]

Using closed loop initial value table?

This behaviour sounds to me that the initial values are too high for cool coolant/air temp (depending on what you have set the y axis to).

Arran

[piledriver]

Using closed loop initial value table?

This behaviour sounds to me that the initial values are too high for cool coolant/air temp (depending on what you have set the y axis to).

Arran

Yes, it needed more fuel when the IAC was open.

[iti_uk]

Okay, I've switched over to open-loop idle control and done a bunch of tuning using this guide; http://www.diyautotune.com/tech_article ... uning.html

I've aimed at 950RPM idle speed and have done some tuning of the fuelling to get it running at 14.5-14.7 AFR. I haven't touched the WUE settings since the change yet, so I'm sure they're out.

I've gone for a run (on a private road) and taken a log. I've zipped the log along with the msq in the following directory (size limitations on this site). I sat idling with a couple of throttle blips at the end of the log.

One thing I noticed is that, although it seems to drive fine and return to idle happily while moving, a throttle blip in idle will result in a droop to ~600-700RPM before it returns to 950RPM. Is there anything which can be done about this before I change over to closed loop?

https://www.dropbox.com/s/v4gxww0rhw8o2 ... 9.zip?dl=0

Any critiques of my tune and log would be most appreciated.

For reference, my setup is;

MS3 running firmware 1.3.4
Mazda MX5 2003 1.8 Sport
Lightweight flywheel
550cc injectors (ready for turbo build)
Cone filter with shorty intake (purely as a stop-gap until I get the turbo+intercooler installed)
Stock everything else

[muythaibxr]

I will have to look at your log and msq.

[iti_uk]

I will have to look at your log and msq.

Thanks.

[iti_uk]

Update:

I've changed my ignition table back to a less advanced one I initially had, as it has been mentioned that my current one was very advanced all over the table.

Here's a pic of the new (old) table;

https://www.dropbox.com/s/zk5lyg3qjqg4d ... 1.jpg?dl=0

[muythaibxr]

As I am overseas at the moment, it will take me until mid next week to look at it. Just to let you know.

[sd1nl]

Wouldn't it be great if we could get rid of all the surplus of settings concerning the idle control/decay/initial value tables etc and have a system that could regulate all this by itself.

Take the initial value table: it takes forever to tune and is impossible to get right. Isn't there any way to improve this?

Rene

[jsmcortina]

Take the initial value table: it takes forever to tune and is impossible to get right. Isn't there any way to improve this?
Rene

I think you need to post more data to back that up!
On my engine (lazy V8) it was no big deal.

James

[iti_uk]

A thought of mine is the following;

I've been previously told that CL idle is best tuned using an IAT-based initial value table.

The OL Idle (warmup) is tuned using a curve based on CLT, not IAT.

Why?

Shouldn't it be better to use a CLT-based initial value table on CL Idle?

And regarless of problems inexact idle tuning, my initial question regarding the Dashpot decay rate (or just its time active) still applies - an engine will a perfectly setup idle will still have an rpm bounce if the dashpot stays active/applied for too long. We need to be able to adjust the active time. It looks like it's currently set to 1.4s (according to my interpretation of my log), this parameter must be buried somewhere within the firmware, even if it's not currently adjustable...?

[iti_uk]

Wouldn't it be great if we could get rid of all the surplus of settings concerning the idle control/decay/initial value tables etc and have a system that could regulate all this by itself.

Take the initial value table: it takes forever to tune and is impossible to get right. Isn't there any way to improve this?

Rene

I think something halfway between an initial value table and the "last value" method should be the best way to do it. I suppose that would end up as a sort-of-idle-autotune. Just set the target idle RPM curve as required, and then let the MS take an "initial value" as it enters idle, but then as it corrects to meet the rpm, it updates the table cell for the CLT/IAT/RPM for the next time it returns to idle...?

edit:
Regardless of my wish-thinking, how exactly does the standard MX5 ECU handle idle control, and why can't we emulate that with the MS? It seems to me that we can plug-n-play any MX5 ECU of the correct year into any other MX5 of the correct year and it'll run fine, but we can't seem to get a working idle setup on the MS which can be transferred between MS-based setups running the same engine. Why does the standard ECU seem so capable of compensating for non-fuel-related engine changes (referring mainly to my lightweight flywheel here)?

[sd1nl]

edit:
Regardless of my wish-thinking, how exactly does the standard MX5 ECU handle idle control, and why can't we emulate that with the MS? It seems to me that we can plug-n-play any MX5 ECU of the correct year into any other MX5 of the correct year and it'll run fine, but we can't seem to get a working idle setup on the MS which can be transferred between MS-based setups running the same engine. Why does the standard ECU seem so capable of compensating for non-fuel-related engine changes (referring mainly to my lightweight flywheel here)?

I was wondering the same thing. How do factory ecu's manage to get a rocksteady decay to idle and idle speed under any circumstance?

[jsmcortina]

A thought of mine is the following;

I've been previously told that CL idle is best tuned using an IAT-based initial value table.

The OL Idle (warmup) is tuned using a curve based on CLT, not IAT.

Why?

Shouldn't it be better to use a CLT-based initial value table on CL Idle?

No.
They are two different things.

With open loop idle you are saying "At this engine temperature open the idle valve this much"

With closed loop idle, there is a target curve that says "I want X RPM and Y CLT"
The initial valve table says "To get X RPM at Y air temperature you need to open the valve this much"

James

[iti_uk]

A thought of mine is the following;

I've been previously told that CL idle is best tuned using an IAT-based initial value table.

The OL Idle (warmup) is tuned using a curve based on CLT, not IAT.

Why?

Shouldn't it be better to use a CLT-based initial value table on CL Idle?

No.
They are two different things.

With open loop idle you are saying "At this engine temperature open the idle valve this much"

With closed loop idle, there is a target curve that says "I want X RPM and Y CLT"
The initial valve table says "To get X RPM at Y air temperature you need to open the valve this much"

James

I understand, but why? It seems to me to be a bit of a band-aid fix to change target idle speed according to CLT, rather than adjusting the IACV.

I know it's asking a lot, but it would be cool to have a 3D target map, with RPM on the x-axis, IAT on the y-axis and CLT on the z-axis. Maybe a 6x6x6 table lol.

Still wonder how the original ECU handles all of this.

edit:
I had previously performed a mapping log where I manually controlled the IACV using the Testing function in TS. I started the logger, then started the engine and once idle had settled, I immediately started stepping down the IACV until it reached 700rpm, waiting a second or so for the rpm to stabilise between each step, then I stepped it up (allowing stabilisation with each step) up to 1300rpm, then back down to 700, up and down again and again until I got up to 90degC CLT. I then repeated the process after opening the bonnet (colder IAT at higher CLT). Once done, I took the log, and mapped the IACV duty, CLT and IAT at each 100rpm interval in the log.
The result of this study showed very little change in IACV duty requirement for each RPM step according to IAT when compared to the required change according to CLT.
I would guess that regardless of target RPM (and a cold engine may well want to idle faster than a cold one), there should be a table covering required IACV duties for various RPMs taking CLT into account.