Occasional rough idle?
 

Hi everyone,

I'm having a little problem here and need to know if anyone else is having it as well. My MS is fully tuned, fuel and spark, and strictly speaking runs great.

Except for this one issue: When the coolant is warmer than 50C (122F), the idle sometimes feels a little rough, like a cylinder is randomly misfiring once every one or two seconds. If the MS is also controlling the ignition, the revs will bounce back and forth from 600 to 1200rpm, like I have lumpy cams installed

By "sometimes", I mean that I may start the car, warm it up and it runs great until I switch it off and restart. Then it may (or may not) exhibit the problem. It is mostly an on-off problem. It will either exist from start to finish, or not at all.

If you gentlemen could be so kind as to say "Yes, I think I have it." or "No, I have experienced anything like it." then I woul be most grateful. I need to know if I am the only unlucky bloke, otherwise maybe I can talk James and the other good MSnSe developers into having a look.

I have a complete description in the MS-extra forums, other setups seems to have it as well: http://www.msextra.com/viewtopic.php?t=22625

Thanks,
Jim

It happens to me every once in a while also. I've tuned most of it out by slightly modifying both my warmup enrichments, and the ve table. I doubt I'll ever get it completely ironed out but now when it does it it only fluctuates between 700-900. I don't think my tables would be of much use to you though since I have a 1.6L, and I had to modify my values from the 1.6L base settings from other 1.6L .msq's.

Brian

The key symptom is that it randomly appears and disappears (completely) as you switch off the engine and restart it. It may be warmed up all the way and running fine, you switch it off and restart, it idles badly. Switch it off again and restart, it runs fine. Repeat as many times with randomly selected bahaviour when restarted (idle fine, idles roughly).

my car does something like this. it'll idle really rough randomly. it doesn't dip below an indicated 700 rpm or above 1k ever, but it IS inconsistent.

and it's running on the stock ECU.

and I've cleaned out my intake manifold.

do you use the stock IAC? what about VTCS?

It's not doing it with the stock ECU (I can switch the injectors back to the stock ECU one by one). The IAC and everything else is still handled by the stock ECU.

Reverant, I get the exact same symptoms on my MS'd Roadster.

I also get the ignition spikes that Foundsoul gets and mentions on MSExtra

Richy, I get the ignition spikes as well. Not good. I think we need to "escalate" the importance of these issues in msextra.com...

I don't get that problem.

On real cold starts, the engine will sometimes die or idle really low (400rpm), but a little tap of the gas wakes it up.

Al, which version do you run? MSnSe (029q2, 029t), HR...?

Yeah I've been meaning to post to add an extra voice, it doesn't worry me *too* much atm as I'm n/a but I will have a supercharger in a month or two and I don't want to grenade my engine :(

I'm also N/A at the moment, that doesn't make me any less concerned, there aren't only spikes "out there" but dips as well, you can feel them pretty hard when starting from a red light on an upward slope, you feel the engine bogging down as if you instantly lifted your foot off the throttle. Very irritating and bad for the drivetrain as well.

Anyone else? Yes or no...?

29q2

After looking back at some datalogs, I realized I have this problem also. Ignition timing dips to 13 when 22 is called for, and the idle varies. It took me a while to realize that it isn't always there, and that it's all or nothing. Either it runs great, or it has both problems. I'm updating the code today which should solve the timing issue.

Mschlang,

check your PMs at msextra.com. Thanks,

Jim

I'm having the same problem. I'm still at a very rough tune, though. Anyone tried a different trigger predictor algorithm?

Arga,

you have this issue on the MS2/extra?

Jim

Just Bruce&Al code v2.8. Yes, same problem.

I was thinking about it and if spark control is causing this then either
a) The dwell is toggling between values
b) Spark advance is fluctuating
c) The timing base is fluctuating

I'd guess 'b' is the most likely. My advance table is evenly distributed across the rpm band and the biggest deltas are at low rpms. It might be worth it to move some of the bins at higher rpms down around 1k rpm for better resolution at idle.

I pulled my wiring harness out last night for modification so it will be a few days before I can try anything.

Can one of you post a data log of this sort of event? *EDIT never mind... saw it in your msextra post

One thing to remember is that the EGO sensor senses oxygen in the exhaust. If it's reading very lean that can mean a misfire (extra unburnt oxygen), not necessarily that it's getting too little fuel. There were some posts in the msextra forum which seem to forget this little fact. So basically you can ignore to O2 sensor when the car is misfiring, it's meaningless.

Anyone "new" to the Megasquirt scene having this problem as well?

I've not noticed this problem as much latey, It's still happening, but I'm no longer running stock injectors, they're now (cleaned and flow-tested) 600cc injectors. I don't know if because of the loss in resolution because they're high imp injectors it's masking the problem slightly?

As for your spikes and dips on the ignition output, that was taken VERY seriously and fixed quick. 029v or later fixes that right up. As for the random minor misfire at idle I'm not sure if I've experienced the same thing you guys have or not-- I do sometimes hear what sounds like a very minor miss at idle that I've never been too concerned about, I don't see evidence of a miss in the AFR datalogs (which would look like a lean spike in the AFR logs due to the unconsumed air in the pipe). At any rate it looks like James is looking into it based on the thread linked at the beginning of this thread. My experience thus far has been A) it doesn't effect my AFR so I'm not sure it's a miss, and B) it really doesn't seem to effect anything. But the bottom like is James is a wiz at this stuff and I'm sure he'll nail it. Definitely go ahead and run 029v or later though to get rid of those spikes and dips...

Jerry,

I already run 029v, I also had the spikes and I participated in the "Ignition spikes under boost - scary" thread. As for the AFR, whenever the problem appears I do get leaner, from 14.7 to about 16-16.5ish. I'll scope the injector outputs today.

Jim

Cool-- if you get anything useful scoping the ignition I/O I'd suggest posting the traces in that thread on msextra.com- James appears to be watching it pretty closely. At one point I had this going on with my MR2 but I haven't noticed it lately-- I'll see if it is and if so I'll bust out the PicoScope and get some traces for James myself...

I have to bring this thread back up because I had the same problem for the longest time.

I have an MSI Extra High Res running on an Escort with a GTX motor (essentially a BP with lower compression) and GTR turbo pushing 16psi of boost.

Here's the issue with the rough idle. Our motors like to run rich at idle. And I mean 13-12.5:1 rich unless you oscillate from lean to rich. The MS O2 feedback controller is not setup like the factory one and will not oscillate properly causing idle roughness. Don't worry, your gas costs won't go up dramatically if you have a rich idle.

Do not run closed loop with the O2 sensor at idle. Turn on closed loop at 1200-1500rpm

Tune your VE table slightly on the rich side.

Watch your intake air temp/density compensation. MS starts to pull fuel above intake temps of 70F. This can be seen in the gammaE if I remember correctly. As you warm up the car, the sensor heat soaks not being representative of the actual intake temp. MS pulls fuel and you start to have a rough and unpredictable idle. Correct this by doing coolant temperature correction and ADDING fuel when the coolant is hot up to 1500rpm. Do this until the car runs smoothly.

At this point you should never have a rough idle anymore ... I certainly don't anymore. Hoeve, if you need to pass emissions, you have to change things a little or you will fail on idle.

When tuning the MS in general be aware of all its compensations ... little problems drove me nuts until I learned to watch all compensations and then everything became really simple. Hope this helps out.

I agree completely that the BP likes a rich idle. I have had much better results w/ 13.5:1. The thing that kills me, though, is that the stock ECU manages 14.7 and holds the idle rock steady. Last night I started setting up a blended Alpha-N just for idle (<1250rpm). It is rock steady but my tune is still very rough. Alpha-N seems really touchy and a lot less error tolerant.

I've also found I had to enrichen the mixture at intakes temps above 100 F. I see than a lot when I idle for long periods in my garage.

My stock ECU oscillated from 14.5-15.3 and kept the idle smooth. Does the miata ECU keep a steady 14.7?

My stocker oscillated from 14.5-15.3 or somewhere close to that.

So the control is similar to the GTX engine ECU .... try what I suggested and your idle should be a lot smoother.

i like to have mine at 17:1 and pretend that it is purring at me.:pitlab:

I might just do that.

Poor wording on my part. The AF ratio oscillates but the rpm stays rock steady.

the idle is maintained by the IAC valve ... but even without one the idle is pretty steady.

Sorry, but you are plain wrong. The rough idle issue is tracked to a firmware bug. Specifically, one of the our injectors is randomly not opened/squirting, same goes for the coils/ignition. It has nothing to do with AFRs. I DO have a perfect 14.6-14.9 AFR at idle and the car runs BEAUTIFULLY. Just like stock, maybe even better at times. Perform a reset and it gets fucked up. Reset again and it propably is ok. Repeat N times within S seconds (with the IAT and CLT not changing more than 3* F) and the results vary tremendously. Its a software bug, not a tuning issue.

Jim

Jim, maybe it's MS2/Extra time? ;)

Yes. I'm building a new block right now so I'm cash-strapped, but it's already planned.

Jim

Having it run richer I never have the occasional rough idle issue anymore. So what is your suggestion for a fix and how did you get your car to run consistently well at 14.7 afr?

I bet MSII might be part of his answer.

I'd love to know a different solution to a consistent idle. I have read on MS forums that even people with MSII suffer from a rough idle and no one mentions a bug.

Reverant, if you have some insight it would be great!

follow his link ;) I have the same issue.

What link??

Read the first post in this thread.

Jim

Posted in the forum ... my feeling is that it has to do with the IAT correction on hot restarts and the physical nature of the engine itself .... while the PW is so small (idle) fuel delivery that is interrupted by restarting the MS can enlean or richen the mixture.

Running without the IAT correction at idle and rich I eliminated the problem completely.

Keen to see if it can be eliminated otherwise.

Hrm. Since I'm going full standalone with my MSEFI, I was curious what the solenoid on the FPR vacuum line was for. Here is the response I got from my local miata gearheads community:

"Factory ’95 manual indicates the solenoid valve raises fuel pressure to reduce the chance of vapor lock during hot restarts & hot idle, hot being coolant over 194f, intake air over 158f in the MAF."

It seems like that could point the need to enrich the fuel mixture on hot idle, at least on the older cars. I'm guessing 99+ doesn't have that solenoid since it's a returnless system.

-Mike

Then you are not experiencing this specific problem but something similar in symptoms. It has nothing to do with IAT correction - read my post above ["Repeat N times within S seconds (with the IAT and CLT not changing more than 3* F) and the results vary tremendously."].

Jim

I've had the exact problem. Resetting the MS by doing a "Burn" brought the car back to a very smooth idle or kept it in poor idle. It was quite random and unpredictable while trying to maintain a ~14.7 AFR idle.

I am just sharing what cured the problem for me .... running rich at idle and most importantly not allowing the IAT to pull fuel at idle as the sensor heat soaked. It runs smooth every time now.

This leads me to believe that at idle since the pulse width is small enough and depending on your injector wiring the injection event happens at different times in the engine cycle (ie pistons are at different positions with respect to each other in the 4 stroke cycle depending on which of the 4 triggers injection happens on) a physical characteristic of the injection process and exhaust gas scavenging coupled with an injector timing issue is causing the random idle issues.

I run 2 alternating squirts myself but I have heard that if you are able to run 4 alternating you will be in much better shape. The idle PW becomes too small at this point for me. 4 alternating would at least ensure that fuel is injected at the same time within the cycle to each cylinder.

Those of you experience this issue-- please email me your .msq file with the subject "Miata idle issue-- slight misfire"

I got together with James, Ken, Bruce and actually a crowd of people developed that helped some, this past weekend at the meet. I was able to reproduce the issue on my '91 and I think we've got it nailed... before I spill the beans though I want to see your .msq files to verify something.

Please only send me you .msq's if you feel you're having the exact issue described here.... send them to websales AT diyautotune DOT com

I might be experiencing this issue, my idle just sucks all the time though :)

Please do post your hypothesis, I can't wait to try it out!

You have email!

Jim

I think mines also missed once at idle, and that was mainly due to the a/c as it would do that before the MS....

from me too!

Dammit so what's this new theory??

Chill the hell out man. They're probably not going to release any info until they have the time to analyze everyone's MSQs to make sure there is consistency amongst people that have this problem.

This is not some commercial product release, a bunch of people using a DIY controller got together and came up with a theory and mentioned it on an internet forum - there is nothing wrong with sharing their findings and saying they are untested.

Perhaps you should be taking things a little less seriously.

Sorry if that came off serious. I want to know as much as you and i understand what you're saying. Hopefully we'll get to see what's up before they release the PnP. I need good idle now and can't seem to get it right :(

No worries, have you tried my suggestions? I have a great idle now, but it is in the 12's A/F wise. So it no longer depends on "how I start the car" or when I reset the ECU - the idle is always smooth but not emissions friendly.

I really do believe that the issue is with the idle control in that the feedback controller. The O2 sensor feedback controller at idle is not designed to respond like the OEM one does and oscillate the A/Fs as consistently causing the bad idle..

Wow... Patience guys ;). I just wanted to compare some of your .msq's to see if it lined up with our findings over the weekend. I had some more testing to do before I could share anything that would be universally helpful -- but I think I can do that now, hopefully.

I can tell you that Ken and James both feel pretty confident it's not a code issue. We had a scope out in the parking lot for a while, along with using the trigger monitor tool in MegaTunix to view the triggers watching for anything funky in the ignition triggering. They couldn't find anything out of the ordinary there. James drew out the wheel pattern and traced it out, and we did change the wheel decoder settings at one point to what he felt may be the better way to set it up, but it didn't change the 'slight misfire' issue for me. It continued.


They kept going back to something in the config/tune. We checked out dwell, played with it a bit, it didn't seem to be the issue for me. Though for someone else it could cause a similar issue. It was running around stoich at idle, mid 14's really. Richened it up a bit to high 13's, better but still missing some. Tried 16deg ign adv, also 12, and 10, no change. Found something that should have been set long ago differently-- I had originally grabbed someone's map off the forums for the base settings back probably 6 months ago-- it was set to 1 squirt, simultaneous. That often isn't best on most motors-- generally 2 squirts alternating is better. We set it to that, and immediately the issue went away on my car. Purred like a kitten.

Looking at the 2 .msq's I recieved (yep, that's it) in response to this thread-- one of them had a REQ_FUEL setting in the recommended range of 6-15ms, and the other didn't. This can make idling a challenge if you're outside of this range. Both users DID have 2 squirts, alternating. So immediately I knew more experimenting was going to be needed. James and Ken are pretty confident this isn't a code issue-- and pretty confident it is a tuning issue, but I know they do want to see it resolved for people. What we found in VA was that going to 2 squirts alternating fixed it for me. The only other change we made was the wheel decoder settings, which I'll be glad to share. That could be related too though in my case it didn't seem to make a difference. On paper it sortof made sense why it could play a role, and in an intermittent sort of fashion like had been reported, and if that's it the new settings should resolve it.

Going further on the theory that worked for me. Different setups are going to have slightly different needs. If you're running different injectors, with different opening times (real opening times, not what's been guesstimated into MT), as well as vastly different size meaning vastly different REQ_FUEL numbers you're just going to have to play with this a bit. The theory is that in a batch/bank fire setup like this, particularly with 1 squirt simultaneous, you pretty much are always running two cylinders potentially lean as the fuel is firing at the same time on all 4 cylinders, but only 2 are close enough to firing time (at idle speed) to get good even fuel atomization in the cylinders-- the other two cylinders, especially the last one in the cycle, is getting fuel that's had time to 'settle' a bit, isn't as well atomized, etc.

But the stock ECU idles just fine right? Why not the MS? Well... it will idle just fine. What we have here I believe is something in the config/tune. On a stock car it's easy to nail-- my car now idles like stock. Purrs like a kitten. Every time. AFRs are now in the low 13's at idle once warmed up.

I leaned it back out to 14.7:1 and it started missing again. Hmmm... that's strange. No other changes, still 2 squirts alternating. So what the heck is the stock ECU commanding at idle-- well that's easy to check. 10 minutes later the stock ECU is in the car, I fire it up, let it settle in fully warmed up, and it's idling at 13.2-13.3:1. Rev the crap out of it, let it settle.... right back to 13.2-13.4:1 range. Leave and come back 10 minutes later (letting it idle), still at 13.2-13.4:1. Purring like a kitten.

The stock computer commands low 13's at idle too. My MR2 on the stock computer liked low 13's. Talking to Justin who works here now but is an ex-Subaru tech he said Subarus on the stock ECU idle in the mid-13's. This is pretty common as I've seen it repeatedly on different makes of cars. Engines make just a bit more power there than at 14.7:1, it makes for a smoother idle, which also makes for better recovery in the event of a stumble for any reason, or increased load, etc. All of which overall contribute to the bottom line of a smoother idle....

My belief-- after spending a good bit of time on this and working with the coders and hardware designers themselves on it (I should have taken a picture, Ken, James, Bruce, DieselGeek, Peter Florance, Dave (megatunix), myself, and several others all standing around this car with a scope, two timing lights, and multiple laptops.) is this--- we just need to tune a bit different here. 2 squirts alternating (for some 4 may work better), I'll get you the new wheel decoder settings in case that does play a role though it didn't seem to for us, and richen your mixture to the 13.2-13.3:1 range at idle once the car is warm.

As for the wheel decoder settings-- pic below-- make sure to check your base timing after changing this just in case it moves!

http://www.diyautotune.com/images/ca...eel_decode.gif

So after all I was correct - a motor needs a richer mixture at idle to be smooth IF (and this is a big IF) the A/F ratios are kept constant.

Now, I know this because I've done a lot of experimenting and a rich idle will always be smooth, especially since there is so much more vacuum at idle causing exhaust gas scavenging. Here's another tip, wire your injectors in the way that they come up in the bores. In other words, put 1,4 and 2,3 on the same bank, not like the stock configuration which follows the firing order. If you can, run as many squirts alternating as possible without having too small of a PW as well.

Stock ECU does another trick. The engineers figured out that if you oscillate the PW and timing at a certain rate while always starting injection at the same point in the cycle, it will maintain a smooth idle. The MS can start at more than a single point in the cycle causing for these erratic sometimes it works sometimes it doesn't cases.

This was programmed by Mazda and this is why a stock ECU keeps a steady, stoich idle. In order to get this right on the MS, a lot of experimentation with the feedback sensor must be done or some way of oscillating the PW in open loop. Furthermore, you'd have to start injection at the same point in the cycle all the time - can't specify this in MS.

Easiest solution is to just stick to a richer idle. It will always be smooth. I wanted to see what you guys come up with, but when I brought this up, I did it with experience behind it and lots of thinking. I work as a controls engineer in the automotive field, so I do have an idea of what's going on ;)

Jerry, I assume the guy whp's Req_Fuel is out of scope is me :( I'm running RX7 550cc injectors that I had cleaned and flowtested. they actually flow 600cc. A little more than I was expecting :(

I guess that's half my problem then!