Misfire
 

I've been chasing this misfire problem for awhile now. I've gone back and forth between stock and MS controlled spark several times now, and it only happens with the squirt. No problems with stock ignition. The misfire is accompanied by a quick flutter down on the tach, usually a drop of 200-500 rpm. In an average 20 minute drive I'll get it maybe once or twice and usually more at the beginning when colder.

So far I double checked the wiring and looks like I have good connections although I only stripped the wires and twisted them together(1/2" of wire tighly twisted). Ignitiong wire is grounded on the MS side only. I added 3 more smaller ground wires to the DB37 that all go to the same stock spot on the block. I also cleaned up the grounds across the whole car to completely rule out a ground problem. I've tried a running dwell between 3.5 and 5 with no change. Took apart the MS today also to double check insides and all soldering looks good. Also checked the resistors that were added in the DIYautotune mods and again all check out.

Over the past few weeks I've cought it several times in datalogs but all that shows up is a small lean spike on the WB. Nothing else indicates anything happening and I've checked every other variable that gets logged.

Anyone have any ideas what else to check? I'd hate to not be able to use MS for ignition control but I might end up that way :td:

Could the ignitor be going out? And it is only showing on the system when under heavy laod obviusly you dont run the car as hard on stock ecu with boost etc as you do with MS right?

It usually happens during cruise under lighter load (probably because I just happen to be there most while driving), but has happend under boost once and occasional high load. Seems like it just misses whenever it feels like it. I also run just as hard on the stock ecu as I'm only at the wastegate 4.5 psi right now and it runs smooth as well.... stock :)

You really need to crimp your connections or solder them in order to guarantee that they are a good connection. Rule out all the variables you can.
Are you running the newer decoder settings that DIY autotune determined were better than what we were all using before? That seemed to help a few people.

RPM holds strong in the datalogs so I figured the connections were good. Hope my board is screwed up =/

And yea I'm running the updated settings. I check back on their site all the time to see if things are updated as they seem to always be trying to tweak the settings to get it just right.

I'll try soldering after work tomorrow if I have time just to make sure

I had the EXACT same problem. I added ground wires (7 total) from the DB37 to the block, and added a 6 ga ground from the battery directly to the block. This improved the situation, but it still missed occasionally. Finally I replaced the coils and the problem was solved. I found a coil from a Hyundai that fit, and I reused all of the stock coil wiring. That was a few months ago, and it has run great up to 10 psi.

I'd bet on the grounds myself. The situation you're describing is one I've seen to the letter, and corrected with additional grounding, never to return.

Individual grounds back to the same point on the block that the stock ECU grounds at (assuming you're using the stock ECUs grounds). Bottom line is make sure they all go to the same place, and the block is where you want them.

Could be dirty grounding point as well, a few swipes of emry cloth wouldn't hurt.

Do you think it makes much difference if I dont have the screws on the DB37 connector to hold it to MS? Does the extra pressure help?

They're really just they're to keep the harness from coming loose-- I'd recommend using them, but I wouldn't think it's related to your misfires...

So I think I am ahving the same problem. Very occasionly it misses and always seems to be right when i press the gas at a green light and am about to start going. Even more so if I am at a hill. I have atached a datalog and you can see at 31.523 seconds my rpms drop then go abck up quickly. Is that a misfire? I plan on redoing my grounds on this weekend.

mason.gmu.edu/~lsnead/drivehome.xls

In your log that mark is labled as "reset 1". Doesn't look like it comes back, so not sure about that. Other than that it looks pretty clean. It doesn't look like you are short on accel enrichment, but it would be an easy thing to increase and play with.

So that dosent look like a misfire? Any clue as to what it is or what causes it?

Yes sir-- that is a reset, as indicated by the red line in MegaLogViewer. Also if you highlight that instant in the log viewer, and look at the text near the bottom left, you'll see it marked as a reset as well.

Most often this is caused by either a noisy 12v power supply or inadequate grounds. It can also be caused by high current noise effecting the low-voltage signal lines. Check out this section of the MegaTune Manual for more info: http://www.megasquirt.info/megatune.htm#mt

It's a few lines down in this section ^^

You ever use a car audio power filter jerry?

No I haven't-- if the source is noisy 12v then I can see how it would help. If it's inadequate grounding it won't. Better grounding will help in either situation.

I understand, just curious if you'd used one. I'll use that as a last resort, still fabbing my boomslang.

Well I added 5 more wires to the block total of 8 plus stock two :eek:. I also took a wire brush to the ground strap on the passenger side of the motor, the stock ecu ground bolts and all parts they are connected to and the battery posts. Also added an extra 8ga wire from the battery to the power plant frame and cleaned that spot up as well.

Still get the misfire. :cry:

Are there any checks I can do to the MS board itself?

Where does your car misfire at? Idle? Light throttle/cruise? WOT? It's usually one of these three, and not all three, and each can have a different cause.

What are the indications of a misfire? Do you feel it? Do you see the tach drop out?

(Every question is here for a specific reason)

I've never noticed it happen at idle. Usually happens during cruise but has happened a couple times a WOT. Seems pretty random though.

Indications of the misfire are a small spike on the WB and a quick drop on the tach and I definatly feel it. Sometimes its small with a little flutter of the tach and barely feel it, sometimes bucks the car a little with a big drop on the tach of about 500 or so.

Checked over many datalogs and see no other sign on any other channels either

Here's what you can do to check out your car's grounds:

A) At the back of the fuel rail, hanging off the hook near the CAS, are two wires connected to a bracket. Just under that connected to a different bracket are two more wires. All 4 wires are either black or black with a green stripe. These four wires are your ECU grounds. They MUST have a very good connection back to the block/head. That means the bolt/stud/nut they are threaded onto, and any terminal the wires themselves use, must be clean and contaminant free (no oil, oxidation, rust, etc.). Make sure you have a good clean metal to metal connection.
B) Then trace these back the next step towards the block/head. Make sure the connection of the bracket itself is clean metal to metal all the way to the block or head. Clean it up and make sure there's a good connection.
C) Now, to make sure the battery is securely grounded to the engine block. In an ideal world, the engine block would be directly grounded to the battery negative with a nice fat cable. That's not the case on the Miata. The battery is grounded to the chassis, and to the drivetrain brace. Pop your trunk, and make sure the connection from the battery negative to the chassis (right by the battery terminal) is nice and sound. Clean it up if needed. Then you'll notice that ground continues with a fat black/yellow wire, on down through the floor of the trunk. If you lift the car up and look at the right side (pass) of the brace that runs from the rear end to the tranny you'll see where this large black/yellow ground wire hits the drivetrain brace. That's the last connection I can think of to clean up. Other than that it's the drivetrain itself (brace, transmission, and engine) that are the ground. So instead of the battery being grounded directly to the block it's grounded to the rear of the drivetrain on a brace and has to 'go the distance'. Not too bad if 'the distance' is clean and clear of corrosion. Worst case scenario you could run an equal sized ground from that point on the brace on up to a large bolt somewhere on the engine block.

The goal here is to have a solid, clean, unobstructed ground from the battery to the block, and then solid, clean, unobstructed grounds from the block to the ECU. Normally you wouldn't have to mess with any of this with the MSPNP but I'm willing to bet that you find something along the way here that is likely the source of weak grounding. And if so it may make the car ultimately run better period as weak grounding to the block can cause weak spark and all kinds of issues.


Other possibilities--
-----------------------
The only other possibilities I can dream up are:

A) If you've got some sort of aftermarket equipment that's causing alot of noise and interfering with the low voltage (sensor) signals. The most likely culprit would be an ignition box of some sort, but some fans have even been known to cause problems.
B) Another thought is a bad/dead/dying alternator can introduce noise into the system.
C) Last possibility I can come up with is it could have something to do with your USB adapter, an incompatible adapter or driver possibly. There are some out there that don't work very well with MegaTune, and some that don't, and yes they can cause resets, and communications problems including scrambled tables and such. I recommend the USB-2920 that we offer as it seems to be problem free. Note that if you're having problems (resets, tach dropouts, etc) when the USB adapter is not connected you can rule this out. A tach dropout you could see, a reset you can't really see unless you're logging though, which you need a serial connection for. You could use a direct serial connection if possible to test this, or try a different model USB adapter, preferably the USB-2920. NOTE-- item C here, the USB adapter, could only be a factor if the problem is ONLY occurring when the USB adapter is plugged into the MS.

I also forgot to unplug my coils when flashing the MS to the extra code. Heard thats supposed to fry things up a little.

Originally thought my ignitor went on me but I have no problems running spark on the stock ecu.

Also just got my 460cc's in so Im nice and motivated to get this thing fixed now :)

That is, of course, assuming your ignition settings are correct, including dwell, and that your ignitor isn't giving up the ghost. But unless you've overheated it at some point in the past I wouldn't suspect the ignitor.

Heh... we posted that at the same time, and it does at least bring the ignitor into question. But it's not a definite-- I think you probably have a list of things to consider now ;).

I'm gonna run a nice fat wire right to the block then just to make absolutely sure the grounds are good.

When I soldered the ground wires on the DB37 they asked for it to be connected to 3 pins. So I laid it across three pins and layed on a lot of solder. Thinkin about it now the wire itself doesnt make very much contact to the pins. It IS completely covered in solder at least. But does solder conduct well enought?

Ever heard of reflashing the MS-Extra code to fix a problem?

You should redo those grounds on the DB37. Remove the mess you have on there now. If it was me I would solder 6 wires from 13-19. I think 14 is for a couple of the sensors. Just one ground wire attached to 3 pins is not enough you will overload that wire and it could cause serious problems.
So i vote, clean up the mess, and solder 6 individual 18 gauge wires to pins 13-19 and attach them to the wires you've mounted on the block already. Make them all connect to the same place.

I thought you said there were 8 grounds from the block to the MS? If they don't go to the MS where do they go?

Reflashing code can fix a corrupt firmware/map problem, feel free to try it, but I'm not optimistic.

I have a feeling he has that one wire connected to 8 somehow.
I think i read in the megamanual last week that the way MS works it doesn't draw much power but it does sink alot of power into the grounds and that is why they are so important. Pretty sure I read that pin 14 is specifically for the CLT, IAT and some other sensor, though i could be mistaken.

If so that won't do you much good. They need to be individually run back to their own pins on the DB37 for maximum benefit. It wouldn't hurt so bad if the pins were tied together at the DB37 probably, though ideally they wouldn't be. Definitely not ideal to have them joining into fewer wires that then come into the DB37.

Sorry I mean the 8ga wire is from the battery directly to the block. And I actually have 8 individual wires from the MS to the block. I took 1 small piece of wire completely stripped and laid across all ground pins and soldered quite generously. Then laid wires across it and soldered them.

It was never inteded to be that way but I messed it up when I did the first two and accidently filled in the almost all ground pins with solder.

Then it's more than likely something else as mentioned above-- noise or ignitor...

it's the ignitor.

anyone wanna take bets?

I had first thought the ignitor was the problem as well but when I switched back to stock ignition and the misfire went away I started looking elsewhere.

How much dwell are you running?

DIYAutoTune spark mods or Kags mods?

DIY mods, non inverted. Currently at 5ms running now, .5 discharge time. I found I got lots of tiny misfires when I went as low as 3.5, anything higher and it went away. Too bad the big misfire didnt too.

Thanks for the info, just confirming there wasn't a possiblity of something going on there, but that sounds right.

Hate to keep repeating myself, so instead I'll quote ;)

Thanks a lot for all your help. I have a camping trip this weekend so I wont get a chance to check out too much more. When I get back I'll mess around with it and let you know if I figure it out.

Thanks again :bigtu:

No problem-- FYI I'll be out of town all next week, gotta take a vacation sometime ;) -- But Matt will be in the office, you can always shoot an email to websales@diyautotune.com and he's quick to answer during the week. I'll see about having him drop by here as well.

During my whole trip out I ran stock ignition and never had any problems at all. So that confirms its a MS related problem.

I have gone over all grounds and the battery terminals and everything is shiny clean. I've added an 8ga wire from the battery negative terminal straight to the bottom of the intake manifold. I also checked voltage at the battery with the car off, running and with various things on and off and everything looks good. 12v off, steady 14 on. Small spike of about 1 volt when I kick something on like the headlights that quickly settles back to 14. Then I tried disconnecting everything aftermarket which includes an alarm and stereo/amp.

After all this and I still have the misfire. My next step is to check over the MS board very thoroughly. I've already tripple checked all diodes, capacitors and anything with polarity. Now I'll make sure all resistors are good.

Does anyone know what components make up the ignition circuit?

ECU-2-Ignitor-2-Coils

Kind of out of off the wall but did you clean your board after soldering it together to get the flux off? The megamanual recommends isopropyl alcohol, i would use atleast 90% or better. Sometime you have to scrap the flux off with a little wooden toothpick. I used alcohol and a toothbrush. Remove the MS chip when doing it and rinse with warm water once you have it all completed. Then let it dry completely.

It sounds like you need an aftermarket misfire system.

Yeah I did clean up the board decently. Didnt scrub very much but it looked clean. I'll do it again to be safe while I have it out to look everything over.

and uh, aftermarket misfire system :confused:

^ https://www.miataturbo.net/showthrea...market+misfire

Dont quite see how a misfire system would cure my misfire problem

It was a joke. You'll catch on, just let it marinate.

mmmmm.... marinate

Schematics are in the Megamanual so you can check out any specific circuit--

V3 Board Schematics Online

Or here's a pdf download, easier to search for signal tags to follow them around: http://www.diyautotune.com/downloads...Schematics.pdf


You're probably pulling spark from the LED's like most all of us right? You'd want to look at those circuits... There's not much to them.

I'm honestly not real optimistic that this is where your problem lies, but it's certainly a possibility and worth looking into...

I checked the board over and I found a lot of flux between all the LEDs. Chipped it all off easily. I must have gone back and resoldered them after cleaning or something. I really hope that was it as I cant find anything else wrong.

Ok Ben you win, I'll try swappin the ignitor.