Hey guys I started my car up today to try to do some tuning to the bottom of my map, but it won't stay running. Though I'd post here to see if you have any ideas.

Engine:
'88 Turbo - All emmisions recently removed, last night I took the Air Supply Valve (from the back of the UIM) off.

Megasquirt:
MS2 Extra on the 3.0 board (I believe? It's been a year since I played with it)

Yesterday it was running great after I fixed my injector wiring (Now ALL of the wires in my loom are soldered and shrink wrapped).

Today I started it up and it died after ~1 minute of running. Started it back up, same thing. Brought the idle up, same thing.
Did a "Trigger logger" and kept track of the sync loss count, everything seemed perfectly normal (triggers looked like a nice pattern, only lost sync when the engine came to a complete stop).
Did a data log and everything looks perfectly normal to me until it starts dying. It all the sudden just goes lean and the car dies.

If you look at the data log, you'll see that the pulse width stays constant right up until it dies, so I thought maybe there was a kink in one of the new fuel lines I put in, but I can't find anything.

When I bumped the idle up (idle screw on the TB), I started it and it ran fine for a minute or two, and then right as I let out the clutch (still in neutral) it died almost instantly, and from the same reason.

I thought maybe there was a bin in the VE table right above where it was that when the clutch was let out, the load was increased just enough to bring it into that bin, but all of the bins above either of those idle speeds are quite a bit richer than the ones below..

I just don't get it. Any ideas? I've attached two data logs: one with regular idle speed, one with the high idle I mentioned; a trigger logger excel file; and my .msq file.

 

So to try to counteract the problem, I made all of the bins around the current idle a lot richer (increased VE value by 10), and it's staying running a bit better now... although it's running at 11-12:1 AFR. If I try to take any fuel away to bring it into the low 13's, it'll lean right out and die. I'm a bit confused of what's happening as yesterday I could bring it up to the high 13's, if not going into 14:1, and it was running fine at about 1300 rpm which is what i have the idle screw at to try to tune my map at that location.

Maybe it's my lack of tuning experience, but if I'm at say 12.8:1 AFR, and decrease all 4 bins surrounding the current location by a value of 1 or 2, why would it completely lean out (ie 16:1 and up until it dies)?

 

So to try to counteract the problem, I made all of the bins around the current idle a lot richer (increased VE value by 10), and it's staying running a bit better now... although it's running at 11-12:1 AFR. If I try to take any fuel away to bring it into the low 13's, it'll lean right out and die. I'm a bit confused of what's happening as yesterday I could bring it up to the high 13's, if not going into 14:1, and it was running fine at about 1300 rpm which is what i have the idle screw at to try to tune my map at that location.

Maybe it's my lack of tuning experience, but if I'm at say 12.8:1 AFR, and decrease all 4 bins surrounding the current location by a value of 1 or 2, why would it completely lean out (ie 16:1 and up until it dies)?

 

I've experienced the same thing... my engine refused to idle any leaner than a 12 AFR... never figured it out. Anything leaner than that and it would just die.

 

I actually don't see anything dreadfully wrong with that datalog. There's a lot of noise on your IAT sensor but that shouldn't cause your problem.

Don't expect to be idling leaner than 13.0 on an engine that hasn't fully warmed up.

Additionally if you don't have a BAC valve, you'll need to tune the idle higher and richer than you would with the valve.

Make sure to rule out a mechanical problem. For example, is the fuel pump sucking air? Filter clogged?

Maybe someone else will see something in the datalogs that I'm not. Your MSQ also looks reasonable for a work in progress and I don't see anything obvious that would kill the engine.

Wideband readings can be confusing sometimes. For example, if the engine is misfiring, it will read leaner. However I'd lean towards a mechanical problem in your case. Fuel pressure, lack of fuel, etc.

 

Wow I didn't know that misfiring would show lean.... I wonder if I could be going TOO rich is some spots, which is what's causing it to misfire when I rev it....

Could someone take a look at this new datalog and tell me in your opinion, if you think that it's misfiring because it's too lean, or too rich? It seems to be adding up that I just need to keep adding fuel, but my intuition is telling me that 70-80% VE at high vacuum and 5000 rpm, and needing an AFR of 11.5:1 just to stop it from sputtering, just seems like too much fuel.

In the log, you'll see that I revved it up twice, but both times it would get to a certain rpm and just sputter and not rev past that point.. my wideband showed a lean spike, so I made the bins that it was stuck at and the ones to the right of them all richer by a value of 5. The next two times, it revved pretty cleanly (so that would make sense that it was too lean before).
I've also attached the .msq file that I ended up with, to show you how much fuel I'm giving it. (just look at the bins under 59 %fuelload)

 

Yes, 80% VE at light load anywhere is too much fuel. Typically in those high RPM areas you'd see around 40%. Depends of course on the car.

Can you trust the wideband? Is it properly calibrated and is TunerStudio/MegaSquirt set up with the correct calibrations?

 

My wideband WAS definitely working properly last year, and seemed to be the day before this started happening... so unless something changed in the last couple days (I do have it mounted at a bad angle).

The one thing that makes me feel like my wideband is working is that it is showing me that there's too much fuel also; my target AFR for that section is 14:1, and my wideband is showing 11:1. I would have tuned leaner, but the only way I could get the engine to get into those bins is to make it richer, otherwise it would just studder like there was a rev limit at those cells.

In a "normal" car, would it be normal that the car doesn't like to rev in these cells? The only time I really should be getting into those is when it's SLOWLY free revving (as I've been doing) or decelerating down a hill.

 

Another thing that might be worth noting...
It is next to impossible to rev this thing without killing it. Literally if you move the throttle ANY faster than the slowest you can (on the gas pedal or with your hand on the throttle) it'll just lean out and stop running. It's almost like the throttle is a fuel cut :P

Now I know that quick throttle movements will create this effect when the pressure suddenly changes causing the fuel to stick to the walls. Hence why we have Acceleration Enrichment.
I currently do not have AE set up, while trying to tune the map, but this seems a little bit extreme to me.

I've read around about this problem, and people say that you can help fix this by adding fuel to the bins directly above your idle/where your having this problem. I've tried this, watching where it goes when I try to rev it, and then making those bins rich. I've brought them all the way up into 70's (%VE), yet this problem is still pretty much just as bad.

Last year when I thought I had my fuel map ok, I tried working with AE, and had it adding a HUGE amount of fuel, which made it better to the point where it would drop a couple hundred RPM, then rev up.

 

So the car was running at some point and just started this behavior?

It sounds like a mechanical problem. Bad injector, extremely low fuel pressure, or something like that. If it has been sitting at all, suspect injectors.

 

Bad injectors could be something to check I guess, it DID sit all winter. I know the fuel pressure is good.
They all go click when i put it in output test mode, should I maybe take the injectors out, put some fuel pressure on them, put it in output test mode, and just make sure that they're spraying a constant amount each time? Or if I had a sticky injector would it stick EVERY time?

 

In my experience injectors are the #1 cause of weird behavior like this. Yes, they can stick one day then work perfectly fine the next, or take twice the duty cycle to inject the same fuel, or only open half way, or stick open, etc. And after sitting for the winter then that would be tops on my list to check. Assuming everything worked fine before storage, that is. Also fuel sock and pump.

Pull the spark plugs and see if you can determine which side of the engine isn't firing properly.

 

Good idea.
I'll pull the plugs and see if I can find any clues. I guess I could have seen this coming because these injectors have sat for extended periods of time at least twice, quite a while with the previous owner, as well as over winter with me.

I am almost certain it's not the fuel pump or sock, as those are only ~2 years old, I put in a brand new walbro 255 when I did my engine swap, and I took a video of my fuel pressure gauge as I ran the car and it was steady. But... if it turns out not to be injectors, that will be the next place to look I guess.

 

Hey guys, sorry it took me so long to get back here, got really distracted by another one of my hobbies.

I have returned with some useful information though.
I believe with the test I've done today, I can rule out 3 things: Sticky or sub-par injectors, injectors not receiving a proper signal, and inconstant fuel pressure.
Here's what I did:

I strapped my injectors to my primary fuel rail, took my fuel rail out of position, and put a few containers (which i marked in small cc increments) under the injectors. I put the MegaSquirt in output test mode, turned my fuel pump on, and set the injector parameters to:
Output Interval (ms) = 20
Pulsewidth (ms) = 4, 6, and 10
I calculated how many total injections I would need for each pulsewidth in order to have e final volume of exactly 200cc. I put these numbers in and compared the volume I got to 200.

I did these 3 different tests for both injectors (6 tests in total). The biggest discrepancy I saw was 3.5% (7cc out on 200cc).

When you take into account the fact that my measurements were most likely not exact... that's pretty dang close in my opinion. My fuel pressure also stayed perfect the whole time at right around 47psi.

 

While that test does help, it doesn't prove the injectors work. 200 total injections is less than 10 seconds of operation (rough guessing math).

At this point I don't have any other suggestions off hand. If the ECU is still receiving sync, and the injectors are firing, and there is adequate fuel pressure, the engine just shouldn't stop. Some other external factor is causing this. Have you been able to verify fuel pressure is appropriate even after the engine stops? The pump should still run for a second or two. If there isn't fuel pressure, that's why it's not running.

 

Sorry I should have been more specific, the total number of injections for the tests were 2600, 3260, and 6520 injections, for each injector.

Anyways, you've given me some insight and some ideas to try, so thanks for your help so far. I'll try to get the car out (probably have to take it off the property) some day soon so that I can actually get some more time with it running.

I'll post some updates when that time comes, and if I come across one, my solution.

 

Not a solution, just a log of some things I've done that may be helpful:

I've been fighting this "only idles below 12:1afr problem" since day one. I'm in the process of actually tackling it, and I believe it to be the fuel injectors. Here's why:

MS2, V3. Lastest F/W as of 7/19/13.
4x800cc Venom racing fuel injectors (most likely the problem; reman'ed, re-pintle'd injectors rarely function correctly, I'm learning). Strange, since I've flow tested them on a Bosch flow testing/cleaning tool multiple times, and they are all within 2% of each other; excellent. But do they ALWAYS work? That, I cannot verify. Low duty cycles produce excellent results. High duty cycles produce excellent results.

Verified timing. It's always exactly where it should be, T1 is always on the red mark with the squirt set to fixed timing mode. T2 is always 180degrees out (I marked the pulley for this). Leading alternates between the two marks (-5 and 180-degree self-made mark), as it should. With table mode enabled, these results are repeated, as they always match the table's values.

S4 13bt engine, center iron port-matched to secondary ports (this is what I originally believed my problem was), exhaust ports matched to sleeves, ebay turbo header, holset hx35w turbo, FC ignition AEM wideband.

Essentially, the engine will not idle above a 12:1ar. I could get a smooth idle if I knocked the timing to 0degreese advance in the idling bins. Low-load scenarios are extremely bucky above 12:1 as well, but could not be remedied with timing modifications. I have always had pronounced problems with staged injection hesitation, especially when using transition. If I set the staging params to engage at, say, 2500rpm, as soon as the staging kicks in, it cuts both rotors, going lean, as though the secondaries are not working at all. In truth, they are being told to operate exactly as they should; exactly the same PWM as the primaries, exactly the same waveform amplitude; both are wonderful little 13v square waves. If I richen the secondary injectors (by lying and saying they're 50% the size of the primaries), it'll actually run okay and transition smoothly, but the PWM scaling ratio is thrown off at higher duty cycles due to this (runs crazy rich to the point of cutting out under loads).

Now, I've already done one test: On my daily driver, I made an adapter harness from an old FC computer so I could easily switch between the stock computer and this MS2. I added the same AEM wideband to it. This car is a S4 T2 with a bone-stock S5 T2 engine with no emissions.

Typically, the stock computer will run the car at around 10:1 when cold at idle, 13.5:1 when warm, no loads. With the A/C on, it idles between 14:1 and 16:1 (15.5:1 nominal). I can achieve these numbers with the MS2 easily. Staged injection works flawlessly no matter what I do or where I set it, as long as it doesn't over-drive the primary injectors first.

I've heard good things about the disc-type Siemens Deka injectors; $200 for a set of 4 new ones. I'll buy a set sometime soon (too many projects/financial obligations to plan around at the moment!), and I'll post an update as soon as I do.

 

Question....did you plumb the fresh air line to the primary injector bleeds and are the bleeds functional?

Without them, the engine won't be capable of idling properly much leaner than 12-12.5:1.

Air velocity in the port runner is low at idle and the injectors are spraying 90 degrees into the air stream. The air jets from the bleeds greatly help and Mazda has continued to improve them year after year. In fact, the RX-8 has little nozzles coming from the bottom of the runners which blow onto the injectors.