jottinger22

Hi all,

I also posted this here: https://www.msextra.com/forums/viewt...?f=132&t=73171, but I noticed there was only one other person logged onto the entire site at the time and I'd kind of like to get my car running soon, as it is my daily.

Anyways,

About a month ago I installed my DIYPNP in my '91 Miata (quarantine project) and aside from one circuit that I forgot to wire initially, I have had no issues. That is, until Thursday afternoon driving home from work. Down here in NC we were getting a ton of rain from a nearby tropical storm, and suddenly my car lost power and shut off as I was slowing down approaching a red light. Restarting it, it would barely idle and I limped it into the entrance of a neighborhood, and after a few tries at some different things, I called AAA and got a tow home. The spark plugs on cylinders 2 and 4 were both dripping with fuel. This was troubling because the wasted spark system is designed such that 1&4 share spark, as well as 2&3. After checking the plugs, I wondered if my megasquirt had gotten wet in the torrential downpour, but I checked it and it was dry. I opened it up and didn't see any glaring issues. Then I plugged in the Stock ECU and the misfires persisted.

This weekend I went through all of the ignition components, swapped in working plugs, plug wires, swapped and bench-tested the coil pack and even used a spark tester to verify that I have spark on all cylinders. I swapped in a working Cam Angle Sensor to see if that had failed, but all of these components passed the test. A cold compression test yielded 155 or more on all cylinders. When I do try get it to sort of run on two cylinders, extra fuel drips out of my exhaust where my headers meet my cat. The headers which connect to 1 and 3 both warm-up normally, while 2 and 4 are completely cold.

As a last-ditch effort this afternoon, with no other clear path, I tried plugging in the stock ECU one more time, and finally I had all four cylinders firing. This is enough for me to say that my DIYPNP has some type of electrical failure. I'm guessing the engine was too flooded to start when I first tried swapping in the stock ECU. Aother start and it idled perfectly at 1000rpm for almost a minute until I shut it off, even with no afm, a wideband (which doesn't seem to be working), and a vtps. To verify that the ECU fixed it, I plugged in the megasquirt again- more misfiring on 2 and 4. Then I swapped back to stock- idling on all four. Again, it ran fine for almost a month before this problem occurred rather suddenly

My best guess as to what is happening is that only half of the ignition signals are getting received by the coil pack. The firing order is 1-3-4-2, so something is causing the 4 and 2 signal to either not get triggered or not reach the coil pack. I have no way (that I can think of) of actually testing this, but it is my best hypothesis given that the cylinders have air and fuel, the timing and compression are good, and the spark tester shows that there is spark. The problem is that there may not be spark every time that there should be.

Long story short, where do I even start to look for an issue? This is my first electronics project, and although I can use a soldering iron and a multimeter with elementary school level precision and I have a basic calculus-based understanding of simple RLC circuits, I have no clue what to even look for here.

Oh, and attached is a log of it running pretty terribly, in case that helps. Let me know if anyone has questions that can help you help me.

Thanks!

sixshooter

iTrader: (12)

I'm not on my computer to look at the tune but what do those two cylinders have in common? Do they share an injector driver? There are ignition test modes in TS and I'm not sure if there are also injector test modes (waiting for coffee to brew and can't brain yet).

Is it possible the O2 sensor failed or there was a misfire and the tune is set to add fuel to compensate if it thinks you are going lean? Many tunes are set that way. Could dump fuel if given the authority.

jottinger22

The O2 sensor is currently reading a full-lean condition all the time. Since there is a very strong gas smell with it "running", I interpreted this to mean that the air was not being used up in the combustion reaction and as it passed by the sensor it registered lean. However, after startup and running with the stock ECU with all four cylinders firing, it still reads this, although it may be actually running lean because of the various changes in sensors. I am doubtful that it is running at 18.6 afr though. It's also very possible that the sensor failed as a result of whatever my real problem is, where it was up to temperature and then doused in cold fuel, heat shocking the sensor. At the time that I recorded the log, almost immediately after the failure, it appears as though it was working, with the outputs fluctuating rather than showing one consistent reading. Either way, I will look into that "lean-fuel-dump" as you described.

SpartanSV

How confident are you that the misfiring is on 4 and 2? It would rule out a lot of things including the signal from ecu to coil packs. You're still wired for batch injection right?

You can also use output test mode to check each cylinder for spark and also test the injectors.

I believe six is referring to EGO correction. It wouldn't be a bad idea to shut it off right now but it isn't your main issue.

jottinger22

After running for only a few seconds, the headers coming from 1&3 became too hot to touch, while the headers coming off 2&4 were still room temperature. These cylinders had gasoline dripping from the plugs. If the problem were in the plugs, wires, or ignition output system, then the issue would present itself on 1&4 or 2&3, not 2&4. It can't be a bad CAS input, because then I wouldn't have any reading on the tach.