Strange intermittent cylinder cutout on a Megasquirted TR6

Occasionally after the car is warmed up, a cylinder or two will go on vacation. The AFR goes up to ~16 (so running lean), and it revs like one of the spark plug leads came loose. I can usually reproduce to problem after a drive (or an autocross run) by shutting down the car and restarting after a few minutes (heat soak?). But it's also happened while driving normally on the street and during an autocross run.

I've pulled injector plugs to determine which cylinder is missing. Oddly, this has occurred on 4, 5, 6 and (just today) on 2. So it isn't simply one bad injector (and I already tried replacing #4). When the miss happened this morning, I verified that all of the injector plugs had a signal using an LED. What's really strange is that the injector on the cylinder that didn't change the way the engine ran when disconnected was still ticking when connected. Is it possible for an injector to make noise and still be stuck closed? Maybe it's just partially opening. It could be blocked, but I have a hard time believing a blockage in an injector could come and go like this.

The original injectors were Weber Picos. We swapped to the equivalent Edelbrock injectors to increase the flow rate (from 15 to 19 lbs/hr). If the Edelbrock injectors are cheap Chinese knock-offs, that might be my problem.

Sure it's not ignition? Are you still using the Lucas ignition system?

Ignore the O2 sensor if you have a miss. A misfire will send raw oxygen out the exhaust and the sensor will read wildly lean.

Log file?

Dr. Hess wrote: Sure it's not ignition? Are you still using the Lucas ignition system?

At this point, no. Except for the wiper motor, The Prince of Darkness has been exorcised from this car. The Megasquirt is triggering DIYautotune parts for the ignition. I've logged the crank trigger while this was happening and it was rock solid. I've also replaced the plugs and swapped a known good ignition wire to all cylinders one at a time, but the problem didn't go away. It could be the coils (wasted spark), but that should affect the paired cylinders (which is doesn't). The main reason I suspect fuel is that pulling one (and sometimes two) injector plugs doesn't effect the way the engine runs nor the AFR.

I had one log from when I was tuning the car that only shows the wideband going lean. I'll see if I can dig it up.

When it is missing, you can let it idle and it's missing, right? Do a compression test right then, starting with whatever cylinder isn't firing. Also, one of the pen-like LED ended spark testers is real handy to make sure a plug/wire is firing. I used one of those to track down a miss when I MS'ed Dr.Linda's Sportster. Thought it was ignition when it cut out, so riding down the road making it cut out, held the pen-like tool to the spark plug wires and it was flashing. Not spark. That one was actually low battery/charging system voltage not allowing the MS to sense the trigger event.

It's not as obvious at idle, but still noticeable. I'll verify ignition with a spark tester. I'll also have the compression gauge handy next time. If anything, looking at the spark plug may tell me something. Thanks.

Double post coils can fail in such a way that one end still gets spark, so don't rule out ignition just yet. Got some insulated spark plug wire pliers? Yank each wire off when it's missing and listen for the miss to get worse with each wire. You should also be able to hear the spark arcing when you hold the wire close to the plug.

I can't stress enough that it can be a trap to get hung up on air/fuel ratio readings when there is a misfire. Reading lean while misfiring just means that the O2 sensor works. The sensor is in the exhaust, it is not measuring air and fuel, it is inferring it based on exhaust gas oxygen content, so anything that adds oxygen to the exhaust (misfires, exhaust leaks at the manifold flange) will make it read incorrectly.

Knurled wrote: Double post coils can fail in such a way that one end still gets spark, so don't rule out ignition just yet.

I guess one advantage of a single coil is that when it goes, you know it. I'll work on diagnosing the ignition the next time this happens. Worst case, the coil pack was only $60 and we could always use a backup. At least I'm fairly certain it isn't the wiring/connectors or the coil driver.

My other Triumph is a 2000 Legend TT motorcycle. The stock coils are a known weak point, and I just upgraded all three over the winter because one was behaving much like this. I'd lose a cylinder, pull over, and it would fire right up after a minute.

JoeTR6 wrote: I had one log from when I was tuning the car that only shows the wideband going lean. I'll see if I can dig it up.

If you can post that log and an MSQ, I'll be glad to have a look at them.