#5 not firing right

[Walt Fricke]

Scotty - for sure I'll check cam timing. With a big cam I'd not have thought it would turn over without interference, but easy enough to check. The one time I got an SC (964 cam) wrong, we discovered it while installing rockers. Waste spark system might help it run this way?

Fuel is over a year old, but why would one hole not fire when the rest do?

Out with the pair of dial indicators and Z blocks. I'll feel stupid, but relieved if this is what is ailing it.

[Lapkritis]

Originally Posted by Walt Fricke

Fuel is over a year old, but why would one hole not fire when the rest do?

Pulled a few cars from extended rests with aged gas... even without gumming it burns terribly. Uneven firing, stink, way down on power etc. No off the shelf additives help bring it back. Siphon out and burn off in a lawn mower or wife's Camry.

[WERK I]

Walt,
You're running batch injector firing, correct? I know you've swapped injectors on #5, but have you tried swapping injector leads on #5? Using the IR gun, you should see the bad cylinder shift from #5 to whichever cylinder injector lead you swapped. There is the possibility that you could have a bad injector driver in the Tec3r. If that is the case, the good news is you have 8 channels for injectors and your using 6 of them.
You will only get black, sooty exhaust if there's a PARTIAL burn. Going beyond that, there is no ignition, just wet plugs.
Can you operate #5 butterfly separately? I suspect if you open #5 butterfly all the way, with the engine idling, #5 will start firing.

[Walt Fricke]

Dave - I had been wondering if a driver (or something) was bad inside the TEC3R ECU. Sending the ECU back to Electromotive is part of the plan down the road if I don't turn up something. But here are reasons to think ECU failure is not the case.

1) The #2 and #5 injectors share a common driver. The wire from the ECU splits, so to speak, where the loom splits to serve the two sides. So if the driver is bad for 5, it should be bad for 2. #2 is firing normally.

2) I opened the butterfly with spark disabled and plugs out to make spinning the motor easier, and watched the injector fire, like the other injectors I reconnected for comparison. Unless a load affects something, it is hard to see how injector timing is part of this deal.

3) I inherited the loom, which is from the factory, and I inferred that the connection where the driver (ground) wires split was good. Testing with ohmmeter also said things were good. Alas, as a bad connection to just that injector would have explained all.

Electromotive's manual is pretty good. My engineer friend who installed Megasquirt on his 928 says it is better than the MS information. However, it is not perfect. For instance, you can chose from three different basic setups for a 6 cylinder dual plug engine on a basic setup screen. Unfortunately, the terminology on the screen is not the same as the terminology in the manual. Basic, staged, phased, sequential, TBI, maybe more terms. Grrrrrr.

By checking back on many saved maps, I see I am using the same setup as previously, when all worked just fine. Which is part of the frustration here - worked fine, no wiring changes, put rebuilt engine in, now this. Rebuild included new bearings, rebushed and refurbished rockers, valve job, and aluminum flywheel with Tilton 5.5" clutch. Shouldn't have affected anything other than perhaps fine tuning.

If I have it right, first off, sort of like waste spark, but not wasting anything, as I have it set up, each injector fires twice per 720 degrees of crank rotation - once timed around when intake opens, and once one revolution later (during ignition). And with cylinders firing at 360 degrees from each other, one driver, like one spark coil firing, works fine for both cylinders.

Not having a cam signal, I can't use full sequential, and think I don't want to anyway because I'd need huge injectors to get enough in with one squirt at high RPMs. So I think what I am running is called phased sequential. If it involves a "batch," it is batches of only two injectors times three.

The ECU has 8 injector drivers in two clusters. I am using three (which are siamesed) of four from one cluster, the 4th being for 8 cylinder motors. I confirmed I am using the proper three. Had I found something obviously wrong with injector firing as such, I would have looked into using the appropriate lead from the other cluster. The second cluster is for use with a two injector per cylinder system, which I don't have.

I can't easily switch injector leads around the way they are loomed. I could switch 4 to 5, and 5 to 6, but 6 would need a patch cable to get back to 4. And it would be a stretch to go the other way, omitting 6. I have some spare female connectors from old CIS looms, but nothing to simulate the male connector of an injector, so it would be quite a cobble.

However, the fact that I got fire in that hole when I squirted brake cleaner into the throat suggests there is something about the timing of at least that one injector which is off. Is brake cleaner so volatile that it could fire off from compression like a diesel, so spark being off timing also wouldn't prevent it having some effect?

So I need to stop thinking, and Monaco FI is over, so I need to be getting out to the trailer to check cam timing. If that is off on the right side, it could explain all. In any event, one basic parameter of a motor which needs rechecking.

Because it is in the trailer, which is stored in a rural area, I've been running it there (lots of people, new baby, etc. where I live, and open exhaust). Which means I haven't had an exhaust valve cover off yet, as I can't jack it way way up on one side in the trailer. Jack parallel to car = ripe for disaster. To back car out of trailer I need to start it (or enlist someone to push), and I am using the orange silicone or whatnot valve cover gaskets. These are the bee's knees, but the motor has to be stone cold when you remove the gaskets from the studs or they tear. Always some complication.

Thank you for trying to help me dope this out.

[rsscotty]

You are a sharp guy Walt. You not only understand what makes your engine & EFI tick, you didn't say who won the F1 race!!! I appreciate that. I'm watching it later.

[Walt Fricke]

I'm not feeling all that sharp right now.

The good news is that the cams were not installed with the right side 180 degrees out. That's also the bad news, as I'm not really getting closer to solving this.

Pulled the intake valve covers, watched what was happening. All six intakes started opening in their proper order. Set up the dial indicators (had to get new batteries for both) and took measurements for #1 and #4. Got basically same lift. Cam is GE80, and spec is 6mm.

When I was assembling the engine, I noticed that the Clewett crank pulley and 60 (-2) toothed wheel are pinned together (as I recall), so no manual setting of the wheel. This is the offset type using the distributor hole and stud to hold the sensor holder off to the left side. I wondered if this was an artifact of machining, as the guy who originally built this motor had a machine shop combine two different good halves into one good whole. That might offset things some?

But I also tested to see where true piston TDC was, and found it was not lining up with Z1 on the fan housing marking (or maybe it was with the case centerline). The pulley has three marks around Z1. The middle is marked Z1, and there is a mark at what I guess is 5 degrees off on each side of the marked Z1. I'm regretting not writing this down, but if I use the right side mark, which would be 5 degrees to the right, that is about where TDC is, and the reading is at 6mm more or less.

If I read the dial indicators at the marked Z1, I get about 6.5mm of lift at overlap.

However, the sensor, reading the trailing edge of the 11th tooth, is (if the right side mark is, indeed, TDC) is reading the "true" 12th tooth, which is a 6 degree change.

The TEC3r manual says you can use the software to adjust for mechanical tooth variances, and gives examples. In my case, if I figured this correctly, I'd want to tell it to use the 10th tooth, which would get things only one degree out (one degree "retarded", so to speak). Alas, when I use the software, it won't allow me to enter "10". The software listed range of adjustment is 11 (standard) to 18 (way retarded)!! I'm going to have to contact Electromotive about this.

I can easily enough pull 5 degrees of timing out of the ignition map, and try that. However, the toothed wheel and sensor also control the timing of the injectors. I'm not finding any discussion of how to adjust injector timing. All discussion about injectors is about a percentage change in pulse widths, not pulse initiation. Perhaps Electromotive can enlighten me there as well.

It has been suggested that old fuel may be part of this. About 5 gallons in the tank is year old 110 octane mixed about 75% of that with 25% of pump premium (that for another motor with higher CR). I added 5 gallons of 91 octane pump fuel last month when I first tried running this motor on the track. So it isn't all all that old, and in any event the fuel is good enough for the other 5 cylinders, and because the #5 injector sprays like the others visually, I can't see jellied old bad fuel somehow interfering with the #5 (which used to be the #6 and working, so it seems it can't be internal either). The injectors are on a fuel ring, so a single obstruction somewhere ought not to affect things either.

Here is something to ponder, though. Using the Unisyn I found that the idle air bleed for the misbehaving #5 didn't really have any effect at all on where the little plastic ball was in its tube. Naturally it didn't affect idle, because the cylinder wasn't firing. But the ball was sort of close to where I had adjusted all the others. It was a bit short of the middle mark, which I was using. So air was being sucked in at almost the rate it was on the others.

I would have thought that I'd see some effect - less vacuum with more bleed, and thus a lower ball? (or is it the other way? - in any case, some effect).

So that's where I am. I sure am glad I have those orange silicone gaskets - as long as the engine is cold, I can reuse them forever.

[WERK I]

Hey Walt,
This is certainly a challenging bug, wouldn't you say? Could you post a pic of the sensor and Clewett pulley on your engine?
The older saw-toothed pulleys were pinned together, but the newer Clewetts are billet/CNC machined as one piece. I'll check when I get home. Are you using the older saw-toothed pulley or the square-toothed pulley?

The Unisyn on #5 was operating properly. An engine is nothing more than one big pump and since the Unisyn is only monitoring the intake valve opening creating vacuum while the piston is retreating.

Have you retraced all the spark plug leads back to the DFU's?

[Walt Fricke]

Dave

This has baffled better minds than mine.

The Clewett pulley has the toothed wheel attached to it with, I think, four screws into the pulley. So the only misalignment there would be at least 90 degrees off. While my other Electromotive setups use the wheel on the back (rear of car) side, with a case centerline sensor location, and thus are easy to check for 11th tooth register, the distributor side Clewett mounting is not so obscured that I can't verify that I'm not 90 degrees off. It is just a bit harder to count teeth despite using magic marker to help.

In addition to the fact that the spark wire (and injector wire) loom/lengths are set up so you can't easily make mistakes, and the fact that none of that has changed since 2010 when first I ran this motor, and the fact that waste spark should mean if one is wrong, so is another, I have, indeed, retraced many times. And found nothing, alas.

I kind of figured the Unisyn is just measuring intake "suction," which ought not to be much, if at all, influenced by whether a cylinder is firing or not. So I thought that the idle air bleed screw ought to have the same effect on the Unisyn whether the hole is firing or not. And it doesn't.

So I need to bring my propane torch with a piece of hose attached to it to use propane, and not sprayed brake cleaner, as a better reverse sniffer tool to check for massive air leaks in the intake tract. Nothing suggests itself, but a better check is a better check.

Now if I can figure out whether I should tell the TEC3r that it should use tooth 10 or tooth 12 in order to retard the timing of all events (spark and injection initiation) by 6 degrees, I can see if that brings #5 to life. I think (from reading the manual, which isn't quite as clear as I would wish on this point, it is 10 I want for retard, but I may have that backward.

And I will reinsert the piston stop in a spark plug hole to recheck where TDC really is on this setup. One theory I have come up with is that the right bank is right on the ragged edge of something, and since #5 has the best flow (straight, not curved, manifold) it is over that edge. And that a small cam setting variation for the left bank has that side on the good side of ragged.

[WERK I]

Have you confirmed that plugs 5A and 5B are firing? If you have a clamp style timing light, you can at least narrow the scope of your problem to ignition or fueling. Just don't know if the 6 degree shift in events could make any kind of difference for cylinder #5. That is more in line with improving engine efficiency than whether a cylinder fires or not.

Now if you have an injector firing AFTER an ignition event, that could be a problem. But since the injectors are fired phase sequential. that can't be it. #2 is firing both ignition and fueling in sequence confirm by your IR gun.

Is the trigger pulley a transplant from the 2.8LS engine? I'm trying to figure why the tooth-offset is different than before. Could earlier cranks have a different woodruff key offsets than later cranks?

[cstreit]

Those coils do go bad. Id replace the offending one and try it. I had similar symptoms where one cylinder would randomly not fire and then BaNG would fire strangely.

[Walt Fricke]

I took a spare (unused, been sitting trailer for 10 or more years just in case) coil. Replaced the upper 2/5 coil. No change. Put old 2/5 coil back, put spare in lower 2/5. No change. I can dig out my DFX-1 plug only crank fire, and keep substituting coils off of it, but I don't hold out much hope there.

In addition to just pulling the plug, attaching the coil wire, and turning the motor over watching it spark, I used my clamp timing light. The engine runs a bit uneven at that level, but every one of the 12 wires from the coils appears to light off the timing light the same (as long as I get the arrow on the clamp right). Jumps around a bit when shone on the pulley, and pulley has no marks in the vicinity of 16 degrees advance or so, but all that part looks OK.

I got out my hard piston stop, and redid what I had done on the engine stand to locate "true" TDC. This time I used #4 instead of #1, but surprise - results the same. This time I measured things and wrote them down. However, the 6mm cam timing is set using the "true" tdc, but because the difference between the fan timing groove and the pulley is very close to 5 degrees, and true is about 5 degrees counterclockwise from the groove (to the left), I can use the first or most clockwise pulley mark, and the fan housing groove, and come up with the same overlap readings - at least enough for present purposes.



I was all excited about the TEC's ability to alter, electronically, the mechanical (tooth) advance. The crank wheel sensor is basically set to the pulley, so I'm assuming that timing is off. So I made up a bin (program) using putting in 10 instead of 11 where this is done on the screens. Didn't help. Well, maybe I had that backward - maybe it needs to be 12 instead of 11. Other than idling at around 1,000 RPM instead of the 1,200, nothing was different. #6 exhaust at idle 400 F. #4 same. #5 160F. #5 upper plug inner insulator looks white as new (is new plug), and the steel has zero deposits.

I didn't recheck the left bank, but all my previous readings showed header temps all to be in that vicinity and fairly even. And all I have to do is look at the dual EGT. Left needle is up some, right needle just sits down at the bottom. But it works - I've confirmed that.

Everything seems to point to fuel (because when brake cleaner is squirted into #5, things pick up for a while. So after I do a propane check to see if that finds an air leak the brake cleaner didn't reveal, I think I'll switch all the left bank injectors over to the right bank. No matter that switching #s 4 and 5 didn't do anything, etc. And if the loom will stretch, I'll see if I can have the left loom fire the right bank, and vice versa.

I think I'll pull the Buckley headers off, so I can look up the exhaust port, and confirm there isn't somehow something obstructing the exhaust. Seems implausible, but they are relatively easy to R&R due to careful design.

So what's different about the motor after the rebuild?

Lightweight clutch/flywheel? Wouldn't do this. Maybe cams were originally timed using the normal marks? That would mean a nominal 6mm overlap cam (that's a lot, by the way - almost half total lift!) was actually timed at around 5.5mm? That might explain why the torque curve up high is so healthy - I only dynoed up to 8,200 RPM, but my optimum shift RPM calculator says I should be shifting some unknown number of RPMs higher. Advancing the cam ought to bring that top end down a bit, which wouldn't necessarily be bad. But how could any of that affect only one hole?

My using 10, 11, and 12 tooth parameters means five of six fire reasonably well over a 12 degree range.

I am also running a fuel pressure regulator with a vacuum port. Hard to see how that would do this. Pinching off the hose to the regulator port doesn't change anything.

I had used a slightly altered program when I was dynoing my ill fated 2.8 short stroke, but I've put back in various previous LS programs, and checked all the sensor inputs, modifying the fractious TPS as needed. Don't think that is what's doing this.

Something mechanical with the #5 valves that I can't spot?

[Lapkritis]

Which cylinder quit previously?

[tknobby]

Walt,

Did you say you were running sequential injection and now changed to batch? If so the timing of the batch fire on #5 may be wrong. If the injector fires when the intake valve is open the fuel doesn't get a chance to rest on the back side and evaporate. Makes for a poor burn in that cylinder.

Just a thought I had.

Tom Knoblauch

[Lapkritis]

Good thinking Tom. Depends on the injector style. The Siemens Deka are a fine spray rather than a jet spray so injector timing is less important... many injectors are mounted at an angle so they impact the port wall first with the jet spray rather than the valve. Good thinking outside the box. Could contribute if the injector is aimed and timed wrong and is jet spray pattern.

[Walt Fricke]

I'm not running sequential (or sometimes "full sequential"). Electromotive is not 100% consistent in their terminology. The option I use has injection pulses coming once per engine revolution, like the waste spark. This is the same setup which worked fine before. Electromotive is also set up to use a secondary or high speed injector, but I'm not doing that.

The confounding factor to normal diagnosis here (other than my hard head) is symmetry. Not only are all five other cylinders firing, but because they are paired, why would one cylinder in a pair work, but the other not? You'd think swapping components around would reveal one as bad, but that's not the case.

At least so far. I keep hoping I'll hit pay dirt.

[HawgRyder]

Walt...the more I read...the closer I come to the idea that you have an air leak.
Cracked head?...or intake runner?
Perhaps a stuck exhaust valve?...letting some of the charge out?
Maybe a piece of crud on the exhaust seat?
Bob

[Walt Fricke]

Bob

I am thinking similar black thoughts.

Problem is, some of the possible mechanical causes should also show up as low compression or bad leakdowns, or both. And the PMO manifolds are pretty darn stout, as are the TWM throttle bodies.

Still to come, a more calculated infusion of propane around all parts (except the throats) of the #5 intake tract. Squirting brake cleaner did not show up any leakage.

[HawgRyder]

Walt...we had a problem one time with an engine that ran lean on one cylinder.
After about a week of chasing our tails...we tore it down partially.
We found a piece of gasket hanging in the chamber near the intake valve...it must have upset the flow of the incoming charge just enough to make it run leaner than the other 7.
I know that Porsches don't have head gaskets as such...but the small piece of gasket leads me to believe that any slight difference in configuration in only one cylinder can upset the apple cart.
Any possibility that one piston has some sort of profile difference?
Or..rings...top ring on #5 different?
Reaching for straws here.
Bob

[Walt Fricke]

Me too.

Leakdown arguably slightly better than #4 suggests not ring issue. All of intake basically to valve visible down ITB throat with butterfly open. Plus I borescoped it. Nothing in there.

That leaves exhaust, so pulling header is on the list. Because I installed the headers with engine in car, the possibility of a problem with a gasket there exists, though inspection of outside doesn't show anything amiss, and I put a dab of grease on the gaskets to hold them to the port while putting the individual primaries up there and getting nuts on to keep things from falling down. Pretty fail save, but something is wrong.

And it has been suggested that I disconnect the #2 injector, and see if allowing all the current for the injector solenoids for that pair to run through #5 makes it fire (even though this means #2 won't be working. Easy enough to do.

[WERK I]

I thought #5's spark plugs were wet when you pulled them out?

Could you check the crank sensor to toothed pulley air gap?

Oh to have an o'scope at this time would be invaluable. The engine is acting like it is in some sort of "limp mode" doesn't it? Are there on-board diagnostics where you can monitor MAP, O2, crank sensor in real-time?