[cerbera]

Hi all,

I've been trying to exorcise the demons out of a 92 Civic hatch with a D16Z6 5 speed swapped into it. I've been lurking the forums for years, but I'm running out of ideas and things to test. Apologies in advance for the long post, it's been a long, long running problem

The main issue is that it doesn't reliably start on its own, but once it's started, it runs completely fine. When cold, it will usually start after cranking a bunch, pausing, and then it'll either fire up instantly or it'll struggle some more, either cranking with no response, or rumbling for a half second and then giving up. When hot, it'll usually refuse to start. It'll crank, rumble, give up, crank, rumble, give up, crank, rumble, give up, etc. etc. for several tries, either until your patience runs out or it finally kicks for no apparent reason. The only reliable way to start the car when it's hot is with starter fluid shot into the intake manifold. Sometimes when you do this, it really feels like it's the ECU getting confused and giving up, but once the engine picks up the starting fluid and clatters to life on its own, the ECU thinks "Oh, OK, I guess I'll work now." Afterwards, the car runs totally normally, rips straight to redline, gets average-ish MPG, etc. etc. So why it doesn't like starting on its own is really baffling.

As far as things I've done to the car, I've mostly returned it to completely stock mechanically and replaced nearly everything in doing so. It used to run rich, which is the reason for some of the seemingly unrelated work I've done. It would be easier listing out what I haven't done at this point, but here's the full shakedown anyways:

• Engine yanked and rebuilt. All new gaskets, crank ground and bearings replaced. Rings were left alone - crosshatch on the cylinders were fine, so I made sure the pistons were clean, rings moved freely, and the gaps were oriented 90 degrees to each other on install. Comp test showed 175psi +/- 5 across the cylinders across 3 different tests so I don't think it's a factor
• Cylinder head was rebuilt. It had 5 bent intake valves, a bent exhaust valve, and a burned exhaust valve, so I handed it off to a shop to take care of. Lashed the valves after I got it back, and have lashed it again about 3k miles afterwards
• During the rebuild, all the usual mechanical maintenance was done - timing belt, water pump, thermostat, serpentine belts, air filter, etc. etc.
• Also during rebuild, went through and cleaned the TB, intake manifold, etc.
• Quadruple checked mechanical and ignition timing to be dead on
• Verified I had spark with a tester. Got new plugs gapped to .044, ignition wires, cap, rotor, and swapped the ignition coil for a junkyard one anyways. Someone gave me the advice that the only reliable ignition components on these cars is Honda OEM or NGK, so I replaced them again with brand new OEM cap and rotor and NGK wires, but that didn't make a difference
• Injectors were flowtested and cleaned at a local fuel injection specialty shop. Before and after, I had the rail off and checked to see if any of the injectors were stuck open, but none of them leaked
• New fuel filter, fuel pump, and pump sock. Pump has been wired directly to a switch bypassing the main relay, so I could play with how long the pump primes. Priming duration makes no difference
• Main relay was replaced as well. I've tried it with both a ISO relay with a bodgy adapter harness and a known working, used OEM main relay, and it hasn't made a difference to the start behavior
• Battery has been replaced relatively recently. I checked it just now and it reads 12.6V
• Replaced the alternator
• Replaced a couple ground cables, and made sure all of them were bolted to clean surfaces with new bolts
• Replaced the ignition switch. I believe it's an EK one, but it works just fine with wires swapped around to match the EG
• Swapped in a junkyard FPR because the old one had a bent vacuum nipple. Verified with a fuel pressure test kit the pressure read 43psi at the filter, both while priming and with the car running
• Swapped in a stock exhaust manifold that I verified had no cracks, along with fresh gaskets
• Went around the engine checking for vacuum leaks with carb cleaner. Couldn't find anything external
• PCV valve, all the hoses, the baffle box fitting grommet, and baffle box o-ring were replaced. Baffle box I also cleaned
• IACV cleaned, FITV deleted
• New O2 sensor (at first had a Bosch one before putting in a Denso unit), new IAT, new ECT
• Backprobed TPS and MAP at the sensors and from the ECU pins to check per FSM that the wiring and sensors were good
• Likewise backprobed IAT, ECT, and O2 at the ECU to verify the wiring was good (was particularly concerned about the swapped IAT/EVAP matter but that turned out not to be the case)
• Spent a bunch of time sanity checking everything. Some previous owner did the swap, so I was paranoid about making sure everything was actually Z6 hardware. Everything checked out
• Checked the ECU was correct (P28) and not chipped. I ended up finding some leaky caps and swapped the ECU for another P28. The guy who sold it to me threw it in his car to show it worked, and it fired right up and ran flawlessly. Still had issues with my car
• For good measure, I sent the new ECU to be refreshed at a local shop that replaced the caps and everything. When I reported back that my car still wasn't starting reliably, they took it back, threw it in another car, and verified the ECU worked again
• I took the car to the shop that did the ECU. They temporarily swapped in a Hondata S300 just to check on their laptop that everything the ECU was seeing looked kosher. According to the guy, it did, but it kicked back a distributor error of some kind. He figured it should have started anyways, but sent me off to go replace the distributor. When I told him it only starts with starting fluid, he tried bumping start enrichment (or something like that), but that didn't work either
• I swapped it with a known working used distributor. It worked perfectly the first go so I bought it, but the problems came back. I started thinking this was CKP sensor related, but the distributor was within spec when probed directly or backprobing the ECU connector with the distributor plugged into the harness (~400ohms which is within the 350-700 I've read on forums). The old distributor read low (~320ohms) so I'm wondering if that was the original problem, but something new is now the issue

Rounding off all of that, I have no CEL (it does come on for a couple seconds when you turn the key to ON, so it's not just a missing bulb). The only other problem this car seems to have is the speedometer will randomly stop working, which then trips a CEL for that. Sometimes I'll hit a bump and the speedometer will suddenly start working again, sometimes it stays unresponsive. This, along with the inconsistency of how the car behaves when you try to start it, made me think the wiring was to blame. I took out the engine bay wiring harness and unwrapped it, but found basically nothing. No wires worn through, nothing falling out of the connectors, no green death, etc. The only thing I found was that the distributor harness shielding was kind of falling apart and was barely attached to the ground cable. It ended up fully ripping apart when I cut back the insulation to check if the distributor wires were OK. Either way, I can't imagine a way that the shielding would cause my problems.

Later this weekend, I'll be properly inspecting the wiring that's still in the car, but my initial checks didn't turn up anything obvious. I'm going to be replacing the distributor wiring anyways now that I've effectively destroyed the shielded cable from unwrapping it, putting the harness back in the car, and testing the injectors with a NOID light. I want to see if my theory that the ECU is turning off the injectors for some reason is true and that's why the starting fluid works. Even if that was the case, I don't really know what would trigger that behavior.

I've included youtube links of videos I've taken of its starting behavior. I wish I had a video of the car starting with the starting fluid, but I never bothered.

And also here's pictures of the engine bay (somewhat old pic before I got the NGK wires) and that questionable shielding fraying

So... any suggestions at this point would be greatly appreciated

 

[robgoof]

FITV deleted

pop the brake booster vacuum hose off the intake manifold and give it a try.

 

[mattliston]

Loosen up the distributor mounting bolts, rotate the top towards the radiator. This retards the timing. Temporarily it should help make engine starts far easier.

If it is far more reliably starting, you likely have an ignition control module that is lagging a bit, or resistance in the wiring harness of the distributor somewhere.

Also do spark checks and be sure they are consistently firing. Checking for spark is one thing. Make sure it is consistent. Unplug fuel injector big plug near brake booster prior to doing this so you do not flood the engine with fuel.

Be sure NGK or Denso plugs are looking good and not showing any signs of coolant or oil contamination.


Double check VALVE LASH ADJUSTMENT. Super important. There are physical marks on the head and cam gear. Use a short metal ruler to line the teeth up during each cylinder rotation. 1-3-4-2 firing order makes it super simple.

Running fine once started means you are dealing with proper components, but something is not happy under the harsh environment of starting.


Temporarily add a battery ground that bolts to the alternator bracket.
Double check alternator and starter wires are in good condition.

Remove distributor cap, disconnect ignition coil, and test the ignition positive for battery voltage during cranking. If it drops below 11.0 volts, it can expose a tired battery, or hurt wiring.