Hilton

I've been trying for a while to get my '87 through an emissions test. Failed the last attempt running with no O2 sensor - using non-cat maps per coding plug (car was originally non-cat), so I've decided to take the idle mixture adjustment guesswork out of the equation and set it up to run closed loop, with a cat coding plug, Bosch narrow band sensor etc.

I've just installed a new narrow-band O2 sensor, and chucked on a Motorsport X-pipe so I had two bungs, so I can run my wideband in the other bung. Once I can verify the car is running properly I'll stick the factory cats back on with the O2 sensor.

What I've just found is that I get different readings from the wideband on either bank (the separate narrow-band sensor is feeding the ECU) - but that the AFR hunts around in the sort of range I'd expect.

Findings are:

Wideband on 1-4 side, NB on 5-8 - reading 15.0-15.3 AFR
Wideband on 5-8 side, NB on 1-4 - reading 14.0-14.2 AFR

It looks to me like the regular Narrow Band O2 sensor is doing its job - but that the O2 adjust is pushing the fuelling out of whack because the sensor gets different readings on each side.

So - where to start diagnosis? There's no obvious misfire or lumpiness to the engine.

I have a new test booked for next Wednesday (it takes 4 weeks to get a booking at the test center), so figured I need to get this sorted in the next couple of days.

Background:

• Temp II, idle, load switches test ok at the LH plug. Temp II matches specs in WSM at 25, 40 and 80 degrees C.
• MAF is a 2012 JDS rebuild - I have other spare MAF's to swap in if needed
• Plug wires are 2yo and show no fireflies in the dark
• Spark plugs are brand new Bosch WR7DC+
• New dist caps and rotors
• Cams are timed correctly (verified with 32vr)
• No ECU faults and all actuator tests seem to be ok.
• I pulled all plugs after running the engine a bit - none wet or smelling of gas
• I have a sharktuner, JDS Spanner, etc. at my disposal.

Things I'm planning on trying:

Any suggestions on how else to diagnose this?

Speedtoys

A misfire on the 1-4 side could show a lean reading.

?

Hilton

Could be. Trouble is I don't know whether its running lean on the 1-4 side, or rich on the 5-8 side?

How to check a misfire?

I've got an inductive light I'll go try on each plug wire as a first step.

Suddenly thinking I should get my fuel pressure gauge out and attach it too - see if there's a leaky injector.

I'd like to get to the exact bottom of this, before resorting to bulk-swapping ignition wires, dist caps, injectors etc.

Possible causes I can come up with so far:

• Misfire due to plug wires/boots/dist cap?
• Leaky injector making it rich on one side
• Clogged injector making it lean on the other side

Hilton

Am I right in thinking that a wiring fault in the injector harness would knock out 4 injectors and be pretty bloody obvious?

Speedtoys

Thats why It probly isnt a large fault. A weak old wire, cracked rotor cap..?

I dont suppose you pulled the battery and cleared out the LH memory between each wide band bank test did ya?

A trim lean will fatten the other side, or a trip fat will lean the other side...

john gill

Hello Hilton

your issue is camshaft position , or to be prcies the difference between them you will find that the belt moves them around a lot in its usual operation.
I have spent hours and hours playing with this position , and I have settled on h 1 degree difference between each bank. The porken tool is good for this , but you need to know where each camshaft is before you use it . What I mean by this , is that you need to measure the lift of the cams to know where they are positioned , (there is a difference). Once this is established and you note it down the Porken tool can be used to make the final adjustments , as once positioned you have a baseline measurement .
You should end up within about 3 % difference in Lambda between each bank , I Thoroughly reccomend that you use Lambda instead of AF .

Hilton

Thanks John, I have the porken tool, and verified that the cams are correct (1 degree retarded on 1-4, 0 on 5-8), as it did occur to me that cam timing could have been the culprit.

I've just listened to the injectors with a stethoscope and all are firing. Next I'll be adding a fuel pressure gauge to the rail to check how long it holds pressure and whether an injector is leaking - although I did check fuel pressures when I was starting the car up a couple of months ago after intake refresh, so expect it to be ok.

I'm leaning towards a misfire on 1-4 at this stage - the exhaust smells a lot less rich with the O2 sensor running the ECU from the left side (5-8), so I'm guessing the lean readings I'm getting are the correct ones.

Hilton

Good sanity check - I was pretty sure I had, but went and disconnected the battery again to make sure, and after a few mins running its back at the same 15.0-15.3 oscillation as measured on the 1-4 side.

So.. next steps:

• Check fuel pressure for leaky injectors
• Take apart 1-4 side plug wires, inspect the joints, and measure each part as well as each complete reassembled wire.

If after the above I can't find anything telling, I'll check 5-8 wires for good measure, as well as both distributor caps.

Hilton

Fuel pressure holds for 30 mins with less than 0.5bar drop (52 to 47 psi) - so not a leaking/freeflowing injector causing additional fuel.

I took off the dist caps, both looked fine, so I put them on the opposite sides, which would move the problem to the opposite cylinder bank if one cap has a fault. Also inspected the plug wire boots at the caps - all look fine.

Started 'er up so I could remove pressure from the fuel rail and take off the pressure gauge, and the problem is still there, and on the same side, so its not the distributor caps.

Time to remove and disassemble the plug wires.. ug, fiddly little M5 bolts in hard to reach places.

Hilton

Well, one of the distributor-end boots on the plug wires seems to be broken (cylinder 4). I struggle to get a resistance reading - if I compress it hard I can get 2-2.4k Ohms, but its open circuit most of the time. Those boots are meant to be 1kOhm.

All the other bits measured ok - so I'll replace just that boot (I have a couple of used spares around) and reassemble tomorrow, and see if that fixes the problem.

docmirror

Swap the O2 sensors from left to right and verify the delta stays the same, or follows the sensor. The left dist feed two cyls on the right bank and vice-versa, discounting ignition. I believe the injectors fire in the same group with the dist.

Hilton

Yep - I'll be taking measurements with the WBO2 sensor on both sides, and the narrow-band on the opposite side running the ECU.

Hopefully the plug wire issue is the only problem in there - its 105F today and 40% humidity, so the prospect of doing much over a hot engine isn't too appealing.

Bill Ball

A short anywhere in the harness takes out all the injectors since they all share a common ground. A cut wire to one injector just takes out that one injector.

I had a similar dual O2 sensor reading disparity and I finally located a corroded spark plug wire plug.

IcemanG17

Interesting.....initially I was leaning towards a gross camshaft miss timing....but Bill's answer about 1 bad spark wire also makes sense....

Easy check....pull the spark plugs and look that them....is one grossly different in color than the other 7?

john gill

Hello all

Just re read my post again , from last night and I am appalled about my lack of diction and lack of spell correction previous., I was having few drinks at the time , sorry no excuses.
ANother thought that I had, was that one of the cams may be a tooth out , a quick way to check , is to pull each of the injector plugs off on one bank one at a time , you obviously should notice a significant change in the engine idle .

If you do not get a change on one bank , this is a good indication, the engine will run fine on the old engine managemnt system 1 tooth out .
WHen you change to sequential and aftermarket different story.

Just something else to muddy the waters though .