Classic Hotwire V8 Runs Rough at Idle

I think the silicone RTV ones that release acetic acid fumes as they cure should be avoided upstream of Lambdas.
Never used RTV on exhausts, just firegum or trade equivalent - the name of which escapes me but comes in one of the cartridges that requires a gun, same as household sealant stuff - but both are a white paste.

IME of even small leaks for MOT purposes they usually result in a fail.
Did see one occasion on Mrs r's car where there was a small hole in the rear silencer & it failed, but just putting a nitrile glove into the tail pipe to create a small obstruction turned the result to a pass.
The tester did explain but it went a bit over my head!

One time i had a car that was in the red on the analyzer ,the exhaust prob was only half way in ,When the mot guy noticed ,well i noticed n pointed it out ,and the prob was pushed in all the way ,and it was perfect......If its not in properly it gets to much air up the tailpipe

Ok, so I’ve acquired a better manifold and swapped that with new gaskets to the head and down pipe - no more leak there. Lambda signals still haywire and idle no better. May actually be worse to be honest. Filled with fuel which I have noticed seems to make it a little better. Goes from about to stall when nearly on red to only stumbling with a full tank but could be placebo.

Done a fuel pressure test this afternoon which seems to indicate that the pump and filter are fine although my pressure tester (USPro Tools from Amazon although almost certainly the same as all the other Chinese brands) is so bad I’m not sure to trust it. I guess if it’s leaking it can’t be reading more pressure than is really there. Definitely don’t recommend getting one the same yourself. Despite the leaks I was able to verify that I have 2.6bar with the pump running and pressure regulator disconnected or engine off. At idle with the reg connected to vacuum this drops to around 2.2-2.4bar. Pressure drops off much quicker than the specified 0.7bar in a minute but the tester was leaking so that would be expected.

I think I need to look elsewhere but where next I wonder…

Rich

RAW111111 said:

Ok, so I’ve acquired a better manifold and swapped that with new gaskets to the head and down pipe - no more leak there. Lambda signals still haywire and idle no better. May actually be worse to be honest. Filled with fuel which I have noticed seems to make it a little better. Goes from about to stall when nearly on red to only stumbling with a full tank but could be placebo.

Done a fuel pressure test this afternoon which seems to indicate that the pump and filter are fine although my pressure tester (USPro Tools from Amazon although almost certainly the same as all the other Chinese brands) is so bad I’m not sure to trust it. I guess if it’s leaking it can’t be reading more pressure than is really there. Definitely don’t recommend getting one the same yourself. Despite the leaks I was able to verify that I have 2.6bar with the pump running and pressure regulator disconnected or engine off. At idle with the reg connected to vacuum this drops to around 2.2-2.4bar. Pressure drops off much quicker than the specified 0.7bar in a minute but the tester was leaking so that would be expected.

I think I need to look elsewhere but where next I wonder…

Rich

Click to expand...

I know nothing about the V8, but have you checked the in tank pump? The variation with the tank level may be an indicator of a faulty pump.

Datatek said:

I know nothing about the V8, but have you checked the in tank pump? The variation with the tank level may be an indicator of a faulty pump.

Click to expand...

Thanks for the comment. I’m going to test the pressure again when I’ve run this tank down to see if there’s any difference in pressure. If so I’d guess that it’s an issue with the pick up in the tank drawing air or something perhaps. At the moment I’m leaning towards placebo to be honest on that one. Even low on fuel it runs fine off idle which wouldn’t make sense for low fuel pressure anyway I think as it would get worse the more fuel the ecu called for.

I think my suspicion is back to the lambda sensors at the moment. Will check them again next unless someone has any better ideas.

Rich

What are the fuel trims now?

You said previously "Both are +100% at idle and within about 10% either side of neutral on the move"... which suggests it was adding fuel at idle (perhaps due to an air leak in the exhaust?) but were OK on the move - possibly from more exhaust flow sucking in less air?

You need to record the long and short term though, as the short term is just a fine tuning of the long, really. That means the short term could stay around the "zero" mark, but long could be anything from -100 to +100 on either bank. You can't use one figure by itself.

If you unplug the ECU for 30 seconds, then let it idle while hot for a few minutes, it'll reset the long term trims. They'll start at zero and adjust based on the short term readings. It takes a minute or so to start, then takes another minute or so to adjust. You can also just drive around, if you want, but it has to be in the idle and low load zone to adjust much.

As Ratae mentioned, partially blocking the exhaust can create enough back pressure that any leaks at the front don't suck in as much air (at idle, HEAPS of air is sucked in leaks, between exhaust pulses.

My manifold to downpipe oinms are wrapped ih exhaust paste/header wrap, and alloy tape, to get them well sealed and I still run -40 and -60 for long term trims, possibly due to lambda sensor or injector differences (though the injectors are brand new genuine Bosch, the Lamdas look original!).

A leaking injector won't show on an individual plug either, because all the cylinders near each other pretty much all share fuel from each other - the batch firing is not timed to open valves or individual cylinders - it's ENTIRELY off when the ECU sees ANY spark... so lots of injection events spray onto closed valves and the fuel is shared all through the plenum.

Thanks for the input. Done a little more diagnosing over an early lunch break...

I've gone through the ETM diagnostic procedures for the oxygen sensors. I've verified that the fuel pump relay is working correctly on both the lambda heater and fuel pump circuits, and is of the correct type. Apparently people sometimes fit normal green switching relays by mistake meaning that the heater is on all the time except when the engine is running rather than all the time the engine is running. I verified the ground for the heaters to the ECU and warmed the engine up. Checking the signals from the sensors I discovered that there was no signal from the left bank at all at idle (0.0V) and the right bank is reading lean (0.1V).

So I moved on to testing the sensors them selves. First I verified that the heater circuit is working all the way to the sensors themselves. If you're doing this your self you should bear in mind that the heaters run with the fuel pump. While the ETM says check for battery voltage with the ignition in position II, the fuel pump only runs for two seconds so I used RoverGauge to run the pump continuously. Next test is to check the impedance of the heater coils is in spec which is tricky as it's hard to get the probes on them but both mine were in spec. In trying to do this test I found the left sensor to be particularly hard to get a probe to since I've managed to push the pin for the left lambda sensor heater out of the plug last time I had that off. That wouldn't have been helping matters and explains why the left bank has been reading lean even more since I had the downpipe off a few weeks ago.

Still doesn't fix matters though. Here's what RoverGauge shows with the engine idling at operating temperature and then at raised revs but under no load. You can see that the short term trims are +100% on both banks at idle and more the odd than the even when off idle. Under load it gets down to closer to neutral. Long term don't seem to be moving at the moment but with it getting reset a lot at the moment I think I need to drive it for a bit now that they're both heated so it shouldn't throw a code again and see where that one goes over the next bit of use.




I'm getting towards my wits end with this after 4 months. I'm considering the options for fitting a Thor 4.6 and gearbox from a P38 instead as I don't want to waste too much time and money trying to get the wrong V8 to work if I could fit a better one instead. Looks like I could probably manage that swap for around £1500 maybe and possibly get something back for the engine and box I have so...

Any help appreciated.

Rich

Just because it was playing on my mind, I had to go back out there and repeat another test I've done before. AFM diagnostic procedures. The AFM on the car was an unlabelled one with no cap on the idle CO screw. I've set this back to 1.8V before as it's supposedly meant to be for Lambda equipped cars, but for some reason it was all the way up at 4.8V. I didn't think this was meant to affect lambda equipped cars anyway except if there is a fault, but I remembered I had another AFM which came with the car and I've never tested before. It's a genuine Lucas one albeit a bit battered looking and with the gauze removed for some reason but it has the cap on the Idle CO screw.

I swapped the gauze over, ran the voltage checks and it appeared like it might be functional and the Idle CO was set to 1.8V, so I started the car up. It reads about 1.5V at idle rather than more like 1.65V where the other was, but straight away the idle is much better and the Lambda sensors were actually switching! Still not running perfectly but it's smoother than it's ever been before for sure. Took the car just around the block and parked back up and the long term trims have adjusted. Now sitting at +100% on the odd bank and +25%ish on the even bank but that has the short term trims within 25% at idle. Still not right but better than ever before.

I'll drive it this week and report back. I have some towing to do for my Mom on Friday and am thinking of going to the Classic and Restoration show at the NEC on Saturday and then bringing the trailer up to Nottingham so should be able to put some mileage on it.

Rich

Maybe reset the ECU, after the MAF change and O2 reapirs and let it redo the long term trims again. I'm not sure on a Classic but on a Disco 1, the engine bay fusebox, fuse 7 is parallel and closest to the front wing... pulling that one (marked "Fuel Injectors" on the legend) will instantly reset the ECU/trims.

AllanWorms said:

Maybe reset the ECU, after the MAF change and O2 reapirs and let it redo the long term trims again. I'm not sure on a Classic but on a Disco 1, the engine bay fusebox, fuse 7 is parallel and closest to the front wing... pulling that one (marked "Fuel Injectors" on the legend) will instantly reset the ECU/trims.

Click to expand...

Hitting the button in RoverGauge to clear the fault codes seems to reset it all too even if there aren’t any codes I think.

I don't think mine does that?

I know it doesn't read them until after a spark event to kick the ECU into life, i.e. starting the engine. If you just turn on the ignition, they all read zero until the spark is registered by the ECU and the start of the ECU program is triggered.

AllanWorms said:

I don't think mine does that?

I know it doesn't read them until after a spark event to kick the ECU into life, i.e. starting the engine. If you just turn on the ignition, they all read zero until the spark is registered by the ECU and the start of the ECU program is triggered.

Click to expand...

You've got me wondering now. I know mine were definitely reset yesterday and updated while it was idling after I swapped the AFM.

Rich

Well after a few thousand miles I can report it was definitely the MAF. Running lovely now and fuel trims are back to sensible if not optimal values (+-20% of 0). I’m guessing this is because I have an unknown mileage 3.9 running on a 4.2 tune with no catalyst in the cats.

Thanks for the help and suggestions.

Rich