[Non Ford] Intermittent Rough / Sluggish Running

[Citroënbender]

Got a car that’s doing my head in.

Alfa 2 litre “Twin Spark” 147 hatch that intermittently runs terribly and is then weak, almost stumbling, on light throttle. 246K on the chassis, a bit less on the motor.

Number 3 cylinder seems to be the culprit, based on its plugs. It has headers with 1+4 and 2+3 and separate primary oxygen sensors. The 2+3 sensor is reading over 3V when at stabilised operating temperature. This flags a fault that doesn’t illuminate the engine light.

New coils
New correct NGK plugs
ECU tested by substitution
Fuel pump (genuine Bosch from reputed vendor) six months old
Full tank of 98
Hot compression 205-220 PSI on all cylinders

On #3 I have tried used plugs of good colour. I have swapped in a spare injector. I have swapped good running plug boots/leads from my other car with the same engine.

Wiring checked
Charging voltage checked
Throttle percentage seems to correlate and doesn’t misread (glitch) in live data. Clutch is not slipping.

At road speed and an indicated 8-12% throttle the effect (when it occurs) is of losing power as though compression braking - but a shade rougher. Instantaneous fuel consumption rises heavily. At idle, if it’s “having a moment” it will rattle and shake heavily - like when the Spirit of Tasmania is berthing. Yet it almost seems that briefly flicking off any of the coil connectors worsens this behaviour.

When you blip the throttle, you can see the intake trunking flex in response to throttle blade movement, but there is a small lag in engine response - a “soft” response as though either spark timing is wrong or fuelling isn’t adjusted quickly to compensate.

What I haven’t done (car is out on loan again):
Substitute tested lambda sensor
Leakdown testing
Fuel rail pressure check (odd fitting like an AC valve)
Vacuum read

I’ve also googled heavily, to no avail.

Any suggestions for diagnosis? When it’s running well, it’s a high sevens per hundred car at freeway speeds.

[Peter B - CV8]

Air flow sensor ??

[Citroënbender]

Good point to consider. The one in it has about 160K.

Normally their elevated location in these motors and air intake trunking size means they are trouble free - but I neither inspected this nor the air filter.

Exhaust cats are all fine, checked ten months ago.

[Citroënbender]

There are two cats from the collectors, each with their own sensor before the cat. Sensor 1 is cylinders 1&4, Sensor 2 is cylinders 2&3.

There is only one catalyst efficiency sensor, this the after-cat Sensor 1. Hence a couple of inaccurate data lines.

Cylinder 2 plugs on the left, cylinder 3 on the right. Probably less than 3000km of mostly urban driving.

This is not a great photo, I tried to show the fairly visible clean(ish) piston top with oil film.

When I changed the belts in January 2022, I checked the cam lobes for wear by visual inspection. All looked normal. I did not remove any lifters, hence not check the valve seals or guides. The engine does enjoy a sip of oil, but I’d assumed it was the usual reason - weak scraper rings.

Is it possible, that a valve guide or stem seal have become faulty (ie; split/worn guide or deconstructed seal) and at light throttle (high vacuum) it’s sucking oil which is causing the stumble and whacked lambda reading? I’ve no personal experience to comment about the real world effect of oil burning on oxygen sensors.

[oldel]

The piston looks too clean, the spark plug has white discolouration.
I'd suspect head gasket.

[Citroënbender]

There’s no excess or rapid pressurising of the coolant, or coolant loss beyond the known weeps (since 22 year old o-rings don’t take kindly to being disturbed). The whiter plug is from the adjacent cylinder to the one whose piston top is shown. I’d more think the ECU is leaning out the mix for that cylinder pair until it pings, as a response to the over-rich oxygen sensor readings.

These cars are a huge hassle to get the cylinder head off in-situ, so that’s another reason I really hope it’s not a gasket issue.

[Dr Terry]

AFAIK the 2 pre-cat oxygen sensors on 147 twin-sparks are the same as each other.

Why not try swapping these 2 sensors & see if the result is the same.

Dr Terry

[Citroënbender]

Yes, they are the same part. I can try this next weekend when I see the car again. Going to pop in a new air filter as well to see if that changes the lag.

You’ve seen plenty of cambelt failures across brands, I presume. How often do you see “knock on” damage like ovalling or cracking to guides in an alloy head?

(This motor was a bitser exchange from ASV in a previous car that got T-boned, so I presumed they did the bare minimum to get it saleable. The timing cover was damaged from prior belt failure and head casting date too far from the block casting to have been built originally that way.)

[Dr Terry]

Originally Posted by Citroënbender

(This motor was a bitser exchange from ASV in a previous car that got T-boned, so I presumed they did the bare minimum to get it saleable. The timing cover was damaged from prior belt failure and head casting date too far from the block casting to have been built originally that way.)

Beautiful, just beautiful !!

Dr Terry

[Citroënbender]

Originally Posted by Dr Terry

Beautiful, just beautiful !!

Dr Terry

I can be naturally blunt. Not out of a desire to be offensive, but because I often miss the cues to be subtle about something.

What about my question on valve guide damage? Your seasoned opinion would be appreciated.

[Dr Terry]

Valve guide damage is a strong possibility, we used to see it regularly in the early days of unleaded petrol (mid-late 80s).

Unfortunately you can't prove it by any external testing, you've got to bite the bullet & strip the heads down, once all other possibilities are exhausted.

Dr Terry