In general you are correct, and on 928s prior to the S4 I would agree that individual intake runner sealing should be at the top of the list since those motors have multiple ways to introduce false air to individual cylinders.

However, on an S4, with one exception, you’d almost have to try, deliberately - with a drill - to get false air to just one cylinder. And if the O2 sensor is working, the MAF is fresh-ish and the false air isn’t pretty massive, the 02 control loop will adjust for a general engine-wide lean condition.

The exception is the lower injector o-ring. On an S4/GT/GTS that o-ring is - of course - a unique part. The ‘standard’ o-ring isn’t thick enough and if used has a high probability of allowing false air into the cylinder. Not 100%, just high.

Your post caused me to re-review this thread. The OP recently (this summer) had his mechanic do a partial intake refresh.

I asked the op (macerl) what was done with the injectors. They had to come out to do the intake. Macerl responded with “injector o-rings”. If we assume that the correct o-rings were used then false air goes to the bottom of the list.

It is not crystal clear if this behavior was noticed before or after the partial intake refresh.

We do not know, for sure, if the intake gaskets were replaced. Macerl didn’t specifically list them. Reuse of old gaskets might result in the same failure-to-seal problem you had. But, damn would it be not smart to reuse 27-year-old intake gaskets. Or even 10-year-old intake gaskets. More like 5. No, really why wouldn’t a professional use new regardless?

So, my logic is:
If we assume that Macerl’s mechanic has typical professional risk adversion and used new intake gaskets (and presumably new pressure washer gaskets too) but isn’t well-versed in ‘87-‘95 928 lore, mis-sized lower injector o-rings are the source of any false air to individual cylinders.

I still bet on bad spray pattern from one or more injectors since they were apparently not (internally) cleaned during the refresh. If Macerl sends his injectors out and specifically specifies/reminds that the correct over-sized o-rings must be used then both potential problems are solved.

On the other hand the WTF? $h1+ I have see done to 928s by ‘professional mechanics’ is a never ending source of amazement.

Your observations are consistent with:
- a failing LH (why it happened right away with the ‘old’ LH)
- the period of time it takes the LH’s adaptive O2 system to ‘adapt’ (why it took 30 minutes the first time)
- the period of time it takes the 02 sensor to warm up (why it takes 5 minutes)

To test, disconnect the battery for a minute. Reconnect and test drive. I’ll bet it takes about 30 minutes to start its missing - assuming you take the same route in the same spirit as before. Vastly different result would be interesting.

Still, all consistent with dirty injector(s) or - as per Marti - “weird” false air due to a screw-up.

Any further updates?