Post-interference event poor idle; not vac'm <?>


Folks, I’m at the end of my wrenching-skills, following a post-belt-break
valve-job. The vehicle is an E30’vert with the M20B25 (non-eta) engine.
I had a few valves break, and couple bent - I replaced bad ones
from a donor head, all were ‘roll-tested’ for straightness, dip cleaned
and polished - and finally lapped in-port upon re-installing. Yet;
I have a “lumpy” idle (made worse by removing the oil-fill cap
on the valve cover), so I first thought blow-by…yet I tested, and:

  1. After doing compression test - all good/spec - I did a leak down…
    A. All cylinders similar, under 5% loss on any one at TDC
    B. No audible hiss from exhaust, throttle or in rad/coolant
    C. However, EVERY cylinder had slight air hiss through vc’s oil cap
    - It was audible, but not enough to even wiggle tissue paper
  2. When rebuilding the head/lapping valves, I cold-adjusted them to spec
    A. Running, the valvetrain is quiet; no audible tapping or clicking
    B. No other anomalies in running; pulls strong, no lagging or smoke
  3. To try cylinder isolation, once plugs were reinstalled, did plug-yanks
    A. Sequentially removing plug leads, tried to find “weak” cylinder
    B. Spark removal equally degraded idle, irrespective of cylinder
  4. In the event that I missed some vacuum leak, I re-looked, sprayed, etc
    A. Double checked the “b*tch tube” which did appear seated in its hole
    B. I had already replaced the intake boot, spritzed hoses, and the usual
    C. …I’ll not say that any subtle TPS issue was missed, but jeez!

It’s official; I am out of ideas. Any next-step tests welcomed. I am
loathe to yank the cover with nothing apparent to look-for <?!>.


Just because you can’t find the vacuum leak, doesn’t mean you can rule it out.

I’d try a smoke test. Google DIY smoke generator, or have a shop do it. Basically what you do is take a steel can and put in 2 air in/outlets. Then you need an air supply throttled down to a couple psi. Soak a rag in oil and throw it in the can. Light it on fire. Once it quits burning, put the cap on the can. Use the air source to blow smoke out. Pipe smoke to intake elbow with the elbow plugged with a rag. Watch for smoke coming out somewhere it shouldn’t.

I’d also swap out the ICV with a known good one, and test that the ICV leads have good continuity to the DME.


Quick update: I struck out on the smoke test (had made a jurry rigged one for past E46 testing), as it indicated no apparent errant wisps.

I finally did a stomp test - was late in considering that, having no CEL illuminated - to see if it reinforced the vacuum thesis; TADA it yielded 1-2-2-2 (Lambda 1) code: This still can be sensors, vacuum, O2, etc; so I guess I am still back to checking final connections made upon finishing the head rebuild ; …never any easy ones, I guess…but all ideas still welcome.


A smoldering oil soaked rag in a can is not substitute for a proper smoke test done with a pro-type system. Small leaks far from the intake may not be found with oil in a can method.


Interesting repair solution to bent valves… Assuming that you don’t have electrical issues and the ICV is functioning, I also think you have a vacuum leak. However, given your repair method, I’m also suspect of your valve train in general. What were your compression numbers? Are you 100% confident in your less than 5% leak down test results? Leak down testing can be tricky and I only ask because we are really shooting in the dark if your testing isn’t reliable. If your valves aren’t sealing, that will certainly cause rough running.

Assuming your “valve job” worked and the valves are in fact sealing, I think you could have other valve train related issues. At valve overlap in an M20B25 (360 degrees after TDC), the exhaust and intake valves are both briefly open about .035" before the exhaust valve fully closes and the intake valve opens. In real racing engines with “big” cams, the overlap period is longer in duration with higher valve lift, but also one of the reasons why their idle is lumpy. However, our engines are like most gas engines and the intake and exhaust lift at overlap is symetric about TDC. Altering the point where exhaust valve and intake valve lift are equal to before or after TDC largely influences the torque curve of the engine (Spec E30 tuning tip). Little changes make a huge difference.

I mention this because moving valves from one head to another and not doing a complete valve job may have introduced some variances in your valve train geometry. Valve seats influence the location of valves in the head and when a good machinist completes a a valve job, he will indvidually adjust the valve stem length the make sure they all have the same height. You can ball park this in your head by looking at your lash adjust eccentrics. Assuming that your rockers aren’t worn,and they generaly don’t wear on the base where you adjust lash, the adjustment holes should all be rotated at about the same angle. If some are rotated more than others, this indicates that you have variances in your valve train that could influence the opening and closing rates of valve events across your 12 valves. It won’t hurt much, but could effect engine smoothness.

Otherwise, you might just have your valves adjusted poorly and be creating a valve overlap condition that degrades your idle. I would suggest re-adjusting your valve lash. It’s much easier to set the lash by placing the feeler gauge between the cam and the rocker arm (instead of between the valve and the eccentric)to get a consistent adjustment. To achieve a specified .010" adjustment, use a .008" feeler gauge on the camshaft to account for the rocker arm ratio. If that doesn’t help, I would open the valve lash up a bit to say .012" (.010" on the cam") to decrease the duration of valve overlap.

A few other questions…

Do you surface the cylinder head?
What head gasket did you use?
Did you change anything in the valve train other than the bent valves?
What else did you replace in the engine?


Continued thanks, guys - I’m seeking someone with a proper smoke test rig, as my homebrew set-up hasn’t evidenced any vacuum leaks. Hope to get that arranged soon!

On the rework questions; I did not surface the head, beyond cleaning; it was flat in all directions per a machinist’s rule. I went OEM on the gasket, and aside from lapping all valves (existing and replaced ones), no other bits were changed.

Meanwhile, I’ll do a re-adjust on the eccentrics this weekend, and report back if that has made a difference.


Machinist’s rule or a precision machinist’s straight edge? It matters which you use.

I’m with Fishman on the valves. It would have been much better to have a shop grind the seats and valves and then lap them in. And the shop should have also checked the guides for wear. After that do a drip check with the head off the car to prove that the valves are sealing.

The cam is also something to consider. You can find gross errors with a micrometer, but the best check is to have the cam checked with a Cam Doctor as it will check the profile of each lobe.

By by all means have a proper smoke test run first. There are other possibilities, see for what I do in a case like this.


All - per the above, I began looking for e30 expertise in Indianapolis: I’ve now tried Merkle (dissatisfied/no-help), Mikar (great guys; too busy) - and it looks like I’m relegated to Reggies…who is well respected by the racers, but I suspect will be VERY costly as a “boutique shop”…

Does anyone have any favorite BMW service experts in the greater Indianapolis area? It’s quite driveable - but wouldn’t relish a trip to Chicago unless my only recourse. 'Continued thanks, guys.


I’ve heard good things about Reggie’s and saw pictures of his shop once. It looks like the kind of place I would want working on my car if I didn’t do it myself. You get what you pay for most of the time and it’s often cheaper to pay the expert premium up front. If I were in your situation, I’d give Reggie’s a call.


I know of people that had good work done at Vaughn Motorwerks in Zionsville as well. Not a first hand review, but I trust the story teller.