Sounds like a dirty or failing MAF to me

I'll drop in a new one in a few days and will let here know if it resolves the issue or not. Thanks!

I don't own a Boxster but was thinking of buying one. Anyway, I was reading the owners manual and remembered about automatic speed control. Don't know if it's related but it seems to fit . From the manual:

"With the speed control system engaged, the
engine speed will not return to idle when re-
moving the foot from
the accelerator pedal"

Hi! Thanks for the advise but this is not the case. Speed control is essentially "cruise control" feature that I do not have have in my car. I'm gonna drop a new MAF today and see if this fixes the issue.

On a side note I've found that turning off AC helps a lot with the issue. I remember somebody with the same issue already found this side effect. Interesting.

The ECU will increase the idle speed when the A/C is on to account for the additional load and keep the RPM's at the normal level. This seems to accentuate the problem that you are experiencing.

Good point, thanks!

Isn't there a sensor/potentiometer for the accelerator pedal that could be faulty?

I have early non-egas model so there is no sensor at the pedal. It's mechanically connected to the throttle plate and I've confirmed that the link does not get stuck and moves freely.

I have just installed MAF and a new air filter. Plus I have found a small air leak in AOS where it gets connected to the throttle body. Somebody before me have damaged the sealing ring on the AOS tube at the side of the throttle body and some air was leaking through it bypassing the throttle plate. I'll change the sealing ring and will try it out.

Big thanks for suggestions!

progress
 

So I have installed the new MAF, changes the engine air filter, and fixed a minor air leak in AOS tube. Did not help at all. But the engine runs better at least :). Next coming up idle control valve. Also I have found an oil leak from the oil filler tube, looks like the tube has a crack in the middle section (the corrugated section) which can also cause a small air leak but I doubt that this is related to my problem.

One interesting observation - RPMs get stuck when I drive faster than 40 mph and will drop back to normal when I decelerate below 38 mph. I think it's the idle valve that does not close properly.

The cracked oil filler tube is a common problem... And a pain to change out too.

I hope you figure out your problem. Sounds like you've discovered others along the way...

I've got the new tube and tools for taking the spring loaded clamps. I read that it's PITA to change but will have to anyway. Better to find and fix it sooner than later ;-). Thanks for follow up and all the suggestions guys! The new idle control valve is on its way, I'll install it hopefully tomorrow or day after and will let you know.

Well it's not the idle control valve either. Dropped a new one yesterday and today I've got the same RPM issue on my way to work. This was my best bet. So the last thing to try is the AOS I guess. And the oil filling tube which is cracked in my case. Will keep working...

No I don't think this is AOS. AOS is connected after the throttle plate and bypass MAF. When I was logging MAF reading while driving I have seen that when the RPM get stuck the MAF sensor reports airflow (and the throttle position is closed at the same time). If somehow air would be leaking through AOS I should not see any MAF readings too. The ICV was my best bet but with the new one doing the same thing I don't know where to look next...

Something doesn't add up here - you say that when the rpm is stuck (open I ssume) you are seeng air flow across the MAF but the throttle position is closed at the same time!!

(I'm playing devils advocate here). If thats the case, then air flow is passing the (supposedly) closed throttle body - which isn't fully closed as the engine would not run at high revs with the it fully closed. So could there be a problem with the throttle body (I know you've checked, but..)?

I'm also assuming that fuel delivery is controlled by the throttle body position, so if it is sticking open you will still get fuel through the plennum. If its not sticking open, hows the fuel entering the engine ???
Can you check your fuel trims when this problem occurs.

Yes, I know. That's why I'm asking some advice for you guys :). Can't figure this one on my own. There is only one way to bypass the throttle plate and get air flow into the body - via the idle control valve. I'll keep digging, will log everything while driving and will try to analyse the data to see where the problem can occur. I found a few forum posts about exactly the same issue but non of them provided any solution or pointed to the problematic area.

Thanks!!!

Has someone actually SEEN the throttle body closed while you are driving and shifting gears? Like, keep engine cover off, passenger stares at throttle body while driving? If its revving up and NOT stumbling or missing while shifting, then it is probably getting proper air and gas, which could only mean ICV (unlikely it can provide enough air flow for redline/high RPMs), or throttle body.

How much flow across the maf are you seeing when the problem happens? The same flow as intended full throttle on the pedal? That answer should debunk a few myths here

The engine runs just fine, no power loss, revs up as usual. I don't think there is lean/rich mixture problem here. Now, the throttle plate does not get stuck otherwise I would redline the engine the second I press the clutch pedal. Rather it's get a little bit of air in to keep the RPMs around 3K when engine is not under load. It totally looks like ICV porblem but I have swapped the ICV already for a new one and the problem persisted. Actually engine's idle became less stable with the new one so I cleaned and swapped back the old ICV yesterday.

MAF registers very little air flow when RPM get stuck, of cause not the same airflow as while I'm driving. Again, if the throttle plate would be stuck I would redline and fry the engine in seconds while shifting :).

Thanks for the reply, still digging in.
I'm investigating AOS/oil fill tube now...

The old oil fill tube was broken in half, installed a new one today but could not get it all the way to the oil filler pipe at the engine side. I read that you can break the pipe that goes to the engine easily and the it's total pain to change it out so I don't want to use excessive force. It stays put just fine and the clamp is holding it in place.

http://986forum.com/forums/uploads01...1392714984.jpg
The old broke old fill tube, I think it's the original '97 one :).

http://986forum.com/forums/uploads01...1392715262.jpg
The new tube is not 100% fit onto the oil fill pipe.

I'll see in a few days if it cured the RPM problem because I was basically driving without the oil cap :eek:. Also I'm wondering is my AOS is not good, the cracked oil filler tube is one of the symptoms of a bad AOS. After I've put the new tube I still can remove the oil fill cap but now the it has considerable higher level of vacuum and the engine dies almost instantly after the cap is removed (cold engine).