Rough idle at startup

I've searched the forum, and I've seen plenty of threads on rough idles, but none that specifically address my situation. Any help is much appreciated. On cold starts, my 996 initially idles at 1,000-1,200 RPM for 5 seconds. Then it will dip down to 700 RPM, then it will briefly drop to ~400 RPM, and just when it feels like the car is about to die, I hear a click that sounds like the a/c compressor or some other component came on. At that point, the idle raises to ~1,000 RPM and stays there for good. It idles a little rough, and the idle moves a little bit (less than 100 RPM) until it gets hot. Once I drive it around for a bit, the idle gets much better. I have just replaced the plugs and coils. Also, I just cleaned the MAF and throttle body. Any ideas what this could be? Thanks

It could be an intake leak... Air is leaking after the maf into the TB.

Quotes source post: 9631536

Was the engine behaving this way before you did the work? My first impression is there's something wrong with the secondary air injectionn system. IIRC, if the car is one of the 1st models it has a ICV (idle control valve) which has to be suspect. Then a malfunctioning coolant temperature sensor or an air intake temperature sensor (part of the MAF) becomes a suspect too. 'course, our old friend the AOS comes in for suspicion too. Or a cracked oil filler tube -- air intake leak -- or even the oil filler tube cap which can leak. In fact there are 'lots' of suspects. You at least need an OBD2 data viewer to look at various sensors/inputs while this behavior is occuring. Also, unless you are willing to get your hands dirty you need to take the car to a qualified shop for a pro tech's expert input. Even if one of us guesses the source of the problem that ain't going to make it go away. Sincerely, Macster.

Thanks for the ideas. Yes, the idle was probably a little worse before I cleaned the intake/MAF. A few questions: 1. Where should the 996 idle at? 2. How much should the idle move? 3. Does the fact that the problem goes away once the car is driven for a bit make you lean in any direction? I have taken it to my Indy, who did coils and plugs. They actually delivered it to my house (~20 miles away) and said the car was driving fine. They are a Porsche specialty repair shop in the New Orleans area. Thanks again

Quotes source post: 9632444

IIRC the idle speed is 740 rpms +/- 20. IIRC. Nothing about the fact the engine gets a bit better after it has run some helps me narrow down the list of suspects. In fact, I think I can add one more suspect to the list... The sudden change of idle speed and the severe degree the idle speed changes does sort of sound like what my 02 Boxster did when what proved to be the passenger side VarioCam solenoid/actuator went bad. No CEL but there was a pending code which pointed to the solenoid/actuator. But the symptoms came on suddenly with the engine up to full temp, after a 30 mile drive at freeways speeds from home to the office. The next day after the engine cooled down overnight I do not recall any odd behavior. In your car's case, the solenoid (or actuator) could be "bad" when they and the engine are cold and for a short time after the cold engine is started then the click is the solenoid/actuator working again the improved idle is the result. VarioCam is used at engine start/cold idle to help reduce cold start emissions. If this isn't working right the DME can have problems fueling the engine until the O2 sensors get hot enough to work and then the DME will really move the short term fuel trims around as it seeks to compensate for a bad or subpar VarioCam system. More things to consider. Sincerely, Macster.

I think I may have solved the rough idle. My question is... does this make sense? I found that the secondary air injection pump hose did not have a hose clamp where it attached to the pump. I added a hose clamp, and the idle seems better so far. My theory is that the secondary air injection valve is only open when the engine is cold. Because there was no hose clamp, when the valve was open, there was a vacuum leak present. Once the engine heated up, the valve closed, which effectively eliminated the path to the vacuum leak. Does that sound like a possibility, or am I missing something. Thanks again

Quotes source post: 9664521

Sounds plausible. The real proof will be if the engine remains symptom free. Sincerely, Macster.

I've had basically the same symptoms. Never concerned me much. My last car (Toyota) did the same thing. I'll see if, one day, I can replicate your fix and I'll post my findings. Any pics?

My symptoms have been getting worse, sometimes even stalling out completely. The problem came to a head yesterday evening when the engine light came on. I don't have a reader to see what the car thinks happened. Anybody have any additional advice before I take it in to the shop? I could put a hose clamp on myself, but I need to wipe the code and make sure nothing else is up.

Quotes source post: 9925124

Knowing the engine code or codes and having the freeze frame data would of some value. If you read this data do not erase it for the shop will want to read the data for its diagnosis. The code could be due to something serious that the old clear the code and drive the car some more to see if the code comes back may not be the best course of action. Sincerely, Macster.

I will bet you dollars to donuts your code is Engine Misfire. When it's choking bad enough to die, it almost always registers a misfire. Welcome to my world, this cold engine stumbling/rough idle has been with me for over a year now.

Quotes source post: 9633685

This is the first time I have been made aware that the Variocam does anything different on engine start. I thought it had only two positions. One for below 2950 RPM, and one for above. Are you saying it moves to the top position on engine start? I have found in the last year or so that I can "feel" the variocam change more than I should. However, I've just doubted myself as being overly sensitive to it. Can you give any more details to what exactly the variocam does on cold starts?

Variocam does indeed operate differently based on operating temp. For the original variocam it has only two positions on and off. With your car Sean and variocam plus it has infinately adjustable positions depending on many operating conditions. Aside from a failure condition, on engine start it is going to be OFF at idle speeds and won't kick in until above 2500-3000 RPM ish. There is a great technical read about variocam here: http://www.renntech.org/forums/topic...plus-a-primer/

Thanks for my lunchtime reading material!

Quotes source post: 9925516

P0306 code The funny thing is, the car started great this morning, for the first time in months. I'll put a hose clamp on the SAI tonight, like CMWebber, and report back in the morning.

Hah, Sean called it. Going back to Macster's post: Assuming no vacuum leaks... If this is a 1999 car check resistance on the IACV (and maybe just clean it): http://www.renntech.org/forums/topic...new-cel-codes/ If this is an egas car (post 1999 or 99C4), then I would think your throttle body is suspect. Do you know someone with an egas porsche and swap the throttle body, see if the problem goes away? Also reset throttle adaption. duh sorry see you have targa, so forget the IACV stuff, egas T/B

so who's buying me donuts?

they serve donuts up there? are they oil flavored?

no, MAPLE

Possible development: let me give you the whole story... The ongoing symptom was a bouncing idle for the first 10 seconds on dead cold start-up. As the weather got colder this season, it actually stalled out on me a couple times. I tried using the accelerator to keep the revs above 2k, but below 3k just 'cause i figure it's hard on a cold engine to rev high. Usually this failed, as giving more gas actually compounded the problem. So, on Tuesday evening (a pretty cold day here) I went to my car in our parking garage at the end of the work day, where I had parked on a slight hill. I let the car coast down the incline a little and then started the car (with the ignition, clutch depressed) but then dropped the clutch once started, just to see if it would hold the idle at around 1200. The car chucked a little and that's when my engine light came on (with a c6 misfire). Wednesday, the car started great. I took the car to a fellow Rennlister who runs a non-porsche shop across the road. He checked the code and wiped it. I applied the hose clamp last night as Cmweber did. Car started great this morning. So, now I don't know what's up. Maybe the clamp worked. Maybe I scared the squirrel nesting in my car (see my other thread about rent) and he jostled a wire. Anyway, this thread may have solved my problem so I finally manned up and paid for Rennlist Membership. Thanks all for your help. I'll continue to post if the problem reappears.