AT wits end- 85 3.2 rough idle

[don911]

I had a similar idling issue once. When cold it would idle ok and then get lazy, bouncy and eventually die. It was a A/F mixture issue. It turned out to be the connector that goes to the AF meter. Two of the contact on mine pushed back and were not making good, if any contact. Simple check, you can pull the boot back and check the 4 wires and make sure the contacts are fully seated in the connector.

[KFC911]

Originally Posted by 84porsche9113.2

After troubleshooting this issue myself, I learned that the ICV's main purpose is to compensate for additional load on the engine such as the a/c, headlights, etc.

There is no need to waste $175 guessing if the ICV is bad. If you do the proper tests, you should be able to determine if the ICV is good. The procedures are in the factory manual.

Part of the test from the manual is to test the ICV is to start the engine with the a/c and lights on and terminals b&c jumpered. The idle at start up should be low b/c the a/c is on. Pull the jupmpers out and the icv should compensate and idle should rise. Turn off the car again, and pull the vacuum hose on the FPR off. Jumper b&c and turn the car on. The idle should be high. When you pull the jumper, the ICV should compensate.

Also the there are electrical tests for the ICV in the manual. Check the ohm specifications. As well as applying 12V (per factory manual) to the outside prongs alternatively, (and grounding the center). Arm inside ICV should rotate to each side alternatively as you move 12V to the other outside contact.

I did some of the tests, but not all (I don't have a/c). I'm not disagreeing with you, but how do you know that the plunger/arm inside has it's full range of motion? My symptoms were consistent, when pulling up to a stop light, my idle would be stable & perfect for around 20 seconds, then I got to practice heel & toe . The electrical tests checked out fine, cleaning the ICV with isoprophyl alcohol helped somewhat initially and I could hear the plunger/arm moving, but it still wasn't "right". I still think if he's got access to an ICV that's "known" to be good, a swap is so easy that I'd give that a try. Good luck!

[JeremyD]

The way it was explained to me by my wrench - basically as soon as the idle microswitch is engaged - then the ICV takes over responsibility for the idle function and fluctuation.

If you don't have any leaks in the intake (and believe me - I did the same thing you did - rebuild my engine - new rubber, etc - i still sprayed a little brake cleaner to cross that off the troubleshooting list. Make sure your vacuum hoses are secure as well - I had one that was about to come off - thought I found the problem, but turned out it had little to do with the idle.

Then the next step is replace the ICV. Hell you probably spent more than $175 in time already.

I replaced mine - and it did smooth out my idle, even when I tested it, cleaned it and assumed it was running fine.

[mackskibum]

Well, after a week to think about this and travel for a while, I am back at this problem. Thanks to all who posted during the last week or so. So to respond to the folks that posted in my absence:

Javers

While I have no way to determine if the ICV is functioning properly, I have to believe that it is making it all the way to the stops when power is applied to both sides- when on the bench it will clunk against the stop on each side when polarity is reversed.

Mark

The AFM is a reman- less than 2500 miles on it. That doesn't mean that its good. To verify, I swapped it with the old one and no difference. I don't think its an AFM problem at this point- although it may be an internal adjustment problem- more on that later.


rallyracer

Not my symptom- twice 2 years ago I got the bouncing idle thing, but it went away. No problem this way since.

ian

At my top end rebuild I replaced all the injectors with remans from a mass rebuilder. I was facing a budgetary crisis at the time, and in hindsight I should have dropped the $800 and went with all new. The ones in the car now are newly cleaned and tested by a local shop that I used to work for. I actually used to run the machine they use, so I am pretty familiar with the process. I have a high degree of confidence that the 4 they called good are actually OK. The 2 replaced with new were due to 1 body leak and one that wouldn't hold pressure without leaking at the pintle. Since the car performed the same with both sets, I am looking elsewhere.

Steve W

I have exhausted the vacuum leak thing unless it is somewhere I can't find external- lets say the ICV won't close for example. The idle is so bad that it has to be an issue- idles like there's a cam in it. No Porsche buyer in '85 would have accepted this.

As far as mixture was concerned, I was trying to "wing it". No success. In the interim, I purchased an LM-1, and did some testing today. I can not get below .81-.82 with the reman AFM installed. This is with the bypass screw dialed all the way in. I need another couple of turns to get this to the .6-.8 you mentioned. Car definitely runs better here that anywhere else I tried.

A couple of procedure questions: Do you set mixture with the ICV not bypassed? I did it with the pins jumped, O2 disconnected.

Could a bad ICV give me unmetered air into the circuit? This could explain my being out of threads on the bypass screw in the AFM.

FWIW, I tried the same procedure with my "old" AFM- results were actually worse- I couldn't get lower than .85. Could this be an internal AFM adjustment? Reindex of the door position to resistance?

84porsche9113.2

Yep, was trying blind- see above.

KC

I may land here, but I'm not ready to make that commitment yet. I would rather complete the troubleshooting process first before condemning parts.

84porsche9113.2

I have done the bench test procedure you describe, and the ICV moves bothe directions. My AC is not installed yet due to the top end rebuild. I have been chasing these issues, and its sure easier to work on it with the compressor and lines out of the way. I am stuck on that part of the procedure.

don911

Connector is OK- pins are in good shape.

Jeremy

I may still have that problem, but it doesn't explain my adjustment issue above.

Thanks for all the help, I think I may be making progress finally. Please keep em coming!!

Dave

[mackskibum]

In another thread, Steve W. Gave me some specs- I need to reset my mixture- I confused Lambda reading with CO%. I'll be back at this tomorrow night.

Dave

[Steve W]

The LM-1 does not measure CO. It displays either air fuel ratio or Lambda, which is the ratio of the current AFR/14.72. You do not want idle Lambda at 0.6-0.8, which would basically be belching fuel out the tailpipe as that would basically be an AFR of 6:1. Idle lambda should be around 1.0, or 14.2-14.7.

Originally Posted by mackskibum

Well, after a week to think about this and travel for a while, I am back at this problem. Thanks to all who posted during the last week or so. So to respond to the folks that posted in my absence:



Steve W

I have exhausted the vacuum leak thing unless it is somewhere I can't find external- lets say the ICV won't close for example. The idle is so bad that it has to be an issue- idles like there's a cam in it. No Porsche buyer in '85 would have accepted this.

As far as mixture was concerned, I was trying to "wing it". No success. In the interim, I purchased an LM-1, and did some testing today. I can not get below .81-.82 with the reman AFM installed. This is with the bypass screw dialed all the way in. I need another couple of turns to get this to the .6-.8 you mentioned. Car definitely runs better here that anywhere else I tried.

A couple of procedure questions: Do you set mixture with the ICV not bypassed? I did it with the pins jumped, O2 disconnected.

Could a bad ICV give me unmetered air into the circuit? This could explain my being out of threads on the bypass screw in the AFM.

FWIW, I tried the same procedure with my "old" AFM- results were actually worse- I couldn't get lower than .85. Could this be an internal AFM adjustment? Reindex of the door position to resistance?



Dave

[mackskibum]

Steve, thanks again for answering. I got the info from my other post that you answered. I will be at the adjustment again tomorrow- mixture and base idle speed. I will post results- hopefully solves my issue.

Thanks

Dave

[Bullet Bob]

Did you check that the distributor's advance mechanism is working correctly? These can get gummed up and not retard the spark very well when the engine drops to idle.

Bob

[mackskibum]

OK- here's the latest. I reset the mixture per Steve W's suggestions. O2 disconnected I am now at 14.1-14.3 AFR at idle. O2 reconnected, I am at 15.7 to 17.2 AFR- lean, rough idle. The interesting thing is that off idle, the mixture falls to 14.5-15.0- a little bouncing, but within that range. As soon as the engine drops back to idle, the mixture is back to the lean condition.

Any ideas??

Thanks

Dave

[Steve W]

What are you using to hold the O2 sensor on the tailpipe? The Innovate exhaust clamp or something else? If a piece of pipe is used, the exhaust reversion will cause false readings at idle. Also exhaust leaks before the O2 sensor can leak air and affect the O2 sensor readings.

[mackskibum]

Steve, we may be on to something here- I have an exhaust leak at #5. I will look at this tomorrow. To get the reading, I was placing the LM1 o2 sensor in the tailpipe by hand and making sure that the cable was bent so it was in open air to the side. I agree that it is not scientific. I will likely order the clamp from Innovate to eliminate the fudge factor here.

FWIW, if you bypass the ICV with the jumper wire, the idle smoothes out some and the readings go back to 14.8-15.5 at idle.

A question- would the exhaust leak cause problems during setup as well? I can get the mixture dialed in OK with the LM1 O2 in the bung for the factory O2, and the idle is pretty rough, but its much worse with the DME O2 connected and in place.

Steve, again, thanks for your help.

Dave

[Steve W]

That won't work at all - not even anywhere close. Either stick it in the factory O2 bung, or buy the Innovate exhaust clamp.

[mackskibum]

Steve, maybe you misunderstood me. I used the bung in the cat bypass to do the settings. I then replaced the OEM sensor in the bung and reconnected the wires to put the system in closed loop. The rough idle only happens when the O2 is connected to the DME. I have a high degree of confidence that the mixture is set correctly at this point, exhaust leak not withstanding. The sensor in the end of the tailpipe was a "sanity check" to see what happened at sensor reconnect. I will order the clamp from Innovate to make sure.

Thanks

Dave

[DRACO A5OG]

Mackskibum,

So it's been over six months, have you found the culprit?

I recently put back my Welmeister Chip, corrected the Fuel/AIr Mixtures using the Connect the Red Lead of my DVM to the O2 Sensor and Black Lead to my Intake Manifold Ground.

DVM reads 0.3 - 0.5.

Base Idle set at 880 RPMs ( 4K DME ) maintained 880 RPM's when jumper was removed.

I thought it was fixed but when I took her out for a test drive, at stop the idle would settle to 880 RPM's then slowly trickles down to 780RPM's (rough idle). She will not stall and AC will not effect the idle ( like it used to when I had the fuel mixture too rich, sometime stalled with the AC on )

Like you I have no Intake Air Leaks.

Replaced:

DME Relay
CHTS
ICV
O2 Sensor
Cat By-Pass with O2 Bung
Distributor, Cap, Rotor & Wires

SW Chip on Standby until I fix this Idle Issue.

Hope you had success,

Jim

[911st]

Maybe re-check the distributor indexing.

[cgarr]

What spark plugs are you running? I tried those fancy pants Bosch plugs once and it ran like crap.

[DRACO A5OG]

Thanks 911st & CGarr

Yeap "Fancy Pants" Platnium 4's. PO put them in there but I did not know that until I serviced the plugs. They were in perfect condition and firing is not an issue. She does not misfire in anyway.

It is the trickle down in idling that is an issue.

[DRACO A5OG]

I think it may be that I am running too lean my DVM auto reads at 2 Volts = d.ddd mV:

My readings show 0.020 - 0.034 mV at the O2 Sensor non BC Jumpered Idle at 810 RPMs.

Will try to bring the mixture up to 0.200 mV+

The wierd thing is when running I assume lean the ICV definitely Kicks in to keep the idle maintained when I turn on the AC or HeadLights or both. When rich it would want to stall when I have both on. It doesn't even blip.

[mackskibum]

Wow, it seems like this has been a hot topic for as while, and I walked away from my original problem over the winter. I am just getting back into this now to try and settle this issue. This has been one of the toughest troubleshooting processes I have ever undertaken.

Where I left it last fall was that I was attempting to tune using the LM1 without the venturi adapter. I have since pruchased the innovate exhaust clamp/venturi and have repeated the tests. I still have the lean idle issue no matter what I do with base idle settings and AFM idle misxture adjustment.

Here's where I am at this point:

Between last fall and now I have replaced another suspect injector from the set I had cleaned last year. The report showed a higher flow by about 7% over the same period of time at 4 bar pressure. I had a spare- the logic was that the flow migh cause an imbalance from 1-2-3 and 4-5-6 that could cause a lean on one side and a rich on the other thus creating a rough idle as the O2 tries to manage the situation.

I found a small vacuum leak at the rubber boot that ties the two intake manifold halves together. This has been corrected, and I can find no other unmetered air leaks in the syste, using a spray liquid at the suspect areas.

I repaired the exhaust leak at #5

I repaired an exhaust leak between the muffler and the Fabspeed bypass

Replaced the ICV with a new unit

Replaced timing chains and sprockets (unrelated, but had to go there due to an oil leak between the chain housing and cylinder head)

As it stands now, using the LM-1, I get approx 14.2 AFR, 1.0 lambda anywhere off idle.

At idle, with ICV bypassed and O2 disconnected, I can get close to those numbers by balancing the base idle adjustment and the AFM adjustment. In this format, AFR and Lambda are 12-13 and .8-.9 respectively- I am assuming this is the richer map in the DME for open loop sensor failure mode.

Reconnecting the O2 by itself makes little change in idle speed or mixture. As soon as I take the test lead off and allow the ICV to come back in, the AFR goes to 15-18 erratic and the lambda numbers bounce between 15.5 and 17, and rough idle returns.

I have verified that the idle switch is completing the circuit at the connector underhood. I have not verified at the DME end of the harness. Full throttle switch is open until full throttle is reached.

I also tested the CHT resistance, and the cold value was 1.087Kohm at 50 degrees F. I did not have the opportunity to test resistance hot yet, but I have a hypothesis and question:

If the sensor is supposed to be 2.2-2.8Kohm at room temperature, and somewhere around 350 ohm hot, and I am starting the process at 1.087K, it would seem to me that I am sending a request for lean mixture to the DME from the get-go. Then as it warms up, and resistance changes, I am still sending a lean request because the resistance curve is functioning at half of the specified resistance. Is this correct? - That is as temp increases and resistance decreases the fuel bias goes lean?

I am hopefully finally on to something here- it would be nice to get this project behind me and actually be able to enjoy this thing when sitting in traffic or at a stoplight.

FWIW, the CHT is relatively new- less that 2K miles, and I had discounted this as a problem. Apparently NEW stands for Never Ever Worked!

Thanks for reading- I welcome anyone's opinion on the newest information I have posted.

Dave

[burgermeister]

O-ring on oil cap? Per a previous post, I replaced mine (quite hardened) and idle smoothed out noticeably.