Do you mean 6 bolts on CAT or what?I am lost now...

Yes i figured that you might have already exempt...

Yes, there are 6 bolts that attach the rear of the cat to the the heat exchangers. There are 3 in front that attach it to the muffler. I was able to remove and replace the 3 to the muffler but can't get the 6 others off.

I see i got a little confused since you did not mention the 1st 3;-)
What i do is every year replace all exhaust nuts to prevent this kind of problem.Every 2nd year i have to go for smog here too, so it is much easier to the replacement of cat for strait pipe and back.
In your case i would use torch and cut them of .Then get new nuts and bolts.also get new gaskets.

Ivan

Resurrecting this as I've not found the issue yet. In Oct-Nov the air temperature was cool and I didn't experience the "surge when engine hot" issue. Then the car was stored over the winter until April - again cool air. But now May is bringing hot temperatures and I'm driving more spirited trips and the surging has reappeared.

I've verified all of my fuel pressures - all in spec and same as I recorded them last fall. RPM drops when the oil cap is removed. Verified the lambda circuit is working (dwide band well meter).

I still can not get those 6 bolts off the front of the cat in order to get a bung installed for a wide band O2 sensor. I got some 1.5" diamond cutoff blades for my Foredom (shaft rotary tool) so I am going to try to cut these off. I think I'm just going to bite the bullet and do a cat bypass and have the O2 wide band welded to it. None of this solves my problem but might provide more data.

Have a Dremel tool ?

The Foredom is a high tech Dremel - with a shaft. That is what I got the diamond blades for.

Just grind the nuts off :eek:

Working on it! It is very tight even with a handpiece to get in there.

Mine was surging because of vacuum leaks. I'd do the vacuum and soapy water in a spray bottle test. Search and you will find boyt911sc's writeup on it.

I have a smoke tester and have run it multiple times. Last year I started eradicating leaks. At the end of the summer I had eradicated enough that I don't see any smoke. Tony was extremely helpful in getting me to the point (and rebuilt my WUR). This is like a "last mile" problem. I highly suspect rich mixture but I recall (I need to go back and reread my log and also re-do the test) that monitoring the lambda via the test port and a dwell meter, that it actually is not running rich. But that was a static test in my garage and I was not able to perform it when this problem happens.

if you have access to an impact gun just put a wrench on the back and hit them with full impact , they will either loosen off or snap ( which is a good thing . )

I can borrow one but there really is no room to get anything into the top two bolts. I can barely get 1 wrench in there. But I do have a plan with the diamond cut off wheel I'll try to execute today or tomorrow.

It was a nice hot and humid day here in New England today and I got some time to work on my hot surge issue. I warmed the car up to get it hot and then removed the airbox cover and filter to do the lift test. Unfortunately, it seems my mixture is set pretty close - just a tiny lift and I get a near stall and a tiny downward pull also nearly stalls. This is "unfortunate" as I hoped the mixture was the issue.

So I'm back to square 1 and have no idea how to diagnose this issue. I did notice something that I've observed before but didn't really think about - that is, if I get the surging and shut the car off, wait a few seconds and then start up, the RPMs are around 1000 for a few seconds and then start to decrease but then the surge starts gently at first and worsens.

Just to recap, 83 SC with lambda. This issue appeared in Sept last year after I worked to diagnose other CIS issues. I went through the CIS system thoroughly with guidance from folks here like Tony. I got fuel pressures that were a little on the border line so got a rebuilt WUR from Tony. I used a smoke tester to check for and fix leaks, I monitored the lambda on the test port and observe all the proper behavior of duty cycle from cold start to warm up to driving and WOT. I also went through each CIS switch and component to verify proper functioning.

One thing I did find during the original investigation was that my air box had separated along the seam between the top and bottom parts. This was most likely one of the culprits for my original odd intermittent CIS problem. It was intermittent as sometimes the 2 halves were "pushed" together and didn't leak but going over a bump or similar would cause them to separate and introduce unmetered air. I carefully reglued the halves together and sealed the seam (still in the car) and verified that the leaks were sealed with a smoke test.

I have acquired all of the parts including a replacement air box, injector seals, and other tubes and seals to rebuild the CIS system but I simply can't bring myself to doing this work until I've exhausted all possibilities for this hot surge issue. I'm also concerned that such major work could make things worse not to mention the loss of driving time during the short driving season here in the north east.

One near-term thing I am going to do is install a port upstream of the cat to install a broadband O2 sensor so I can monitor air fuel mixture directly (using a 14point7 data logger). The new gaskets and hardware I need to re-install the cat after I remove it to have the bung welded in came in today. I'm going to try to cut the rusted bolts at the front of the cat (6 of them) off tomorrow. The 3 bolts that attach the cat to the muffler were replaced with SS bolts last summer so they will be easy to remove. I've also built an interface to monitor the lambda test port and feed the data into the data logger so I can log the lamda duty cycle and mixture at the same time.

I had nearly the exact same symptoms on my 81 (see thread - http://forums.pelicanparts.com/porsche-911-technical-forum/827097-intermittent-high-idle-lambda-behavior.html) - even observed it looking at the duty cycle with a dwell meter. Duty cycle would ramp up and idle would be annoyingly high. Cool temps - generally no problem. Probably struggled with it for a year or two. Gone now.

Not exactly sure what fixed but I suspect it was a bad ECU or my fuel injector seals allowing a little vacuum leak. This will be a little all over the map but here goes...

What I tried...

1. Swapped back to my original ECU vs. a spare I had in there. I had a spare ECU that was in use when the high idle occurred and I swear that thing was possessed. In an unrelated test, I had tried an alternate ECU (same unit/part #) for reasons I dont recall now. I know people will say I'm crazy but the spare ECU (bought it used) seemed to want to pull the mixture rich all the time (including at idle). Under driving conditions, the car felt like it had a turbo or something. Pulled harder but also seemed to have the surging idle and gave me a notably sootier tailpipe. I had the two ECUs in there stacked for awhile and would pop the plug back and forth to see if any difference. One day in September I was experiencing the surging idle - Shut the car off, swapped the big plug back to my original ECU in probably 30 seconds, restarted car and idle was totally normal. After that I chucked that spare ECU out of there. However since it was September, cooler temperatures arrived and due to work and life, car sat until November at least so was kind of unable to confirm that fixed it.

2. In early spring the next year, I replaced O2 sensor with a Genuine Porsche unit to eliminate the crimp connector for the replacement Bosch O2. I had heat shrink on the crimp connection and have heard of issues caused by not allowing the trace amounts of air in the wire strand (deep topic). Shelled out $120 for the full Genuine Porsche O2 sensor with wire, plug, grommet, etc. This was a total shotgun approach using some Christmas parts action to obtain the new sensor. I kind of doubt this was the issue but was fine with eliminating it as a variable via a fresh O2 right out of the porsche box.

After steps 1 and 2, I can't recall having the issue anymore.

3. After having starting to have some bogging performance about a year and a half later, found my injector seals leaking and changed all sleeves and seals. When hot, hit those seals with some carb cleaner and see if engine bogs down just a tad. If so, they're done - no way around it. Replace sleeves and seals (see my posts on those also). 5 minute test that costs nothing (outside of the can of carb cleaner).

I seem to remember the issue pretty much disappearing after steps 1 and 2 above, but 3 could have been it also. Thinking about it, if the seals were leaking a bit, the ECU may have been trying to compensate by richening the mixture, ramping up the idle. On paper that makes the most sense and if your seals, airbox and such are old, this textbook answer (vacuum leaks) is very possibly the culprit. However, something did change when I went back to my original ECU, so may be worthwhile to swap in a known working ECU (try Tony) and see if any change.

All I can say is its been 2-3 years now and all fine. Drove 8 hours to Canada last fall then piled on hard mountain driving to Mont Tremblant and extensive hot idling stuck in Montreal traffic and all rock solid. Now, its CIS, so if you sit idling and ramp the temps up to 230, the mixture will probably lean a tad and idle may rise a bit, but my surge was something else - just like you have. And I had the same thing on shut down and start up - would be normal again after a quick restart but slowly degrade again to high idle condition.

mhackney,

I have some thoughts.

First, Ivan (proporsche) suggested disconnecting the O2 sensor. Maybe you don’t want to leave it that way, but I think it was a good suggestion for troubleshooting. When the O2 sensor is disconnected it will fix the duty cycle. If the surging is due to the lambda system fluctuating trying to find stoich, this would eliminate that variable by ignoring the O2 input. If the problem goes away, it is a good clue.

I will actually be in Littleton tomorrow night through Wednesday for work. I might be able to meet Tuesday evening, work permitting. Happy to put another set of eyes on it if I can. If this trip doesn’t work out, I will be back in the future. My company is HQ’d in Acton.

Inductive heater..........
 

Michael,

The rusted nuts and bolts for your catalytic converter could be removed with ease using heat. I was skeptical about the effectiveness of removing these stubborn and rusted nuts and bolts until I tried the the tool myself. Heated each nut for 2~3 mins. and zapped it with an impact wrench. All six (6) nuts came off instantly. No broken fastener.

http://forums.pelicanparts.com/uploa...1528014642.jpg

http://forums.pelicanparts.com/uploa...1528014642.jpg

The conductive heater borrowed from a Pelicanparts member:
http://forums.pelicanparts.com/uploa...1528014642.jpg

http://forums.pelicanparts.com/uploa...1528014642.jpg

Another option is to bring your car to muffler shop and let them remove the stubborn bolts. Have a pre-muffler modified with OXS bungs (one for the ECU’s OXS and one for AFR meter). The extra bung could be plugged when not needed.

I have several ECU’s that you could use for your tests. Keep us posted.

Tony

Thanks all. Tony, I sent a PM about the ECU. Hadn't really thought that ECUs could act oddly but it is worth a try given everything else I've pulled out.

Tim, email sent. You'll be about 12 miles from my house from your workplace and much less than that from Littleton. Looking forward to hearing from you tomorrow.

Meanwhile, if I disconnect the O2 sensor to test I suppose I wouldn't have to test for long to observe if it is the lambda system fluctuating. I'm going to try to squeeze this test in today (another high school graduation to attend for my niece north of Boston). It's a bit of a pain to "disconnect" the O2 sensor since the original connector is disintegrated and I wondered the connection. I need to clean all that up and use crimped barrel connectors with a loosely fitted rubber sleeve.

Tony, that inductive heater is quite the tool! I need to check my local friends and see if anyone has one to lend!

cheers,
Michael

+1.

I bet it goes away. If so, then you have to determine what is causing the ECU to pull things one way or another. This is the frustrating part of lambda operation. With everything connected, if you can hook your dwell meter to the test port, I bet you can observe what the ECU is trying to do when this happens. This is so helpful for troubleshooting and setting the mixture. Could be a vacuum leak is causing your ECU to sense a lean condition, to which it responds by pulling mixture rich by increasing the duty cycle.

Mix and timing need to be right. Eliminate those as variables. Set timing to spec. And you have to get mix right. To set my mix, I used a Gunson in the test port on the CAT (o2 disconnected), get it close to 0.6 (understanding there is a margin of error with the Gunson), then I fine tune with it the dwell hooked to the CIS test port (o2 connected). My car seems to like the dwell dithering at about 40 on my 30-year-old Sears meter (maybe just a touch rich) when warmed up. Idle solid with very light fluctuation at 900 warm (+/- 50). In essence, follow the procedure on the label on the underside of the engine lid.

Ok, I'm not intentionally being scatterbrained or ignoring sound advice but...

I had very little free time today so I could only do something quick since I had to drive the 911 to get to my niece's graduation. I couldn't risk doing something that might create a problem to drive to the party 50 miles away with 40 miles of highway and 10 miles of in-city driving on a cool (65°F) and dry day. Here's what I did and the results.

I hooked up my Actron CP7677 dwell meter to the lambda test port using two 6' wires to run into the cockpit. I set the meter on 4 cylinder and then hit the road to get to the ceremony.

Once the car was warmed up - it never went past the first white tick mark on the temp gauge - I was reading 58.6 on the meter. It fluctuated between 57 and 59. This was highway cruising at 80MPH in 5th. From the conversion to duty cycle, I get 65 duty cycle.

Getting off the highway and stopping, and stopping around town at lights, the dwell read 86.7 (fluctuating from 85 to 88) which is 98 duty cycle. At idle the engine did not surge - this is what I would have expected based on the cool day and the fact the temp gauge never got over the 1st tick even though I drove for about 45 minutes on the highway at 80MPH.

So to recap (if my conversions above are correct):

highway cruising: 65 duty cycle
idle: 98 duty cycle

This was repeated on the drive home. Can this data be trusted? If so, what does it say?

Those are big numbers. Not sure you can even trust those with 98 duty cycle at idle. That indicates WAY lean mix, so ECU is trying to compensate radically by richening mixture massively. Not sure how it would even run if it were that lean. But since you saw the dithering, maybe it really is an accurate reading.

Ensure your mixture is set accurately. It could be you are way lean (and/or have a big leak) and the ECU is trying to get you back to stoich. But before you go turning screws like crazy, I hate to say it, you may want to try another dwell meter, one that has the red/black/green wires if the actron doesn't.

As a data point, setting my mix to 0.6 CO last time around on a Gunson gave me an idle duty cycle of about 55-60% (Gunson seems to get it close but not perfectly accurate). I then fine tuned it down to just under 50% by going maybe 1/8 of a turn richer on the screw - that's it - very fine tune adjustment on the screw. I will say I reduced my mix from 2.8 to 0.6 going out about only a half a turn on the screw way back when, so if you are setting mix without proper equipment, it doesnt take much turning to get to extreme settings.