My CIS is a complete mess and I need help!

[mhackney]

Ok, there is a long painful back story here. It starts with this thread that I posted on 4/28/2024). https://forums.pelicanparts.com/porsche-911-technical-forum/1160979-its-time-air-box-replacement.html#post12240127

So I had a backfire that separated my air box back in April when I was first bringing the car out of storage. (there was no pop-off valve in the air box) In the thread above I contemplated replacing it myself but I just about to start a new job that would have big time demands so I decided to take it to a local and reputable shop, one that has been around for decades. I had the car flat bedded there on May 3rd along with a bunch of parts including:

* 911 110 904 00 - airbox
* 930 110 197 12 - intake manifold gaskets (6)
* 930 110 885 00 - intake sleeves (6)
* 999 701 423 40 M213 - injector O-ring (6)
* 911 110 886 03 OEM - insert sleeve for fuel injector (6)
* 999 701 446 40 M17 - injector sleeve O-ring (6)
* 999 701 210 4A - cold start injector O-ring (1)
* 999 701 395 50 - cold start injector spacer O-ring (1)
* 999 701 124 40 M17 - throttle body O-ring (1)
* 911 110 394 02 M30 - coffin gasket for sensor plate housing to airbox
and a pop-off valve

I asked them to let me know if anything else should be addressed as this was the first time the car was being looked at by a pro since I've owned it (both in 2008).

A week later, the shop called to check in. They found that the air box had been replaced at some point before I owned it and had loctited the nuts on the intakes. And the intakes were coated with a lot of carbon deposits. [sidebar: the car was running very well for the last few years and I had installed an AFR gauge years ago to keep an eye on AFR and it has never run exceeding rich]

At this point they adjusted valves, did a compression test (I don't have the numbers but it was fine. Engine has 191K miles on it). And did a lot of the other replacement parts on the list and the typical triangle of doom parts and installed the pop-off valve. He also replaced the distributor cap and rotor. Several weeks later they called and had installed the air box and had things back together and were finishing up tuning it.

And that's where this story starts...
On 6/17 I went to pick the car up and was told that it wasn't running as well as it should but that I should put 100+ miles on it with some good gas (remember, it had been sitting over the winter and I had not put gas in it yet this year). Driving home was a nightmare with lots of back fires - both exhaust and in the air box, I could hear the pop-off valve working. It was also at this point that I noticed my AFR was read "---" (this is an AEM gauge and 3 -es typically means out of range too lean. The next day I decided to plow through a 100 mile drive and was able to do that with some minor back firing, especially when decelerating. But it was not a pleasant drive. The shop agreed to come and pick the car up on Sunday (May 19th), which the did.

The car was in the shop from then until last Friday - October 4th - and still was not running well. But I wanted to get my car back and go through it myself. When I picked it up, they told me that they had replaced the OEM O2 sensor and noticed that when it was disconnected, the AFR gauge worked. (I have not verified this myself). They thought maybe the WUR was faulty but after testing, it turned out ok - but I did not get the test data. The car would cold and hot start "fine" but unless I give it gas immediately after starting, the RPMs drop and it dies. When cold: giving it gas and it won't idle below ~2000 rpm. Keeping it above 2000ish RPM keeps it running but it feels rough. I had some back fires on the drive home (15miles) and multiple stalls when stopping at lights and signs, but it always started right back up as long as I fed it gas as soon as it started.

I recalled that I may have spliced the power and ground wires for the AFR under the passenger seat on one of the relays. I took the seat out last Saturday but found that I had maybe spliced the wires there but had rewired at some time in the past.

So today (a week later), I started the process of going through the CIS system to try to resurrect my baby. It has been almost 10 months since I've really driven it so I've forgotten some of the basic things like "does the fuel pump come on when you turn the key to run" and "when do you hear the FV".

So let me start with the basics, I started by doing a fuel pressure test. Here is my data in the Pelican-approved format:

1. Year of engine: 1983
2. US or RoW (Rest of World): US
3. WUR model number: 0 438 140 090
4. Ambient temperature at time of test (in degrees C): 19.4°C
5. WUR Resistance (in Ohms): 25Ω
6. System Pressure (in bars): 4.8 bar
7. Cold Control Pressure (in bars): 2.25 bar
8. Warm Control Pressure (in bars): 3.5 bar
9. Time delta for Cold -> Warm Control Pressure: 3 min
10. Residual Pressure @ 5 min (in bars): 2.25 bar
11. Residual Pressure @ 15 min (in bars): 1.7 bar
12. Residual Pressure @ 30 min (in bars): 1.5 bar
13. Residual Pressure @ 60 min (in bars): 1.5 bar

I had taken the heater fan and large duct out to make some room and I noticed that one of the "CIS connectors" - the green one, was not connected. Green had 2 brown wires, so it is verified to be the AFS switch connector. I plugged that back in. But that did not make any difference in running. I then did a smoke test with a homemade smoke tester and it seems to be very tight, no leaks that I could find anywhere. The shop had replaced the rotten old lines.

Now I need help trying to figure this out. It bothers me that it cold and warm starts with a single key click (no cranking required) but needs gas to keep running. And it runs rough.

One of the few things in the CIS system that I don't have experience with is the FV. I plan to test the relay under the passenger seat - I had put my hand on it when I turned the car on and did not feel the relay activate. Normally you can feel these types of relays. This relay feeds power to the ECU and to the FV. When I turn the key to "run" there is no sound. I don't recall if I should hear the fuel pump and FV in the run position, I seem to recall that the fuel pump did run.

Anyway, I'm out to the garage to test that relay but wanted to write this up in the hopes it might trigger a memory!

[mhackney]

I may be on to something. I tested for voltage on pins 87 and 87A on the FV Relay (aka OXS Relay) with the car running and there was none! I unplugged the relay and there was no change in startup and running behavior.

Fortunately, I have a replacement (correct) relay in my parts that I got years ago after reading a list of parts one should have on hand! Bosch 0-332-019-109

Now, things are very different - not right, but different I think in a better way.

The car started up and I did have to give it just a little gas. But then I could take my foot off the gas and it idled at 2000ish RPM and sound much stronger. Then after 30 seconds (or less) the idle speed started cropping up and I noticed that my AFR was showing in the 17s. I gave it some gas and the AFR came down to high 14s and crept back up and back to "---". Letting my foot off the gas to come back to idle resulted in a larger back fire in the air box - enough to blow the gasket out of the pop-off valve. I shut down and with the key completely off I heard this odd clicking sound from the engine bay. I made a recording of it. Here's the link: https://youtube.com/shorts/66hDgPFD93A?feature=share

Now, a few minutes later I just checked and it is not making any noise.

So, is it possible that the FV relay had died and the shop had not checked that and instead, tried to adjust the idle speed and mixture to richen it up? My understanding is that with lambda cars, the FV is constantly adding fuel to a lean mixture. So if there is no FV adding this fuel, perhaps the shop tried to compensate with the mixture?

In any event, where should I go from here?

[mhackney]

I just went out to startup again and record things. Here is the video: https://youtube.com/shorts/-2tkhedWa_I?feature=share

As you can see, right after startup (and I did have to give it just a bit of gas) it idles at 1800RPM and then at about 30 seconds, it jumps up to 2900RPM and my AFR gauge starts to read in range.

So, I do think I'm heading down the right path. I believe the dead FV relay was the root of the issues that kept me and my car apart for all these months. Now I need to tune the mixture and set idle speed and set the lambda duty cycle. All stuff I've done and forgotten so I'll need to do a lot of reading. Any directions greatly appreciated!

[E Sully]

I have no experience with late lambda systems, but since all your pressure test numbers look good, could you be having issue with the Aux Air Regulator or Aux Air Valve?

[mhackney]

I don't think so Ed - the smoking gun was the inoperative FV relay. Although not impossible it is improbably that multiple components all failed at once! I think I'm just dealing with basic idle and mixture tuning at this point. That's the plan for tomorrow.

[john walker's workshop]

Lots of shops don't know about the FV system. When it fails, usually due to the relay, wiring at the relay, or a popped interior light fuse, the system goes way lean.They crank up the mixture adjustment to get it to idle better, but the engine still runs like crap.

[AndrewCologne]

I'm with John here, unfortunately even many 911 shops don't know anything about the FV system.

First - get a DMM with Duty cycle reading option.

The green plug behind the CIS is responsible for a proper working of the fuel pump safety circuit. If not connected, the fuel pump and therefore also the FV and WUR will be active when key position of ignition switch is on. So you can keep it unplugged for checking purposes, but after that workaround you should re connect it again.
--> So when you hear the fuel pump working, the FV should be buzzing, if not, ... check for proper voltage at the terminals of the plug of the FV. If voltage is ok, then check the FVs resistance, should be between 2 and 2.3 ohms. If not then replace FV for testing.

I did not read that a proper co setting has been done. So when FV is ok, unplug the OXS sensor and start the engine, with engine warm set – for testing purposes the – the CO to 1.2 %! Not 0.4-0.8. With this setting drive the engine and tell us if it runs better or even ok.

Then at idle, re connect the OXS sensor plug in the left engine compartment and check the Duty cycle at idle using the DMM -- explained here:
https://nineelevenheaven.wordpress.com/the-911-sc-3-0-engine-with-catalyst-and-lambda-control/
Read the part "The lambda control – display and check", there its explained how to connect the DMM in a proper way.
Then tell us what duty cycle is present at idle (with still the initial co setting of 1.2% as mentioned above)

[mhackney]

Thanks John and Andrew. I’ve followed John for decades and I’ve studied Andrew’s site for hours!

As of my last post the FV is now indeed working but the idle speed is way high. I have not set the CO yet. I do have an oscilloscope and a DMM with duty cyle/dwell and have measured the duty cycle in the past. I also have an AFR gauge (AEM). I am not sure what the next steps should be given that my idle and mixture are likely way out of whack. I’m thinking I should see if tweaking the idle can get it down below 3000rpm when the engine is warm then set CO using the duty cycle method as described. I can verify that with the AFR too. I believe things are so out of range because they tried to tune with the FV and ECU inoperative due to the faulty relay.

[Dpmulvan]

Found this for you

[mhackney]

Thanks for that video link! I've been working on my car's CIS for 16 years and thought I had found all of the good stuff on CIS! I learned a couple of things from this video for sure, not the least of which is the function and adjustment of the Deceleration Valve.

Wow, is working on the CIS on an engine stand MUCH easier! It is such a pain to get to those components in the car.

I had removed and verified proper functioning of the AAR back in 2016. (I keep a detailed log of everything I do and links to references on my car). I can check it again.

Today is rainy and 48° and my garage is unheated and the tail of my car is at the garage door side so working today is probably not going to happen unless the rain lets up some! I have a few other minor projects to finish like cleaning and lubricating the tracks on the passenger seat before reinstalling it, replacing the front hood shocks, and upgrading to H4 headlights. All stuff that was on my to-do list this year but wasn't able to get to since the car was in the shop from May to October

Based on all of the reading here, Andrew's page, stuff I saved from Jim's CIS page, etc I think the first step would be to adjust the mixture. This was probably the adjustment most out of whack because the shop tried to compensate for the non-functioning FV. Does that make sense? Given that I don't have a gas analyzer but do have a scope and in-dash AFR should I unplug my OEM O2 sensor and set the CO to 1.2 % as Andrew describes - but how do I target 1.2% with no gas analyzer? I am running E10 here in New England also.

[AndrewCologne]

Originally Posted by mhackney

Thanks John and Andrew. I’ve followed John for decades and I’ve studied Andrew’s site for hours!

As of my last post the FV is now indeed working but the idle speed is way high. I have not set the CO yet. I do have an oscilloscope and a DMM with duty cyle/dwell and have measured the duty cycle in the past. I also have an AFR gauge (AEM). I am not sure what the next steps should be given that my idle and mixture are likely way out of whack.
... ... I think the first step would be to adjust the mixture. This was probably the adjustment most out of whack because the shop tried to compensate for the non-functioning FV. Does that make sense?

yes – as it was described in my last post above ...
- With disconnected OCS Sensor (forces ECU to a duty cycle of 50%) set CO to 1.2%
- With engine warm re connect OXS sensor and read duty cycle with DMM

I’m thinking I should see if tweaking the idle can get it down below 3000rpm when the engine is warm then set CO using the duty cycle method as described. I can verify that with the AFR too.

No, do the steps as mentioned above, as with following the steps above you can see how much duty cycle is needed to bring the combustion to result in lambda 1.
Also check if your ignition timing is set to correct 5° BTDC – with both vacuum hoses at the dizzy disconnected and ends of hoses closed/plugged to avoid false air. Dont use the reterd timing vacuum for now, .. this will be acrtivated at the very end if all other things are working well.

This was probably the adjustment most out of whack because the shop tried to compensate for the non-functioning FV. Does that make sense?

I guess that as well - a typical foolish way of "experts" trying to bypass a non working FV.

Given that I don't have a gas analyzer but do have a scope and in-dash AFR should I unplug my OEM O2 sensor and set the CO to 1.2 % as Andrew describes - but how do I target 1.2% with no gas analyzer?

Setting the CO using the duty cycle method id only if all is working properly, which with your setup now it doesn't.
Do let a shop make the measurement with a calibrated CO analyzer. My clear recommendation.


Another question: How is your AFR gauge/controller connected? Do you have a separate connection soldered to the exhaust pipe? Or do you use a bracket at the end of the tail pipe? Or do you even provide a simulated narrowband signal from the AFR gauge controller to the ECU?

[Aztim]

I chased keeping my CIS

tbh I regret trying to keep it

go ITB or pmo, rip the bandaid off and ditch that antiquated system

[mhackney]

Thanks Andrew. So my challenge is going to be finding a gas analyzer. I can't drive the car and I am out in the country where there are no shops that can do this on my car.

I have a dedicated wideband O2 sensor installed in its own bung (what you are calling a connection) that is only connected to the AFR gauge. The car has the OEM narrowband O2 sensor for the lambda control.

[mhackney]

Aztim - I've had minimal problems for 16 years and love this car. Thinking through this issue and timeline, I think what happened is that my FV relay had failed while the car sat over the winter and when I first started it back in late April on a cold, humid morning, the car backfired and blew my airbox. Part of that is shame on me for not having a pop-off valve installed. But, as it turns out, the PO had replaced AND repaired the airbox previously (maybe just repaired). I'm sure if I had installed the airbox myself I would have found the issue with the really working through the CIS troubleshooting steps like I did this week. I just didn't think I had the time to do an engine drop myself given my hectic schedule.

[AndrewCologne]

Originally Posted by mhackney

I have a dedicated wideband O2 sensor installed in its own bung (what you are calling a connection) that is only connected to the AFR gauge. The car has the OEM narrowband O2 sensor for the lambda control.

Ok, but if the Orig OX sensor is not connected to the ECU but to your separate AFR controller, then the ECU will switch to open loop control, resulting in a static 50% reading with a warm engine and a warm OX sensor.
This means that it with this it is not possible to measure a correct AFR or lambda value with your external AFR/controller.

[Dpmulvan]

Some more from the peanut gallery

[mhackney]

Andrew, just to be clear - the original OX sensor is connected to the ECU. It always hqs been. A few years ago I simply installed an additional wideband OX upstream of the catalytic convertor that is dedicated to the AFR gauge that I installed below the dash. The original lambda system is exactly the way it came from the factory.

[AndrewCologne]

ok, thanks for clarifying, ... I was not sure if I understood that correctly.

[mhackney]

No worries Andrew. Other than the fact that the AFR is not as precise as an exhaust gas analyzer, could I at least use it to get much closer to a proper mixture and then use the duty cycle method to dial it in?

As it is now the car starts up both cold and warm easily. If cold, it idles at about 1700 RPM and then after about 30 seconds it increases to 3000 RPM. If warm start, it goes to 3000 RPM idle almost immediately. At 3000 RPM my AFR gauge is showing a VERY lean condition - 16.0 AFR. I should note, when the car is idling at 1700 RPM on a cold start, the AFR gauge is not giving a proper reading, it shows three dashes "----" which is an "out of range" condition. I recorded the entire cold start sequence and show the AFR gauge as well in the video I posted: https://www.youtube.com/shorts/-2tkhedWa_I

It is only a minute long but shows pretty clearly what I am dealing with. And if anyone is interested, this is the AEM gauge I bought with the 4.9 LSU sensor. In the video, you can see I installed it in the console in my cabriolet with a 3D printed mount: https://www.aemelectronics.com/products/dashes_and_gauges/wideband_gauges/classic_wideband/parts/30-4110

[Funracer]

My set up is similar to yours. 83 with full CIS with functional Lambda. If you are above 15C as you stated the only thing that changes at about the the 30 second mark is the O2 sensor(s) heat up enough to start doing their job. The - - - is displayed until the sensor is warm enough to function. Until then the FV is defaulted to 50%.
In your setup (and mine) this AEM sensor is a proxie for the factory O2 so we can assume that the factory O2 begins working at the same time your indicator turns on. So when your numbers pop up on the AEM your OE system is now controlling the FV.
The FV can make pretty large mixture (an thus idle) changes if CO is way off. Either your system is super rich and bogging (low idle) until the FV leans the mixture, or super lean (again, low idle) until the FV enriches the mixture and it surges to 3000.

I may have missed this, but at about the 2-3 minute mark (this from an overnight dead cold start) the WUR should be at full WCP and the AAR should be fully closed or nearly so. What happens to your idle if you wait for those two things to occur?

Regards