1983 911SC - Cold Start Problem

[sigchuck]

Hi folks, I'm experiencing an intermittent cold start issue with my '83 911. The engine is newly rebuilt (case split, top end, cylinders honed, etc.) with approximately 800 miles on it. This included new plugs, wires, cap and rotor, but the car does appear to have the original coil.

I have been driving the car fairly regularly, but after it sits overnight or longer, I occasionally have a cold start issue. The car will fire without any issues then nearly immediately die. It's intermittent, and probably occurs one out of every seven "long-term" cold starts.

I can keep it running by giving it some pedal, but there will be pops as the engine revs (no actual backfires though). The idle doesn't really smooth out until I can get it into gear and get it rolling, after which point it's fine.

The decel valve has been disconnected and the vacuum line is plugged. I haven't done a legitimate vacuum leak check, but removing the oil cap does cause fluctuations in idle speed. I have checked the lambda relay under the passenger seat, and while it's original (dated 1982), it doesn't seem to be the problem as the car runs noticeably rougher without the relay connected.

I have found a wealth of information on the CIS system, but was wondering if anyone had any ideas as to where I should look next. Thanks!

[T77911S]

probably too lean.
check fuel pressures first.

you can have someone start the car while you lift up on the sensor plate to see if it idles better when cold. if so you are lean.
fuel pressure
air leak
mixture setting.

has the mixture been set after the rebuild

[sigchuck]

As far as I know, the mixture was set. The rebuild was done by a very reputable regional Porsche shop, but I will verify.

Would a mixture codition cause intermittent issues? The majority of the time it starts fine, momentarily idles around 900-950, then picks up to 1200 or so while the car warms up.

[Jameel]

If you'll be working on the CIS from here on out, you owe it to yourself to read through the Dummies thread , get the gauges, and measure all the pressures, then read up on how everything works. It's a lot to tackle, but I broke down and studied it when I did my rebuild. Now I can pretty much figure out what's wrong with just a little refresh reading. I work on two CIS cars and both run really nicely. It took some doing and study, but now I like CIS because it seems really stable once its set up properly.

[T77911S]

CIS systems are old. so they can have little "personalities" to them. not saying to not try to fix it but an intermittent issue can be really tough, time consuming and perhaps $.

cold be things like:
residual pressure
flakey fuel pump relay
pop off valve not seating properly
injector(S)
mixture
fuel pressure
lambda relay- yes you looked at it but remember it is intermittent. bad solder joint on it.
wiring. green wire from dist to CD
I am sure the list could go on.

if it starts at 900 or so then goes up to 1200 I would really look at
fuel pressures
mixture
residual pressure
injector

probably in that order

my bet it is too lean.
do the test I said above. see if the idle jumps up to 1200 faster and idles better if you richen it up by lifting up on the sensor plate

[sigchuck]

Thanks for the help guys. One thing I have noticed is that when the engine is cold cold (hasn't run in 12 plus hours) it starts fine if I turn the key and hear a buzzing (almost like the sound of the electric antenna raising, but quieter).

I didn't hear the sound yesterday morning when it was difficult to start, but I did hear it this morning and it started and ran fine.

Any idea what this buzzing is? I don't hear it while the car is running.

I have ordered a set of gauges and will also try the sensor plate test. Thanks again!

[T77911S]

you say you don't hear it when the engine runs, it is possible the fuel pump is running with the key on and engine off, priming the system to help it start better.

my fuel pump in my 77 was noisy and I could hear it with the engine running.

try to locate the sound with key on and engine off.

[Paulporsche]

You don't say where you are located, but a lean condition could cause those symptoms intermittently depending on varying ambient temperatures.

[sigchuck]

Would the fuel pump typically prime the engine with the key on? The car does have a new (chinese) red fuel pump relay, but I do have the original still as a spare. Would it be worth swapping this out to see if it's the issue?

I'm located in Dallas. Morning temperatures have been in the 50s recently.

[Sub8]

Hi - assuming you have CIS lambda; the vibration sound in the engine bay at key on is the frequency valve. At the back of the engine behind the throttle body assembly.

This will affect overall fueling hot or cold.

[brad951]

Hi - I am working on my '83 SC (130k miles) as well, diagnosing starting and warm-up issues. I thought I would add my experience here instead of starting yet another CIS issue thread.

Symptoms:

Cold start (58 degree ambient): Starts immediately, idles smooth enough, but in the 750-900rpm range for the first 20 seconds or so, then idle lowers to 500-600 and idles rough. Giving it throttle during this rough phase causes pop-off valve backfire (definitely lean). Idle gradually improves as it warms up, and after 6-8 minutes the idle is smooth at 950 rpm, revs freely no backfires, all good.

On to the data and and measurements, as I had the pressure gauges hooked up this whole time.

- It's the -090 WUR (correct for 1983 model). Resistance of WU circuit = 25.4 Ohms (correct)

- Cold control pressure (electrical connector disconnected from WUR), engine off, FP jumped = 1.1 bar (a little on low[rich] side but that's ok)

- Warm control pressure (electrical connector connected to WUR), engine on, warmed up = 3.5 bar (correct)

- WUR takes about 1-2 minutes to from 1.1 to 3.5 bar (ok)

- System Pressure = 5.0 bar (correct)

- Frequency Valve operating (pulsing, rattling) - correct

- AAR checked and is operating. Takes about 5 mins to fully close.

- Retard vacuum line is disconnected and plugged at distributor.

- Idle vacuum, engine warm and idle at 950rpm = 14-15 inHg (low)

Here's the behavior mapped to pressures and time:

Initial cold start is good, but high idle (>1000 rpm) is not happening. The AAR is functioning, so it's strange that the idle is not higher, but smooth initial idle suggests engine is happy with the rich condition the 1.1 bar initial control pressure provides.

As the control pressure increases, the idle quality degrades and RPM drops. This starts when the control pressure climbs to about 2 bar and continues as the pressure climbs to is warm pressure of 3.5 bar where the pressure stabilizes. This suggests the engine is running very lean during this time and is not too happy. The backfires when throttle is applied confirm the lean condition.

After several minutes of this rough 500-600 rpm idle, the idle starts to smooth out until it is actually very smooth at 950 rpm and the engine seems to be running perfectly. I assume that the lambda system has taken over at this point and the frequency valve is now in closed loop controlled by the Lambda ECU and fuel mixtures have been brought to a happy place. Presumably this phase is triggered by the 15 degree temperature sensor on the right side chain cover ?

So what I do know:

- WUR is correct and functioning within spec
- AAR is correct and functioning within spec
- Control pressures are good
- Idle speed is in spec
- Lambda system is operating properly, at least once in closed loop mode

What I don't know:

- Why the several minute lean condition during warm up? What is responsible for adding fuel during this period ? Is the Lambda system not going into closed-loop mode early enough ?

- Why is the AAR not causing a faster warm-up idle ? Is this reason for the lean condition during warm-up ?

What I suspect:

- Vacuum leaks. 14 inHg idle vacuum is on the low side. During warmup, possibly the vacuum leaks are larger than the AAR port, rendering it ineffective ? And when warm, the Lambda system compensates ?

Next steps:

- Inspect thoroughly for vacuum leaks.
- Measure the Lambda system FV duty cycle in both open and closed loop modes with dwell meter.

I would love any feedback on this, and suggestions on what I am doing wrong or additional things I might consider, and I hope this helps someone out there !!

/Brad

[pmax]

Brad,
That's some thorough data collection !

Originally Posted by brad951

As the control pressure increases,

- WUR is correct and functioning within spec
- AAR is correct and functioning within spec

Yes, agreed.

- Lambda system is operating properly, at least once in closed loop mode

Couple of ideas to try and things to check during the warmup stumble ... disconnecting the O2 plug, check if the FV is working properly.

Hope this helps.

[brad951]

Originally Posted by pmax

Brad,

Couple of ideas to try and things to check during the warmup stumble ... disconnecting the O2 plug, check if the FV is working properly.

Hope this helps.

Thanks for the advice, I will definitely try that.

I am also dealing with a hot start issue. After the test I just described, and the engine was warmed up and running smoothly, I shut it off and came back 90 minutes later and restarted. It started immediately but idles very rough for several minutes, it even stalled once. Control pressure was max at 3.5 bar from the start as the WUR was still very warm/hot. Residual fuel pressure before starting (i.e. after 90 minutes of sitting) was 1.5 bar which is in spec. It idles rough at about 450-500 rpm for several minutes and gradually improves until it comes back to normal running. Not sure what's going on here, but I'm sure it has the same root cause as the cold warmup stumbling.

[Bob Kontak]

Originally Posted by sigchuck

Would the fuel pump typically prime the engine with the key on? The car does have a new (chinese) red fuel pump relay, but I do have the original still as a spare. Would it be worth swapping this out to see if it's the issue?

No, Sir. No priming. Start the car cold and cold start valve sprays a good deal of fuel into the air box. Car starts and a second later fuel pump catches up and pressurizes system to meet operating demand.

Not like modern cars with that one second fuel pump run to jack in pressure.

Fuel pump on the 83 only initiates when starter is engaged.