Cold start issue with my newly refreshed CIS

[Jameel]

Just finished refreshing my CIS. It's sitting on a freshly rebuild engine that's running great.

Here's my issue. Cold starts are usually fine. But maybe every 5 cold starts I'll get really low rpm, like around 200, so the engine is sputtering like it's starved for fuel, or getting erratic fuel delivery. I just finished testing my entire system except for a couple components. These components were just put back onto the CIS, since they didn't exhibit any faulty behaviour before the refresh.

I've got a lot more experience with CIS than I did a few months ago. This feels like it's being starved for fuel, or maybe running really lean at startup. My WUR tested fine.

Here's the kicker. If it does this cold start-low rpm thing (and I've tried this a few times now--works everytime) I immediately shut it down, then fire it up again. On the second start it revs right up to high idle (1000-1100) for a minute or so, then once the WUR warms it settles back to 900.

[timmy2]

Lambda system getting bad readings from O2 sensor maybe?

[Jameel]

Any idea how I can test that? I did not touch my lambda unit during the rebuild. Promise.

[gazzerr]

Mine does something similar to this because my WUR cold control pressure is too low when the WUR is stone cold. If I start it, let it run for about 45s then start it again, it behaves as it should - revs to 1300 and slowly returns to 950 as the WUR warms up. I think I need to knock the WUR pin in a little bit to adjust it. Have you ran fuel pressure gauges on it? Mine is only 0.8 bar. It should be 1.2 to 1.4 I think.

You can test the O2 sensor by disconnecting it - careful the connector can be very brittle and hard to disconnect. Then run a multimeter on it and you should be getting some like 1 to 1.5 volts when it's hot enough and has woken up. You'll have to look up the value to confirm.

When the engine is warm and idling in closed loop mode it should ever so slightly "dither" the revs back and forth about 25 or 50 rpm as the computer adjusts to stoich. If it's doing that then you know it's working. You can also look at the signal coming from the green wire on the test port with a duty cycle meter or scope and watch it waver back and forth adjusting the frequency valve.

[groovydude]

I wonder if your AAR is getting stuck closed when you drive it? In other words, it's not opening back up when the car cools down. Perhaps the bit of current to the heating strip on your first start attempt shocks it back into place? Just a thought. I have a spare AAR laying around if you want to test.

[Jameel]

WUR is adjustable. Tested great just a few weeks ago. Used the CIS pressure gauges. Sensor is new. Connector is new. Most everything in spec and tested recently.

I definitely have the "dither." Also used the parts classic CDI box tester and the green wire signal tested good.

AAR bench tested fine just a few weeks ago. Opens and closes fine. It's not like it moves fast anyway, so it doesn't seem like it could be "shocked" into place, if I understand its function correctly--in other words that bimetal strip isn't like a solenoid. It's slow to move. Thanks for the offer.

Here's a video I took just a bit ago. I start the engine, it does the thing, then revs up to 1100 on its own (I didn't touch the throttle at all). I then shut the engine off, and immediately restart. It pops right up to 1100 immediately.

[Walt Fricke]

You might try disconnecting the CSV to see if this hesitation has anything to do with it.

Also, when it starts in its momentary slow mode, what if you give the throttle a little blip? Because it sounds and looks like it corrects itself pretty quickly.

Interesting differential diagnosis question, though many would be overjoyed if this was their only intractable CIS issue.

[Jameel]

I will try disconnecting Walt. That shouldn't be too tricky. I am overall thrilled with the refurbed CIS. But now that I've got it running good, it seems small things stand out more!

[gazzerr]

Hmmm, sticky sensor plate or base setting maybe? Does it move freely?

[Walt Fricke]

Interesting idea - I can't imagine the pivot for the sensor arm getting sticky, but if somehow the plate got off center and rubbed a hair at its rest position?? That would mean starting on the CSV's contribution only, though I don't quite think that is enough to make the engine run as long as it did at super low RPMs. Easy to check with the boot and filter off - put a light into the filter area under the plate, and see if you see light all around the round plate's circumference. You can try just pushing up on the plate, but you want to do that only after the residual fuel pressure had fallen pretty low, or you will squirt fuel into a not running motor.

[wreckah]

the lambda sensor has no function when the engine is cold, so you can rule that out...(that is, if your lambda brain functions correctly).

measure cold fuel pressures and go from there. The SC CIS cars need a certain cold fuel pressure + extra air through the AAR, these combined give you the cold running characteristics.

[rwest]

Hi Jameel,

Double check the timing just in case?

Rutager

[Jameel]

Originally Posted by gazzerr

Hmmm, sticky sensor plate or base setting maybe? Does it move freely?

Well, it did when I rebuilt it a few weeks ago. And during startup and initial tuning.

Originally Posted by rwest

Hi Jameel,

Double check the timing just in case?

Rutager

I just timed it a few days ago. It's running great after that startup, just as CIS should. It seems like its just a cold start issue. Never does it after the engine has run for even a short while.

[Niels911SC]

I'll be following your thread, I have exactly the same issue however with me it always happens, not just 1 in 5 cold starts. I have had the problem for a couple of years, have done all the checks on system pressure, rebuild wur etc etc. Have not found a solution. So please report back if you manage to fix it!!
(link to my thread: click)

[schoward]

Mine does something similar maybe once every 5-10 cold starts. It has done this regardless of mixture settings I have tried over the years. When it does this, it generally recovers after a few seconds and idles up. My guess is my cold pressures is just on the cusp of spec but car generally runs great and usually starts fine so I have left it . I also run close to spec mixture and run with the vacuum pod connected with timing at spec and not sure the SC really likes the heavy timing retard at cold starts. My guess is cold pressure related. Following...

I know of others with 81-83 SCs with same or similar experience

[gazzerr]

Ultimately in my admittedly limited experience - I think the only way to really fix is to disassemble the entire CIS and first make sure that every part is correct for your model year of 911 and then test every single part for cleanliness and either electrical/mechanical operation within spec or ability to hold a vacuum within spec. Then reassemble and retest it for vacuum again as you go. I replaced every seal and hose. I installed a new o2 sensor (if required), new plugs, new leads and new injectors. Then make sure your timing and mixture is within spec as per the workshop manual. Other than that I think it's just a game of replace this/replace that and see if it fixes it.

[Jameel]

I hooked up my CIS gauges yesterday to try and fix this problem. Here's what happened.

Control pressure: aok.

Cold control pressure: at the bottom of the spec range (Bentley)

Warm control pressure: right where it needs to be 3.4 bar

I couldn't remember if knocking the plug leans or richens, so I raised it (my WUR is adjustable) until the cold pressure was at the top of the range (I did this before connecting the WUR wire for the warm test). I've since remembered that raising the plug raises pressure and leans out the mixture.

So this morning (temps in the low 60's) I went out to fire it up on cold start. Started right up, but was even worse than before, really low rpm, coughing almost and then died. Turned it off and restarted. Did the same thing again. So yes, raising the plug made it worse, leaned it out.

So I popped the deck lid and backed off my WUR adjustment nut, then knocked the plug back in a little. I didn't bother hooking up the gauges, so I don't know what my cold pressure is right now.

Started the car and it fired right up and idled immediately at about 1100. Again, I don't know what my cold pressure is right now, but I'm guessing its lower than it was before.

So here's the questions. On cold start the WUR is in a rich state (low pressure--mine was 1.7 before I knocked the plug) then as it warms via the electrical connection it leans out as pressure raises (mine is 3.7--in spec). Got that. Here's where I'm confused. Before when I would start the car and it would run at low rpm I would shut it off right away, then restart and it would pop right up to 1100 then settle down to 900 after running for a minute or so. Since the WUR leans as it warms, this theoretically should make the low rpm problem worse, right? But now I'm thinking that starting it once, shutting it off, then immediately restarting I'm giving the engine a double squirt from the cold start injector, so in effect enriching the mixture on the second start. Sort of like pumping the throttle on a carbureted engine to get extra fuel in.

Also, I've got conflicting info from searches on knocking the small plug. I'm fairly certain it only affects cold pressure. Once the metal strip warms its just hanging free, right? So it doesn't affect warm pressure. When I adjusted the plug yesterday and raised my cold pressure the warm pressure didn't change, so I'm pretty certain I'm on the right track here.

If it affected warm pressure too, then the AFR mixture screw would also need to be readjusted.

[groovydude]

I think the cold start "squirt" is consumed almost immediately, so I doubt that the restart is picking up residual fuel. I'm with Gazzer on CIS issues, unless the problem is obvious (ie you have incorrect cold pressure meaning your WUR is not up to spec), your best bet is taking the whole thing apart and rebuilding. I'll add that the FD could be your problem here, which is probably best left to a professional to rebuild. The only one I know of is Larry at CIS Flowtech, but I read recently that he's way backed up.

I believe you're correct in that once the WUR is warmed up it's taken out of the loop for the system pressure. Also, I recall reading that the adjustable WUR was only stock on earlier models. Perhaps a PO swapped your original WUR for an earlier model? I also recall that it's possible to modify a newer model, but I would be skeptical of that mod. Either way I'd bet that's your problem. I"m pretty sure your WUR should have 090 as the last three digits, but I would do a search here to make sure and check.

Of course, there's always the option of ditching the CIS and replacing with carbs (ugh) or efi (sweet). Personally, even though my CIS is now working perfectly, after MANY hours of labor, I feel a Bitz kit in my future.

[Walt Fricke]

Groovy - no Porsche came with an adjustable WUR. Bosch had some guy or gal at its factory carefully tapping the two steel plugs in until all was within their spec, and then they sealed it.

You have to do some drilling and tapping to make any of this adjustable.

However, for the cold start enrichment piece, if you want it richer (lower CP cold), you can disassemble the unit and add a shim where the bimetal spring is bolted to the end of the small rod (the one the poster knocked in). Don't ask me what shim thickness does what, but pretty thin.

Opening up the unit is also the only way to raise this steel rod post up higher if you have not modified things.

The simplest modification externally is to drill and tap the post, and drill a small piece of flat metal a bit larger. You place the metal part over the rod (the rod is conveniently recessed below the aluminum surface), run a nut onto a bolt - I think I used M5 - which fits your tapped hole, and screw the bolt into the rod. Then tighten the nut down. To raise the rod, hold the bolt head so it doesn't move, and tighten the nut. There are instructions and a drawing out there for a different way of doing this, but it calls for drilling a small hole partly in the aluminum and partly in the steel and inserting a thin rod to keep the steel rod from rotating. Not easy to do.

With hot CP within spec, when the bimetal spring which lowers the cold CP curls up as the heating coil or the engine heat or both do their job, the end of this spring rises above the top of the sping assembly which sets the hot pressure, and thus has no effect on hot CP. If you have things all out of whack, it might stay in contact and reduce hot CP.

You can find out more than you might want to know by some Pelican and Google searching, though often you won't find an answer to your own particular problem.

[Walt Fricke]

Jameel

Is your micro switch on the throttle working as it should? With the foot off the throttle that switch should be (closed? - doesn't matter, you can easily test with a VOM which way it works). The point is that the moment you touch the gas pedal and move the throttle up just a hair, that switch should click, and change state. You check this with the engine off in the engine bay by moving the throttle by hand. Just possibly it isn't working as it should and after you have driven the car and let it sit, it doesn't end up in the state it should. Since you don't touch the throttle to start the car, it would stay in that state. Turning off and back on might just correct that. Maybe another test would be to have the car where you have your issue, and before starting blip the throttle (won't hurt a thing)before each start to see if this relieves your symptom? Go to Jim William's website to see his CIS Primer, and read up on the micro switch and what it does.

Never leaving anything alone, Porsche/Bosch also have another switch (back behind the throttle body, much harder to see, feel, access than the micro switch with its two connecting wires staring you in the face) with three connectors and three states, which affect how the FV responds. With the throttle closed it is in state 1, and stays there until the throttle is open some, at which point it goes to state 2, and when open farther state 3, which goes all the way up to WOT. If I recall, states 1 and 3 may be the same - the Lambda system goes into a default setting on the rich side, and the Lambda sensor is cut out of the loop. This allows idle and very low RPM to be rich, and same with WOT, but for mid range, where street cars run about 99% of the time, the Lambda sensor optimizes the air fuel mixture to approximate stoich and give the best fuel mileage and the least pollutants.

Only bright idea I can come up with for this glitch in an otherwise all checked out system, and of course just a guess. Are all the contacts clean on the Lambda system big (14 pin, like the one connecting the engine over to the rear fuse/relay panel?) multi-pin connector? Maybe use some electrical contact cleaner and work in in and out some? You can do the same on the huge connector on the electronics box under the passenger seat, though corrosion there is pretty unlikely. Same with the roundish white plastic plug to the acceleration enrichment relay unit nearby, though I don't think that enters into the idle equation, but when in there, why not?