SaabCentral Forums banner
1 - 15 of 15 Posts

·
Registered
Joined
·
1,190 Posts
Discussion Starter · #1 ·
Lately, the CIS on my '86 sedan has been grumpy to say the least. There seem to be wild swings between running very well and running terrible. When it runs well, you can't even tell the engine is running while sitting at a stoplight and it accelerates like butter. When it's in a bad mood, which is increasingly most of the time, it idles very roughly and stutters under load. It's a problem that's been developing for several months and I'm finally at the point where I need to do something about it.

I've pretty much ruled out the AAV, warmup regulator and fuel distributor. I have not yet tested the injectors. The fuel pressure seems fine, but there is the possibility that it is intermittently not fine. In other words, it's fine when I check it, but is still suspect. I've experimented with other fuel pumps to try and rule out an intermittent problem, but they all seem to behave the same.

The main thing that I've been able to find that is not working properly is that the dwell is reading at the test port is 40%. If I understand correctly in trouble shooting, the dwell should be at 50% +/-5% for my year, and it should go up to 60% if I disconnect the o2 sensor, but it doens't. With o2 distconnected, it stays right at the same 40%. Every now and again, with or without the o2 sensor, the dwell reading will migrate up to 55-60%, then quickly drop back down to 40%.

Am I correct in assuming that the two possiblilities are that either the modulating valve is malfunctioning or that the ECU is not properly controlling it? It seems to me that if other components are faulty, the dwell should go to 60% as a default, not down to 40%. Any thoughts on how to proceed?
 

·
Registered
Joined
·
13,220 Posts
Check the output voltage from the O2 Sensor. Compare it to the Dwell reading.
 

·
Registered
Joined
·
1,190 Posts
Discussion Starter · #3 ·
Check the output voltage from the O2 Sensor. Compare it to the Dwell reading.
Output voltage is approximately .2 volts, though the behavior of the engine during this test was abnormal compared to it's typical recent abnormal operation (if that makes sense). Typically, the idle stays in the 900rpm range sometimes less, whether rough or smooth, sometimes as much as 1000rpms, but not typically much more. After warm-up this evening, the idle stayed upwards of 1400 rpms, which is highly unusual. What is also odd is that the dwell reading was oscillating between 48-54%, which is what I would suppose it should be at if it were idling properly, correct? There are no apparent vacuum leaks and the car was idling normally in terms of RPMs when parked earlier today.

I'm starting to think that I'm going to have to check the fuel pressure again tomorrow.
 

·
Registered
Joined
·
13,220 Posts
Poor running is very hard to handle with only a computer screen at a distance but....
What's the condition of the Sensor?
Do the Pulse Relations change if you hold it at a steady 2500rpm?
Does the 0.2v correspond to the poor performance?
If the Pulses oscillate as they should at 2500 rpm, adjust them to 30-40% (it improves cold running).

It sounds like you have an intermittent lean condition and the Lambda system is trying to compensate until it goes out of range.
Intermittent vacuum leaks could cause that. Check the Brake Booster for a vacuum leak.
Check the Injector spray pattern, clean as needed.
 

·
Registered
Joined
·
1,190 Posts
Discussion Starter · #5 ·
Poor running is very hard to handle with only a computer screen at a distance but....
What's the condition of the Sensor?
Do the Pulse Relations change if you hold it at a steady 2500rpm?
Does the 0.2v correspond to the poor performance?
If the Pulses oscillate as they should at 2500 rpm, adjust them to 30-40% (it improves cold running).

It sounds like you have an intermittent lean condition and the Lambda system is trying to compensate until it goes out of range.
Intermittent vacuum leaks could cause that. Check the Brake Booster for a vacuum leak.
Check the Injector spray pattern, clean as needed.
There is nothing particularly unusual about the physical condition of the o2 sensor, so I would call it typical used condition in terms of soot accumulated. I did have an oil burning issue due to the crankcase being pressurized for about 10 miles worth of driving, but the irregular running predates that by months slowly getting worse with time. The poor running problem did exist over the course of 2 different o2 sensors as well, the first being in the N/A manifold and the second being in the recently installed turbo manifold. The o2 sensor in the turbo manifold was in visibly better condition than the original N/A manifold sensor, so I went with it. Neither one seems to have performed better than the other. Just to be clear, the problem existed prior to and after my change from N/A to turbo, though it is more obvious when boosting during the spells of poor running (4psi is the maximum I have boosted to regularly).

I'm not sure what the .2 volt output correlates to since the running condition at that point differed from any previous running. I'm hoping that proves to be an anomoly. I'll do some more tests today at various points and try to get a measurement at more typical poor running and good running swings in its behavior.

The pulse relations seem to remain constant regardless of RPMs, but I'll make more careful note of this at 2500 RPMs specifically. It's almost as if it has just decided it will remain at 40% regardless of any other input. When it was reading 40%, there was absolutely no variance. The reading was exactly 40% and did not change by more than 1% in either direction. I would not call what little change occuring an oscillation either, it was very gradual and could take 30 seconds to move 1%. The needle was rock solid on its reading.

The brake booster is an interesting thought. I don't notice any performance variation to the brakes, but I'll have a look.

I plan to replace the injectors anyways, would they cause this sort of intemittant problem?
 

·
Registered
Joined
·
13,220 Posts
The pulse relations seem to remain constant regardless of RPMs, but I'll make more careful note of this at 2500 RPMs specifically. It's almost as if it has just decided it will remain at 40% regardless of any other input. When it was reading 40%, there was absolutely no variance. The reading was exactly 40% and did not change by more than 1% in either direction.
Whenever the Lambda system is disabled, out of its range of adjustment, or getting no signal from the O2 Sensor, it will default into "Lambda Lock". Lambda Lock is at 60% Pulse Relation. I'm assuming your dwell meter is reading in reverse. Swap the leads and see if you get the 60% Lambda Lock you should get.
Compare the dwell reading with the O2 reading (they should move in concert) to confirm that the Lambda system is working properly. Typically, on cold start, it would lock at 60%, show a brief effort to adjust when the Sensor gets hot, then locks again when it determines that it's out-of-range.
Dirty Injectors can do all kinds of things, but it seems more likely that you have a vacuum leak causing a lean condition (0.2v O2 Sensor output). The Continuous Injection system, being continuous, rarely has intermittent variations such as you describe. The only thing I can think of that could be an intermittent vacuum leak is the Booster because it moves. If you have another insight into possible sources of a leak, check them with Carb Cleaner, Propane (an unlit torch), what-have-you.
 

·
Registered
Joined
·
7,033 Posts
Jim, if you have a vac leak on the booster does it not often kill or bog the engine on braking?
 

·
Registered
Joined
·
13,220 Posts
Jim, if you have a vac leak on the booster does it not often kill or bog the engine on braking?
It might, but it's not a diagnostic indicator.
If I had that symptom, I'd check the Booster.
I wouldn't exonerate the Booster because it didn't do that.

I only mention the Booster because it gets moved regularly and can leak enough to cause major symptoms. Other vacuum leaks are a possibility, especially with a Turbo conversion.
 

·
Registered
Joined
·
1,190 Posts
Discussion Starter · #9 · (Edited)
Whenever the Lambda system is disabled, out of its range of adjustment, or getting no signal from the O2 Sensor, it will default into "Lambda Lock". Lambda Lock is at 60% Pulse Relation. I'm assuming your dwell meter is reading in reverse. Swap the leads and see if you get the 60% Lambda Lock you should get.
Compare the dwell reading with the O2 reading (they should move in concert) to confirm that the Lambda system is working properly. Typically, on cold start, it would lock at 60%, show a brief effort to adjust when the Sensor gets hot, then locks again when it determines that it's out-of-range.
Dirty Injectors can do all kinds of things, but it seems more likely that you have a vacuum leak causing a lean condition (0.2v O2 Sensor output). The Continuous Injection system, being continuous, rarely has intermittent variations such as you describe. The only thing I can think of that could be an intermittent vacuum leak is the Booster because it moves. If you have another insight into possible sources of a leak, check them with Carb Cleaner, Propane (an unlit torch), what-have-you.
I believe the leads to be properly attached. If you reverse them, the reading is around 15% which would make no sense. I double checked everything for vacuum leaks and I just can't find anything. The vacuum level is currently at 20mm hg when idling at just under 1000 RPMs, which I would typically expect to be higher if there is a leak (I suppose it's actually lower vacuum, I mean higher as in more pressure/closer to atmospheric pressure).

My test results today differed from last night despite the fact that it had not moved since then. The car started up cold and idled at normal <1000 RPM speed very quickly after starting (3 seconds at 1200 or so). The dwell indicated normal start up operation, but settled at 40% shortly afterwards and before I would call the engine "warm". It remained there even after a drive. I decided to experiment with the adjustment screw on the air meter/fuel distributor assembly and found that clockwise turning of the screw would increase the dwell reading until it reached the maximum of 60% where further turning yeilds no change. Turning it back counter clockwise I was able to get a sustained 50% with a +/- 4% oscillation. The RPMs did increase slightly vs. the original reading before this experimenting and have settle at approx. 1000 RPMs on the nose with 20mm hg vaccuum still. Changes to the throttle resulting in 2500 RPMs yeild dwell readings of
around 44%. Perhaps the o2 sensor starts to function better at higher RPMs yeilding an adjusted Pulse Relation? (I suspect from the exhaust smell that typically the car is running rich, not lean)

Measurements at the o2 sensor at idle have been varied and do not seem to correlate to any particular scale paralleling the Pulse Relations. The readings are always low, from between .1 volt and .3 volts, though the smell of the exhaust is somewhat rich (not a lot, just a little). If you unplug the o2 sensor, the dwell readings go to 40%. As I understand it, unplugging the sensor should trigger the "Lambda lock" at 60%, but instead it drops to 40%. The interesting thing is that adjustment to the sensor plate screw seems to generate the correct change to the Pulse Relations (unless my understanding is incorrect), enrichment (CCW) drops the P.R. and reduction in fuel (CW) increases the P.R., so I'm fairly certain that my dwell meter is not simply reading the opposite portion of the signal. Also, there seems to be a ceilling at both 40% and 60%. There have only a couple times where I've been able to coax a reading beyond those points and it's never been more than a flash at them rather than a sustained period (no more than 1 second), but I suppose that could be related to how infrequently the fueling system is needing that significant of an adjustment (sudden acceleration or sudden closing of throttle?).

Also, checked all vacuum lines and grommets for leaks with carb cleaner including the brake booster with no results. Everything seems to be sound in that regard. Checked the injector seals as well with no issues detected. Could an intermittently functional warm up regulator cause problems due to migrating control pressure? As I understand it, control pressure opposes the line pressure which makes subtle change to the plunger position during startup. Surely the lambda system should compensate for variation in control pressure by varying the Pulse Relations, correct?

It seems to me that things are starting to point mostly toward a faulty o2 sensor?
 

·
Registered
Joined
·
13,220 Posts
The purpose of running at 2500 rpm was to see if the O2 Sensor reacted differently at a higher temperature.
Apparently, yours did. That's a certain indication of a failing Sensor. That may not be all of your problems, but it should definitely be replaced.
Replace it and see how many of your problems disappear.
 

·
Registered
Joined
·
1,190 Posts
Discussion Starter · #11 ·
Installed the new o2 sensor and it does seem to behave differently, but it is unclear as of yet if it is better or worse. After initial install and warmup of the sensor, the Pulse Relations were out of range with the readings going from 60% to 40% and back again as if it couldn't make up it's mind. After some adjustment to the sensor plate screw, the readings settle in at 44%-50% with oscillation. Changes to the air sensor adjustment do not generate changes at the Pulse Relations unless they go out of range, where the 60-40 alternating readings occur again. Previously, I was able to move the Pulse Relations based on the adjustment of the air sensor plate screw until the 60% or 40% was reached where the readings seemed to max out. I would have though that this would be the correct function, indicating that the computer was changing the Pulse Relations to compensate for my changes to the sensor plate screw. It seems to me that if the P.R. does not change, but my setting of the sensor plate screw does, then the computer is not properly compensating for my A/F adjustment by matching the P.R. to the o2 sensor's detection of changes to the fuel mixture. P.R. did move slightly higher when rev'd to 2500 RPMs, but would settle back to the same range after a couple moments of continuous running.

The oscillation within the 44%-50% range is noticeable in the idle as well and is a pulse of about 1 second from high to low. I would say that the quality of idle is not any better than it was but have yet to make further tests of the sensor voltage or road test the car. Idle is still around 1000 RPMs with 20mm hg vacuum.
 

·
Registered
Joined
·
13,220 Posts
You've done all the obvious stuff now, it may be time to "test by substitution" the Lambda ECU.
The Pulse Relations should oscillate smoothly, in a 10% range. Your description of an either/or behavior suggests something electronic.
Another general pointer: try to get it to work properly and smoothly at higher rpm (1500-2000) so you don't get confused by fouling Plugs or balky idle issues. Worry about the idle after it's straightened out.

It doesn't have an exhaust leak near the Sensor, does it?
 

·
Registered
Joined
·
1,190 Posts
Discussion Starter · #13 ·
No exhaust leak. The car seems that it is performing better on the road, but it's only been 75 miles or so and I was hauling a 1000 lb. trailer so hard to tell if there was any stuttering under those conditions. There are times when it's been good for that long and does not run as well the following day. I'm optimistic that the fueling is generally better though, so I'll start looking at some other possibilities such as timing and rotor/cap (which were replaced around 30,000 miles ago). The plugs are new since the heat range seemed to change when the turbo system was added and inspection shows the new heat range is working better than the N/A set did with the turbo. There are no signs of poor combustion on the plugs or anything out of the ordinary such as oil or water.

In continued testing of the brake booster, I was able to get the engine to idle higher for a short duration by pumping the brakes. Would this be a sign of problems with the booster? With persistent and quick pumping while idling, the RPMs would go up by around 100-200 RPMs for around 2 seconds and then drop back to the previous idle speed.
 

·
Registered
Joined
·
1,190 Posts
Discussion Starter · #14 ·
I did discover a vacuum leak today, in a place I wouldn't normally look I suppose. The vacuum line running from the throttle body to the vacuum diaphragm on the distributor had a crack in it somewhat in the middle of the line, not at the nipples where I had checked. After replacing the line, the idle did drop a bit and running at highway speeds has definitely smoothed out some. I'd say that between the o2 sensor the this vacuum line, it's now running about as good as it will be getting. There is still a slight stutter at times while idling, but I'm sure I'll be able to iron it out with some further tuning. Otherwise it's not too bad at this point.

One thing I did notice is that the o2 sensor readings are now quite odd. At idle they sit at around .3 and then suddenly drop to 0 for a short time and then back up to .3. What would this behavior indicate? I'll be trying out a alternate ECU just to see if there are any differences, but I'm not a big believer in ECU failures so I'm not optimistic that there will be any change. I've pulled one that is the exact same part number, so it should give a good comparison.
 
1 - 15 of 15 Posts
Top