Lambda System functioning ?

[psalt]

Does this sound like the kind of vacuum input I am looking for for a typical WUR application ? It is the one that is connected now.

Bill,

Your engine's original WUR had no vacuum enrichment, it was a vent hose, so you are throwing another variable into the mix.

The BOSCH lambda switch usually has the temp stamped on the side of the hex, I checked a few and the terminals came both ways. The switch is grounded when the engine block temp is below this number and opens above it. The idea behind the alternate switch is to delay the onset of closed loop so that then engine gets a bit more enrichment until it is warmer. There is also another ECU , throttle switch and 35C switch on the breather to provide enrichment during cold driving. Have you tested those ? All of these gizmos were to squeak through the emissions test and still have reasonable drivability when cold.

Use the factory WUR pressure/temperature chart in the manual and compare the gauge reading with your IR thermometer.

[scarceller]

You can in car test any narrow band O2 sensor, here's how:

Disconnect the sensor from the harness so it's not plugged in at all. If the sensor has a signal line and also heater lines then if possible leave the 12vdc heater lines connected but not the signal line. Then with motor fully warm and running at idle measure the voltage at the signal line it should read somewhere in the range of 0.0 to 1.0 volts. Where anything below 0.5 is lean and above 0.5 is rich.

You won't know what side the engine is running at (lean or rich) before hand but if the sensor reads 0vdc it's either bad or very lean, so what you will do next is richen the mixture intentionally. In the CIS cars this is easy, just slightly push up on the air plate in the air meter funnel, this adds fuel to the mixture and the sensor MUST saturate at above 0.7vdc

If the sensor is already reading rich then go the other direction and lean it by pulling down on the air plate slightly and the sensor must drop below .3vdc

In the CIS cars it's very very easy to alter mixture via this air plate method.

If you find that the sensor does not respond to the mixture changes then the sensor is bad. Once you do this method on a properly working sensor you will really get the hang of how it's done via this very easy method.

[simsalabim]

That makes good sense Sal, thank you for that info.

A couple of things before I dive back in ...

Paul, are you sure that my original WUR did not have a vacuum sensor ? the primer I read on the subject gave me the impression that they added one in the mid-late seventies ?

Also, I stated that the vacuum line coming off my manifold that is now connected to the wur was pulling 15 or more bar and then dropping to zero when rpms rose beyond 2,500 or so. That was in inhg, not bar (sorry !) It is more like .6 bar ....

cheers

[boyt911sc]

Bill,

Your '81 SC USA model's WUR is 0-438-140-090. This particular WUR is not vacuum assisted. If you know and understand how your CIS works particularly the fuel pressure system, using an after market digital WUR is acceptable. Provided you know how to calibrate this unit. Using an OEM WUR that is correctly calibrated would be much simple. If after installing your new digital WUR, and you are still having some sort of problem, contact me and I'll lend you one to use for your test.

Tony

[psalt]

Paul, are you sure that my original WUR did not have a vacuum sensor ? the primer I read on the subject gave me the impression that they added one in the mid-late seventies ?

Yes, full throttle enrichment is handled by the ECU and the hose to the WUR was essentially an overflow tube like on a marine fuel pump. There is very little accurate information on your CIS lambda on the web.

[simsalabim]

OK and Great. Thanks for the feedback Paul and Tony (as well as your kind offer).

Things are starting to make more sense to me now as little by little, both the lambda cis and this digital wurs operation is better understood.

Last night, I installed the new 02 sensor (old one was completely covered with a hard black surface) and tested things again.

This time, I was able to get a signal from the sensor lead while warm, and it was around .85 volts when engine was fully warm and at idle. At this time, the control pressure was steady at 2.9 bar. The FV still runs at 50% duty cycle in open loop, but when the sensor lead is reconnected with a warm motor, it is now around 88%.

Seems odd ? if I understand things right, the 02 sensor is indicating a rich mixture, the control pressure is producing a rich mixture, but the FV duty cycle is responding in a way that would indicate it thinks the mixture is too LEAN ?

[blpetry]

The way I understand it, the duty cycle on a pulsed ground FV would be the INVERSE of what you would expect. It would read "on" when the FV was not receiving a ground and "off" when it is receiving a ground. That would mean that your "real" duty cycle for the FV being on is 12%, which holds hands with what everything else is telling you. Easily verified by leaving the meter hooked up and removing the oil tank cap, which will introduce unmetered air into the system and lean out the mixture, so your duty cycle should go "up" (down on your meter).

Regards,
Brian Petry, Springboro OH

[simsalabim]

Thanks Blpetry,

Well, that would surely be great news ..... I would then seem that once I increased the control pressure to specified values, the FV would then pulse on at a higher frequency and richen the mixture nicely.

But, is it true ?

Thanks very much,
Bill

[psalt]

Bill,

When the system goes into closed loop, the duty cycle dithers back and forth around 10 degrees, say 45-55 or 35-45. If your system is not going into closed loop, it could be the O2 sensor signal (at the ECU with your broken wire), the main switch, or the throttle switch. What is your idle speed like after a 10 mile drive ? Is it rock steady or does it dither up and down slightly ?

Your control pressure is still too low.

The reason I suggested an analog dwell meter is that you can see the quality of the sensor input and the response in the rhythm of the needle swings. A good quality analog meter is a $10 tag sale item. A tight engine with good even compression and a new sensor has a very different pattern than an old motor with vacuum leaks and an old sensor. Look in the archives, other listers have had problems with digital meters that display an averaged reading.

[blpetry]

Decided to run a little experiment this morning to confirm what I was thinking last night.

Hooked my car (warmed up, running in closed loop) up to my inexpensive amazon oscilloscope I purchased as a result of advice here and it showed the following ~50% duty cycle (The "up" and "down" portions of the square wave are roughly equivalent).



Next I removed my oil cap to introduce a lean condition and the duty cycle went "down" (the "down" portion of the square wave is now much bigger than the "up" portion).



Based on these findings, I think its safe to say that the duty cycle you are seeing is probably correct but it is truly the inverse of the effective duty cycle. Your engine is running rich and the computer is responding by reducing that effective duty cycle of the FV as you suspect.

BTW - here is the oscilloscope I have. Its very inexpensive but seems to work very well for the only thing I use it for (adjusting the mixture).

JYE Tech 06203P 062 LCD oscilloscope: Oscilloscope Probes: Amazon.com: Industrial & Scientific

DISCLAIMER: I'm not an electrical engineer, nor have I played one on TV. So, my observations are certainly subject to review and confirmation or rejection by people who actually know what they're talking about .

[mysocal911]

Originally Posted by blpetry

Decided to run a little experiment this morning to confirm what I was thinking last night.

Hooked my car (warmed up, running in closed loop) up to my inexpensive amazon oscilloscope I purchased as a result of advice here and it showed the following ~50% duty cycle (The "up" and "down" portions of the square wave are roughly equivalent).



Next I removed my oil cap to introduce a lean condition and the duty cycle went "down" (the "down" portion of the square wave is now much bigger than the "up" portion).



Based on these findings, I think its safe to say that the duty cycle you are seeing is probably correct but it is truly the inverse of the effective duty cycle. Your engine is running rich and the computer is responding by reducing that effective duty cycle of the FV as you suspect.

BTW - here is the oscilloscope I have. Its very inexpensive but seems to work very well for the only thing I use it for (adjusting the mixture).

JYE Tech 06203P 062 LCD oscilloscope: Oscilloscope Probes: Amazon.com: Industrial & Scientific

DISCLAIMER: I'm not an electrical engineer, nor have I played one on TV. So, my observations are certainly subject to review and confirmation or rejection by people who actually know what they're talking about .

This post should be used as a reference for future problematic issues with a CIS
setup as it provides a good graphical depiction of what occurs electronically:

1. With the oil cap removed, the duty cycle is actually increased, i.e. the ECU
is grounding the FV for a longer period of time to richen the mixture. This
is the same result when one grounds the O2 sensor, i.e. the duty cycle
goes to it's max value attempting to richen the mixture.

2. When the ECU grounds the FV for shorter and shorter times, the duty cycle
is decreased. In the limit when the ECU no longer grounds the FV, the duty
cycle is zero which is the full lean condition as controlled by the ECU.

The term duty cycle typically defines the ratio of an 'on' time (actuation of the FV)
to the total period of the waveform.

[simsalabim]

Great feedback one and all. Thank You.

I am going to attempt to re-map my digital wurs control pressure and warm up rate first, and then do some re-testing.

In the meantime, all of the talk about the closed loop readings and the factors affecting it, has got me re-assessing my 02 sensor connector, the lead running back to the harness, and any shielding issues I might have.

I have attached a couple of photos of the "current" situation (apologies). One photo shows the state of affairs with the connector and harness lead before I attempted to address it. The harness lead (damaged area) was being chaffed over a long period of time due to it's contact with another harness. It looks to me that the outer cover and the mesh shielding under it has been compromised, but the wires under them are still insulated and intact. I was thinking of trimming away the rough stuff and wrapping the area with a shield of some sort ?
The connector itself had only a few strands of copper wire left attached (via solder) to the connector pin, which was exposed and connected to the sensor black wire socket (I pulled it off for the photo).

The other picture shows what I had hoped would be at least a connector repair ?

I broke away more of the (by now pretty brittle) connector housing/mounting cover to expose more of the green wire and its copper wire. Since it was still difficult to work in that space, I butt spliced (crimped) a small piece of similar gauge speaker wire to the green wire and shrink wrapped it. I then soldered the other end to the original connector pin, shrink wrapped and electrical taped it. This is where the sensor lead attaches now (until I am sure it works).

I have read discussions of the importance of proper shielding and also some about "breathing" sensor leads, so my confidence in the connector fix as well as the connectors harness lead damage, has me a bit doubtful about whether the sensors signal is actually making it back to the ECU accurately.

[mysocal911]

Originally Posted by simsalabim

Great feedback one and all. Thank You.

I am going to attempt to re-map my digital wurs control pressure and warm up rate first, and then do some re-testing.

In the meantime, all of the talk about the closed loop readings and the factors affecting it, has got me re-assessing my 02 sensor connector, the lead running back to the harness, and any shielding issues I might have.

I have attached a couple of photos of the "current" situation (apologies). One photo shows the state of affairs with the connector and harness lead before I attempted to address it. The harness lead (damaged area) was being chaffed over a long period of time due to it's contact with another harness. It looks to me that the outer cover and the mesh shielding under it has been compromised, but the wires under them are still insulated and intact. I was thinking of trimming away the rough stuff and wrapping the area with a shield of some sort ?
The connector itself had only a few strands of copper wire left attached (via solder) to the connector pin, which was exposed and connected to the sensor black wire socket (I pulled it off for the photo).

The other picture shows what I had hoped would be at least a connector repair ?

I broke away more of the (by now pretty brittle) connector housing/mounting cover to expose more of the green wire and its copper wire. Since it was still difficult to work in that space, I butt spliced (crimped) a small piece of similar gauge speaker wire to the green wire and shrink wrapped it. I then soldered the other end to the original connector pin, shrink wrapped and electrical taped it. This is where the sensor lead attaches now (until I am sure it works).

I have read discussions of the importance of proper shielding and also some about "breathing" sensor leads, so my confidence in the connector fix as well as the connectors harness lead damage, has me a bit doubtful about whether the sensors signal is actually making it back to the ECU accurately.

A common error typically occurs when attempting to repair an old O2 sensor wire.
In repairing the O2 sensor wire, the shield wire gets connected to the center signal
wire which cases the Lambda ECU to operate in its full rich mode. The shield must
only connect to the ground pin at the ECU and remain unconnected at the O2
connector end.

[simsalabim]

Dave,

How specifically, does this relate to the repair shown ?

Thank You.

Bill

[psalt]

Bill,

There is a lister on this site that makes a kit to repair the O2 sensor connector. Take a look in the archives. There is no breathing going on.

[mysocal911]

Originally Posted by simsalabim

Dave,

How specifically, does this relate to the repair shown ?

Thank You.

Bill

Just do a simple resistance test between pins 2 & 5 of the Lambda ECU with it disconnected.
You should measure an open. Then measure between pin 2 and the O2 connection
in the engine compartment. There you should measure continuity.

[simsalabim]

Thanks again for all of the great feedback.

Paul, to answer your question, after a 10 minute drive the idle dithers. Maybe 25 rpm or so, up and down at a roughly 1 second frequency....

In your opinion, am I wrong in assuming that the single wire narrow band 02 sensors' black wire (from the sensor to the connector pin) is unshielded ? And, if there is a shielding problem, it is between the pin connection and its path to the ECU ?

I am trying to understand why with a low warm CP (around 2.95 bar) and a .85 V 02 sensor signal (both indicating a rich mixture), the measured duty cycle of the FV is over 80% which as I understand it indicates the ECU trying to react to a LEAN mixture ? One thought I had was that the high 02 voltage was somehow being attenuated on its path from the connector to the ECU, and telling it to further richen the mixture ?

If I checked the connection at the ECU at the appropriate pin (number ?), should I be getting that same .85 V ? If not, and it is lower than .5 V, this would explain the inconsistency right ? I could then focus on that critical path.

Thanks as always,
Bill

[psalt]

Bill,

If the idle speed rhythmically oscillates after the engine is fully warm, it sounds like it is going into closed loop. The oscillation is caused by the dithering duty cycle around stoich. If you are still not reading a dithering duty cycle, you need to try a different meter, then test the green/white wire circuit in the test connector. I would not adjust the mixture until you can read the duty cycle correctly. The duty cycle can be calculated by multiplying the dwell reading on the 4 cylinder scale by 90%. An analog dwell meter displays the numbers normally, not inverted, just as described by Probst and in the Bentley manual, and confirmed by many listers on this site.

[blpetry]

If someone could post a model of analog dwell meter that they have gotten to work on a 911 and where to buy it, I think that would be most beneficial (I did a search on our host's site and didn't see one).

I know that all dwell meters are not created equal, and in my case the "Actron" brand meter I have doesn't work properly - reads waaaay too low (and this is the one you will probably find if you walk into any autozone or probably many other FLAPS). I've also read (somewhere here) that you need to use a "powered" meter to get accurate results, but my 60's Sears model that is battery powered yielded similar results. Interestingly, both work perfectly fine on my CIS-equipped '84 VW GTI.

I guess it *could* be the test connector on my car, but the oscilloscope nails it through cold start enrichment, open and closed loop operation on the same wire.

[psalt]

here's an example of one that I know works well (if undamaged)

Sears Inductive Engine Analyzer 28 21423 Testes Volts RPM Dwell Amps Ohms Points | eBay

Look for one with that measures battery voltage with red and black leads and has a 4 cylinder scale