Help - odd 02/frequency valve behaviour
 

Need a little troubleshooting advice with my '81 SC oxygen sensor system. I've searched and read many threads but haven't found a solution to my situation. Sorry for the long info dump but I'm out of ideas...

Symptoms:
- poor gas mileage (last tank was 14 mpg/city)
- slight hunting at idle (warm and cold)

Tests conducted and notes:
- 02 signal not fluctuating - fixed at about 0.8V. Unplugged, the sensor still produces the same output so I assume (at this point) that it's functional.
- Pulled 02 relay - no change to idle or engine running conditions. Relay snaps when plugged/unplugged. Verified proper voltages/grounds are present using Bentley manual diagrams
- Felt frequency valve (engine running) and can't be sure if it's buzzing because the whole engine vibrates right? Inconclusive.

Wiring checks (with ohm meter)
- O2 cable to pin 2 on the lambda computer ok
- Frequency valve wiring between computer (pin 15) and valve connector, and between valve and relay ok
- Frequency valve resistance (2.9 ohms) verified between pin 15 on lambda connector and 87B on the relay

Dwell testing (using an oscilloscope)
- Cold start dwell about 75% (estimate). As engine warms up, dwell settles to 50% with O2 sensor unplugged
- Plug in O2 sensor: dwell jumps to about 80% (again, sensor output is about 0.85V fixed)
- Ground O2 signal wire: dwell gradually (about 10 seconds) decreases and engine starts to die. Assume this means the frequency valve is doing its job?
- Unplug frequency valve - little/no change to idle! This seems strange.
- Tried opening the oil tank cap (to force a lean condition, with frequency valve plugged back in) No change to O2 signal output.

The car starts quickly, pulls pretty smoothly so I don't think anything is really wrong with the fuel pressure (plan to check control and system pressures anyway) but I want the 02 system to function properly.

I have a new 02 sensor but haven't installed it yet, that's probably next but since the current sensor (which has about 10k miles on it) I've heard that 02 sensors can get "lazy" and not respond quickly but mine just seems to put out a constant voltage no matter what.

Any suggestions on what to try next?

Mike

just because the relay clicks does not mean it works right. try a new one. rarely, there's a cracked solder joint in the computer. pull out the guts and use a strong magnifying glass and look for circular cracks. a TV repair store can fix them if you don't have the required mini soldering iron.

John,

Thanks, I'll get a new relay and look at the computer (I'm an engineer and have spent countless hours with soldering irons!)

Mike

Mike, you are right, they do get lazy, espes. if you mix is a Little on the rich side. They slowly get covered in soot and get slower and slower over time. Easily seen on a meter. You can clean these sometimes and get ok results, but a new on wire Bosch is 19 bucks at O'reillys. Your test pretty much backs up a bad 02 as everything else looks good. I would put a new one in, get car at operating temp, and buy that i mean as hot as you normally get it, put on your meter and watch your dwell, and I bet it is just a little to rich for the 02, if it is, just turn your screw counter clock wise just a tad, and by tad I mean 1/8 inch, and then back clockwise about a 1/16, and then let it settle for a few seconds (these prehistoric lambda systems are very slow to react to changes sometime) The closer you get to stoiche, the faster you will see the needle moving on the dwell. EDIT: JW is rite, the relay has a high fail rate because it get very warm (normal) I take them apart and give a good cleaning to see if that the prob, and then get a new one. Get the correct realy, as the uni five pins from the partts stores will last about a week.

Thanks Don.

Forgot to mention one other thing - early this spring I installed a catalytic conveter (car always had a bypass and I wanted to clean up the exhaust smell) I can't really correlate the fuel consumption issues with the cat install, I had this issue last seaaon too but thought I'd mention it. The bypass tube had an O2 sensor and there's a chance I damaged it (don't think so but who knows) when I removed it and installed it on the cat. Cat is a used (20K mile) original Porsche part I bought from another Pelican recently.

Mike, I would take the cat back off if it where my car, An operable 02 will get rid of the bad fumes. The cat just robs power and make the car run a little warmer.

Mike,

It sounds like the system is not going into closed loop. Have you tried testing the main temperature switch that triggers closed loop ? It is on the right timing chain cover and is normally closed, contrary to the Bentley manual. The ground connection should be open above 59 F.

In my experience, 14 mpg in town and a slight hunting at idle are normal for a 81 SC. When the car was new the EPA combined rating was 16 mpg. Highway cruise at 60 mph should be closer to 25 mpg.

not respond quickly but mine just seems to put out a constant voltage no matter what.

Mike,

You may be confused about how the system works. The O2 sensor makes voltage based on the amount of Oxygen in the exhaust compared to free air. It does not know anything else. In open loop, the sensor should put out a constant voltage. When the temp sensor tells the system go closed loop, the FV changes the mixture based on the O2 sensor voltage. This changes the composition of the exhaust and causes the voltage to change in a feedback loop. Engine management systems are not accurate enough to maintain stoich within the 0.1 AFR needed by the cat, so they all dither the mixture back and forth of stoich to get close to a 14.7 AFR average. The CIS lambda on your car was designed to be lean everywhere and needs a default, pulsing FV to get near a basic fuel curve. It is a little confusing at first, because when you adjust the mixture screw you are only effecting open loop, cold start and +35% throttle, because on a operating system, the O2 sensor trim brings it back to stoich within a wide range. A high open loop dwell means a richer mixture, however in closed loop, you are looking at the correction factor and a lower dwell means a richer mixture. Got it ? My experience is that these cars run best with a hot closed loop dwell around 20-30 and the vacuum retard disconnected and pluged with the idle speed reset.

good call Paul. an easy test for the switch is to hook you test light ground to the spade (unplug it) and probe 12v. Light should be on when temp is below switching point, and as engine warms up it will go out. Of course at this time of year, it should be off. (you can put it in the freezer to check) But, that does'nt mean the 02 sensor is going to work right off the bat, as it has to get to several hundred degrees to turn oxogen ions into voltage.

Paul - thanks for the detailed description of how the systems works, I think I understood it correctly before but probably missed a few things. I do know that the sensor is essentially a battery that outputs a DC voltage that depends on the concentration of oxygen. I always thought (and read) that if you monitor the voltage of the sensor (unpluggeed of course) and introduced false intack air (to simulate a lean condition) you should see a change in the output voltage. I did this test by opening the oil filler cap but maybe this doesn't really work on 911 engines.

With the sensor unplugged and the engine warm, the duty cycle (which is not really dwell) is dead on 50% which is what I think it's supposed to do.

It did occur to me that the system might be stuck in open loop mode but since I was able to see the dwell change when I grounded the o2 sensor input to the lambda computer, I assumed it was operating correctly (lambda computer basically ok and doing what it's suppose to do when it detects a low signal level on the input - reducing the dwell/duty cycle to compensate) It would be interesting to connect the sensor input to a low voltage source (1V or less) and see if the dwell goes the other way (as it should?)

Over my lunch hour, I pulled the o2 relay and took it apart (don't have a new one yet) I did check it carefully with an ohm meter and 12V supply and it appear to be functioning properly. Of course there's no load on the terminals so it could still have resistive contacts (clean them up a little with emery paper just for fun)

I'll check out the temperature switch on the right side of the engine. Assume this is the 15C switch and not the 35C switch? Where's the other (35C) switch located?

mike

02 signal not fluctuating

Mike,

The only time you will see the voltage "fluctuate" ( back and forth) is when the system is in closed loop. If you have already verified at the test port that the system pulses the FV at 65% duty cycle (more like 58 dwell) open loop (default), the relay and ECU are powered and working. You need to focus on why it does not go into closed loop above 59F, beginning with the 15C switch. The 35C switch has nothing to do with this, it is for the cold acceleration enrichment system and it is located on the breather cover.

paul - thanks I'll check the switch tonight. Still a little confused though, one of the reasons I tried grounding the input to the ECU (to simulate an o2 sensor producing a low voltage) was to see if the system tried to compensate and sure enough, it did at the test port, which I assumed implied that it was in closed loop mode. Maybe this was an invalid conclusion?

- Ground O2 signal wire: dwell gradually (about 10 seconds) decreases and engine starts to die. Assume this means the frequency valve is doing its job?

Mike,

My suggestion would be to get a manual and try to understand how the system works, then do the basic troubleshooting. The FV is a variable pressure regulator, it is a pulsed injector in a return line. When it is dead, pressure goes up, the more it pulses, the more fuel bled off and pressure goes down. In CIS lambda, there is an inverse relationship between lower chamber pressure and mixture. The higher the duty cycle, the lower the pressure and the richer the mixture. When you ground the sensor wire, you are doing the rich stop test, the ECU see zero voltage and the duty cycle should immediately go to 95%, the opposite of your result. The FV is the actuator, this is a test of the ECU.

Thanks, the relationship between grounding the o2 input and the duty cycle was just the opposite of what I expected but I'm probably missing something. Like any control system, if the feedback between the sensor and the actuator was out of phase, it would probably oscillate...

>My suggestion would be to get a manual and try to understand how the system works, then do the basic troubleshooting.

Will do, I have an old copy of the Probst book and will do some reading, thanks again.

Help - odd O2/frequency valve - resolved
 

Just wanted to follow up on this for the archives.

I think I mentioned it in one of my posts but, I had installed a cat converter early spring and I suspected the mixture was set up with the straight pipe (by the previous owner).

After some reading and little head scratching, I suspected that the system was running very rich with the cat installed and that possibly the lambda system just couldn't adjust the mixture to the stoichiometric area (50% duty cycle) due to the limited range of the of the control system.

I got the system working properly by adjusting the mixture 1/4 turn CCW (lean) while watching the frequency valve duty cycle. Once I turned the screw the duty cycle moved near 50% and started fluctuating.

Car runs great and the exhaust doesn't smell at all. It's rather strange to have the car idling in the garage and not smell anything!

Mike