82 SC Fluctuating Lambda Duty Cycle
 

I've got another one for the brain trust. If anyone has seen my recent posts, I recently rebuild the CIS on my 82 SC and performed a major service. Now I'm trying to set the mixture via the test port on the lambda system. With an oscilloscope attached to the test port (red wire to ground, black to the green wire on the test port), I'm getting the following fluctuating duty cycle. Is this normal?

https://youtu.be/cDs5I0ifs8U

The duty cycle is pinging back and forth between 45% and 60% and minor adjustments to the mixture screw only shift this up or down without changing the fluctuating (at least I think). Engine is warmed up.

Things to note:
- I verified no air leaks by smoke testing the system after CIS rebuild (replaced all vacuum lines, injector seals/buckets, new airbox); when I pop off the oil fill cap the throttle changes significantly, indicating there is no significant air leak
- valve lash is set, head studs are tight
- ignition timing was set, 5 deg BTDC at idle, advances with more throttle
- WUR is rebuilt, system/control pressures at spec
- just replaced 02 sensor, old sensor was original (!!) and was giving the same duty cycle fluctuations

Do I just need to fiddle with the mixture adjustment more and it will even out? It seems like the system is fighting to try to get to lambda 1.

I have an 83 with functioning lambda. It fluctuates, after warmup, usually abut 8, meaning + or- 4%
around whatever I am shooting for. I have seen it fluctuate more like yours does but I just assume the O2 sensor has dropped below the temp it needs to work. Try raising your idle to maybe 1200 the reset and see if it holds a tighter range.

I have not used a scope for the lambda adjustment. I have used a dwell meter and adjusted for a minimum swing on the meter from left to right on a 4 cylinder scale.

Such an amplitude in the duty cycle oscillation is quite huge, but what makes me more wondering is the uneasy voltage shifting of the pwm signal where beside that its voltage is too low and should be present only on the positive voltage level.
You can clearly hear the uneasy and very oscillating idle.

Check the voltage at the OXS Relay under the passenger seat if at least 12v are present, made its too. As you have an oscilloscope also check the signal at the frequency valve port 15. Here a clear peak and hold pWM signal which drives the frequency valve should be present.

This is how it should look like at port 15 of the ECU unit/plug.
Maybe the signal results on your side inverted.

http://forums.pelicanparts.com/uploa...1717101989.jpg

Thanks, guys. I'll check for voltage at relay and look at the signal on port 15 of the ECU, and report back. Hopefully this gives us some idea what is happening.

I use a dwell meter to adjust my car (82 SC). Can't recall how big the swing was last time I did it (years ago) but based on engine sound alone, I'd say your car is completely normal sounding for one that's all warmed up.

Normal. Go drive it!

Interesting feedback. I might be chasing a ghost but I might as well go through the checks. I also bought the car with the O2 system unplugged, which makes me suspect.

A few updates:

1. With the O2 sensor unplugged, the duty cycle is steady at 50%. I thought it was supposed to default to 65% with no O2 sensor...

2. I verified the O2 sensor (which is brand new) is working. The voltage starts around 0.45 and then fluctuates within a few tenths once the frequency valve starts to do its thing.

3. I verified voltage at terminal 30 of the O2 relay, which was 14.4 V at all times. With the car running, terminal 88 was 13.3V. I thought it was strange they were not the same voltage when both are hot. Probably nothing, but let me know what you think.

Andrew: what is the best way to get to pin 15 of the ECU while running? I don't see an easy way to backprobe it or disassemble the connector. Any suggestions?

I have a 1981 SC, Duty cycle varies from 52 to 58 % just checked after a long drive, also the idle doesn't sound like yours, it keeps dead steady, no pulsing whatsoever. Yours seems to surge a little then back, I have no idea what could cause that.


Edit: I see a light surging as yourself when the idle rpm is set high, maybe the advance is bouncing it around too, it's not as rhythmic as yours, more random, but it's there. Only when I drop the rpms to ~ 850 then it's rock steady.. the duty varies somewhere between 52 and 60, so really not that different from yourself, maybe it is somewhat normal which is what other's have said.

Phil

Default is 50% duty cycle when running open loop.

That's one unsettled trace. What does it look like with the O2 unplugged ?
http://forums.pelicanparts.com/uploa...1717293004.jpg

Here's a pic from the last experiment I did playing with this.... leads attached to the same ports. I didn't notice any voltage droop but that could simply be the scope.

http://forums.pelicanparts.com/uploa...1717270709.jpg

Andrew, I opened the lambda computer and tapped into port 15. When I initially started the car it was running great and I got the waveform you provided as an example:

http://forums.pelicanparts.com/uploa...1717287017.jpg

That must have been before it switched to closed loop because after a few minutes the car started running rougher and the idle started oscillating like I showed above. The waveform also went nuts and the voltage dropped way down to tenths of volts. Here is a video:

https://youtu.be/t6j8JmkE7x4

Does this mean the lambda computer could be on faulty, or could something else cause this?

Yes, some can be intermittent, requiring a re-soldering of the circuit board, a 10 minute effort.

That wave form looks like the one in the other recent O2 thread where solder was used where it shouldn't have been, rendering the O2 sensor erratic and useless.

You mean this? https://youtu.be/t6j8JmkE7x4

Most likely the 2901 quad comparator in the Lambda ECU has gone open-loop (hi-freq oscillations), e.g. cold solder joints.

Completely normal sounding? Really ... to me it sounds totally uneasy and never meant to be like that per factory default, and this also is proven by the first video where you can see in the plot that something in the electronics is flawed.

As many of these old ECUs after such many years are now partly flawed, then , ok ... its obviously "normal" but not working as it should.

I didn't see anything about the age of the current oxygen sensor. The symptoms suggest to me that it might be getting lazy and is on its way out. Replace it and see if the duty cycle fluctuations tighten up to where they should be.

Nix

O2 sensor is brand new. Had the same symptoms with the old one too.

It sounds like the lambda computer may be the culprit. Any suggestions for repair? I see the specialized ECU folks fix these and the feedback on the forum is mostly positive. Alternatively, anyone have a known good one they want to sell?

Found this company; https://www.systemsc.com/products.htm