'83 SC runs great; fails smog with high HC emissions

[zeusdog]

My 1983 targa failed smog for the first time ever last week. Here in CA, we're required to be smog tested every 2 years. In all past years, the car has passed with very low emissions of HC, NO and CO - way below the allowed limits. This year, however, the measured hydrocarbons were very high (3-4 fold higher than the max allowed). Here are the 2017 and 2019 test results. HC, CO and NO concentrations are in parts per million (ppm).

MPH | RPM | %CO2 | %O2 | HC | CO_ | NO

2017
15 | 1543 | 13.3 | 2.0 | 24 | 0.01| 477
25 | 2381 | 13.7 | 1.6 | 21 | 0.01| 501

2019
15 | 2550 | 14.4 | 0.4 |561 | 1.06| 100
25 | 2464 | 14.5 | 0.3 |428 | 1.07| 100

For both the 2017 and 2019 smog test, the car was thoroughly warmed up. (The smog shop is 20 miles from home.) Test was done at the same shop, by the same technician. Unlike some other smog test shops I used prior to 2017, I trust this guy. It's a "STAR test" station, meaning they only do the testing and can't offer any repairs or even suggest a repair shop. They will retest for free once I solve the problem.

The 2019 results certainly look like it's running rich. The car runs fine - starts easily (cold or hot), steady idle, runs smooth and pulls strong, although the exhaust smells nasty. No backfiring on deceleration.

In the time between the 2017 smog test and this year, I replaced the fuel pump, fuel filter, and fuel accumulator due to a hot starting problem (fixed). Also changed the oil and replaced plug wires, distributor cap and rotor, and one fuel injector that had a poor spray pattern. I have made no adjustments to the fuel system, but did reset the ignition timing. Removing the oil filler cap produces a VERY slight RPM change (too small a change to show on the tach, but maybe 15 RPM difference by ear).

Here are the measured fuel system values


Year of engine: 1983 (US)
WUR model: 090
Temp during test: 14 C
WUR resistance: 25-26 ohms
System pressure: 4.6-4.65 bar
Cold control pressure: 1.1 bar @ 14 C
Warm control pressure: 3.4 -3.45 bar
Time delta cold to warm: 150-180 seconds
Residual pressure:

5 min: 1.65 bar
15 min: 1.6
30 min: 1.45
60 min: 1.4

Any and all thoughts or suggestions will be most appreciated.

[cpa]

Zeus dog, what is the total mileage and when was if ever was the oxygen sensor replaced?

[zeusdog]

78K total miles; O2 sensor replaced around 60K.

[zeusdog]

I forgot to mention above that the CO emissions (which are about 6x higher than before) also exceed CA standards.

[Solamar]

I noticed your NO dropped substantially also. From what I understand high NO is caused by lean conditions or high combustion chamber temps.

Just another sign pointing to you having a rich running condition.

[zeusdog]

I agree. O2 is also quite low. Except for the car starting and running well, all signs point to it being too rich. But the system, control and residual pressures seem to be within spec. If the cold start valve was leaking, my understanding is that the residual pressure would drop quickly...and it doesn't.

HC concentrations are up more than 20 times what they were 2 years ago, and CO is up 100-fold. Would idle mixture affect the exhaust gases this much at 2500 RPM? Would a dead O2 sensor do it? A dead catalyst?

[zeusdog]

Having found my system pressure, cold and warm control pressures, and residual pressures all to be within spec (as reported in my origial post) I have checked the O2 sensor output, watched the FV signal on a cheap digital oscilloscope, and only then made an adjustment of the mixture screw.

Like others, I found the O2 sensor connector to be in bad shape - the plastic is crumbly, and the rubber boot that goes over it (housing the female connector from the O2 sensor) was difficult to separate from the plastic part. I *think* the sensor as found was still electrically connected to the shielded green wire leading to the ECU - but I can't be sure of that. I checked continuity from the connector to the ECU and it was good. I kluged together a jumper that allowed me to keep the O2 sensor connected while also attaching a lead from a voltmeter. (The other lead was grounded to the engine.)

Fired up the engine and watched the O2 sensor voltage on the meter and FV signal on the scope. Initially, the O2 voltage dithered in the 0.58 - 0.7 V range. The FV signal dithered some, but appeared to always have a duty cycle below 50% (typically appeared to be about 30%). I adjusted the mixture screw about 15 degrees counterclockwise. The next day, I measured the O2 voltage again, and found it to now dither from about 0.3 to 0.75 V.

Upon further reading of this forum and elsewhere, I realized that it is important to get the O2 sensor good and warm to get a good reading. I took the car out for a 10-minute hard drive with a number of full-throttle accelerations (strictly for diagnostic purposes, I assure you). I listened very carefully for any pinging/knocking and there was none. The butt dyno felt that peak torque and acceleration were about the same as before, but was smoother and not as peaky. Afterwards, I found the O2 sensor voltages at idle to dither over a narrower range (0.4 -0.65 V or so), with the FV valve duty cycle varying accordingly, although still never going longer than 50%. (This refers to the low side of the square ware.) If I lifted the flowmeter lever slightly, the O2 voltage increased (as expected) and the duty cycle of the FV valve signal became much shorter (also as expected). If I pulled down on the lever (leaning the mixture), the opposite occurred, and only then did the duty cycle get somewhat longer than 50%. Removing the oil filler cap had the same effect.

So now I am considering taking my car back for a smog retest before doing anything further. (I get one free retest if I bring it back within a month.) My questions to you Pelicans are:

• Is it possible that a small adjustment of the mixture screw would be sufficient to bring tailpipe hydrocarbons down by a factor of 5 and reduce CO by a factor of 100?
• I am hallucinating to think that is all that is needed?
• Should I adjust a bit more to get the FV duty cycle to dither more equally above and below 50% without artificially leaning or enrichening the mixture with the flowmeter lever or introducing unmetered air?
• Given my drastic increase in HC and CO emissions from previous smog tests, is it likely that my catalyst is toast?

[pmax]

Great troubleshooting.

No harm taking it to the test again since adjusting the mixture screw definitely has a big impact on the CO.

In fact, the official manuals recommend using a CO sensor during mixture adjustment.

As to the change from 2017, the refresh could have fixed the air leaks and voila, it's rich.

[john walker's workshop]

A MSD Blaster coil fixed the HC issue on one recently. HC is unburned fuel basically, and the spark wasn't getting the job done. Low CO could be the same thing because fuel has to burn to produce CO.

[rubenmad]

Zeusdog, were you able to pass smog on the second time? If so, how much did your HC's drop?

[Bob Kontak]

Originally Posted by zeusdog

I have made no adjustments to the fuel system, but did reset the ignition timing. Removing the oil filler cap produces a VERY slight RPM change (too small a change to show on the tach, but maybe 15 RPM difference by ear).

Pull the oil cap when warm. If just after cold start, idle will not drop.

If you can get it to dither around 50% you may be good (emission check-wise) at idle but it sounds like you may have an air leak based on the oil cap statement.

Off idle, the leak becomes smaller relative to the air coming in. Fuel delivery is sort of linear in these things so at 15 and 25 mph you probably have excess fuel if you do have an air leak.

Your CO/HC numbers are large and unfortunately, I don't understand the order of magnitude of the numbers. Can tweaking the AF 3MM screw cause this much of a change?

Timing? What did you do when you reset it? Factory settings?

I like JW's coil recommendation. Blaster coils are relatively inexpensive. Not the worst part you could roll the dice on.

Edit: Man, just saw the dates of the posts............

[zeusdog]

Originally Posted by john walker's workshop

A MSD Blaster coil fixed the HC issue on one recently. HC is unburned fuel basically, and the spark wasn't getting the job done. Low CO could be the same thing because fuel has to burn to produce CO.

As it turns out, this suggestion is pretty close to the mark. See post below.

[zeusdog]

Originally Posted by Bob Kontak

Pull the oil cap when warm. If just after cold start, idle will not drop.

If you can get it to dither around 50% you may be good (emission check-wise) at idle but it sounds like you may have an air leak based on the oil cap statement.

Off idle, the leak becomes smaller relative to the air coming in. Fuel delivery is sort of linear in these things so at 15 and 25 mph you probably have excess fuel if you do have an air leak.

Your CO/HC numbers are large and unfortunately, I don't understand the order of magnitude of the numbers. Can tweaking the AF 3MM screw cause this much of a change?

Timing? What did you do when you reset it? Factory settings?

I like JW's coil recommendation. Blaster coils are relatively inexpensive. Not the worst part you could roll the dice on.

Edit: Man, just saw the dates of the posts............

The rest of the story:

After adjusting idle mixture to perfect dithering around 0.5 V on the O2 sensor and 50% duty cycle on the frequency valve, I took the targa back to the same smog station. No dice - HC emissions decreased slightly (a few percent), so I failed horribly once again (tagged as a dreaded "Gross Polluter").

At this point, I reasoned that the lamda system seemed to be working properly, and the mixture going INTO the cat can't be TOO far off with the O2 signal dithering around 0.5 V and reacting as expected to warm-up, WOT and so on. I was already into double-overtime (late registration fees and threatening letters from DMV), so I bit the bullet and bought a new CARB-approved catalytic converter from our host, as well as a new lamda sensor. After installation, the car ran the same. The O2 sensor voltage range was essentially same as before, and the FV signal continued to behave as expected (dithered around 50% at idle after warmup).

Checked the timing once again - no adjustment was needed; it was right at factory spec. I replaced the spark plugs with new, although didn't expect it to matter since the plugs were not that old. The old ones looked fine and fairly consistent from cylinder to cylinder. Took the car for a drive, expecting no change. Whoa! The butt dyno detected an extra 10-15 HP. I can only assume that, despite the plugs appearing OK, at least one of them had been misfiring. Or maybe I had failed to connect one of the plug wire caps properly when I installed the Clewett Engineering plug wires last year.

I took the car back for a third smog test, and VOILA!: passed with flying colors.

Results (@ 15 mph/@ 25 mph)

CO (ppm) 0.00/0.00
HC (ppm) 4/5
NO (ppm) 93/6
CO2 (%) 15.2/15.1
O2 (%) 0.4/0.4

That's a hundred-fold decrease in HC (from about 3x the max allowed down to about 1/30th of the max allowed). CO not even measurable...

So, apparently my catalytic converter had died, leading to the failed previous tests. Excessive exhaust hydrocarbon concentrations are a known cause of catalyst failure. My strong suspicion is that a misfire caused by inadequate spark under some driving conditions caused such a failure. Although the lamda signal and FV duty cycle indicated a reasonable mixture at idle and low throttle settings, I believe I must have had a routine misfire at higher revs. (In fact, I notice that hard acceleration now is smoother than it used to be, as well as stronger.)

I'm very glad to (I hope) be done with this fiasco. I doubt that I would have ever learned how the CIS works if not for this experience, so I guess that's a silver lining.

Thanks for help.

Zeusdog out.