This is by the far the most extensive troubleshooting I have ever done on a car. Everything I've done I learned right here and by using a manual, and for that education I thank the community of Honda Tech.

In regard to cars, there is nothing worse than failing a smog test -- except failing it several times.

1990 Civic Si converted to obd1, running a D16z6, connected to a P28. AEM air intake. DC header. New distributor. No check engine light. Magnaflow California approved CAT with less than 2,000 miles.

See images.

NGK ZFR5F-11 Plugs, 150 miles later. White spark found at 1-4.

Spark plug wires resistance:
1. 6.6ohms
2. 5.5
3. 4.5
4. 4.0

CKP sensor: Resistance at 368 ohms between B and F. No continuity found at ground.

Injectors click with 9v. / 12v found at yel/blk / injector coils, 12ohms / ground to ground, 12v found, clips connected.

Fuel Pressure:
*At filter, engine off, pump primed, 28psi.
*Engine running, 28psi.
*Engine running, vac line disconnected at FPR, 34psi.

Timing: 16 BTDC Idle: 750

Compression, warm engine:
Dry: 120 psi -- Low, but even. Readings specific to gauge?
Wet: 130 psi -- ALSO EVEN.

Valves: intake .007 / exhaust .009 (across)

O2 sensor: New Denso unit. Readings untested.

New ECT sensor, reading 3.6k ohms cold, 267k ohms hot.

IAT reads 3.8k ohms cold, 696 ohms hot.

TPS calibrated to .48 closed. 4.5 open.

Stock PCV system, new valve. No vacuum leaks.

Coolant system works, no leaks or fluid loss. Fan kicks on. Thermostat opens and closes.

Everything seems to test out good, except my emissions. And the car runs smooth with expected power. At times the idle will fall to 500rpms without the introduction of lights, heater, etc. What am I missing?

 

After inspecting the newish o2 sensor, which was very clean, but for a thin ring of carbon buildup around the base, I probed.

Heater circuit on sensor, 16.5ohms

Heater on engine harness, 11.44v

I also monitored the signal with a multimeter. On start up at idle the signal fluctuated from .8v - .9v and cycled down to .1 - .2 and then returned to .8 - .9 it did this several times, which I believe to be expected.

Once fully warm the signal fluctuated at .1 - .2 while idling at 750 - 800 rpm.

Holding the rpm's at 2k, the signal sat below .2v

Do these findings describe a healthy sensor?

 

Anyone have experience here, testing an o2 sensor? Are these readings acceptable?

I'm running a Magnaflow CAT, around 2k miles, do the numbers on the smog check point to an inefficient CAT? I stopped in a local garage and the mechanic believed so. Would you agree?

 

Would a metal matrix CAT bring these numbers down, versus the ceramic matrix CAT which is in use?

 

Or could this rich condition be an after effect of installing a header (DC 4in1) and intake (AEM, short) and not adjusting the fuel/ignition map, getting it tuned?

"Where is the love?" I asked my wife.
"Isn't it out there in the garage?" she answered back.

If I put the stock intake and header back on would I be in business?

 

Why do I feel like I'm talking to myself?

 

You can try it. Any i/h/e do effect performance, in this case, stock specs. I'm not really experienced with tuning, maps and etc, but more air flow into the cylinder would require more gas in the cylinder. Not sure if the ecu responds to more airflow by regulating fuel or if it stays within its map permenantly only effected by throttle. Now I'm just confusing myself.

 

Ceramic core cats do not do as good of a job as the precious metal matrix type. Any muffler shop should be able to shoot the cat with a thermometer and tell you how well it's working.

The stock cast iron exhaust manifold would be likely if you were still obd0, but since the obd1 o2 has a built in heater, then I kind of rule that out as a likely culprit.

You can't test a narrowband o2 sensor with a multimeter. You need a lambda scope. The o2 should be fluctuating up and down constantly. As it reads a little rich or a little lean it constantly adjusts voltage like a sine wave. A properly working o2 does this quickly. An old lazy o2 does this with delay and a completely dead o2 does nothing.

An aftermarket Wideband o2 sensor give a solid voltage for a content reading rather than a high or low reading like the narrowband.

How are you testing?
Do you drive around for a good 20 minutes beforehand to get the cat heated up?
When you get to the test station, are you leaving the car on while you wait?
Are you waiting less than 5 min before they can test you?
Are you using the lowest octane fuel you can run?

 

If you can, put on your oem air intake and exhaust manifold. Make sure your intake filter is clean. If you want, it will help if you retard timing to 14 BTDC. Change your oil before you smog it as well. Running a higher octane fuel will help. And you can even us this product called HEET. And as 4drEF said, drive around for a good amount of time right before you go to the smog station. Get your cat nice and hot. All of these suggestions helped me pass bar/smog my 01 gsr in my 90 crx.

 

Low octane burns more completely so i wouldn't suggest a higher octane if he running rich.

 

I understand a multimeter is great when testing static or near static signals and is not sensitive enough to measure the voltage fluctuation of an o2 sensor, however, it should be able to give me an idea where the sensor is reading. Correct me if I'm wrong…

First of all, I'm dealing with a Denso o2 sensor, which requires splicing. Using a new sensor I left bare the signal and ground wires, allowing me to connect my alligator clips.

From a cold start at idle I got .60v, which held for a good minute and then dropped to .05v for 10 seconds. Naturally the idle dipped each time the mixture went lean. As the engine warmed, the AFR cycled through this rise and fall several times, but each time the voltage increased until it reached the .80v (+ -) -- a rich mixture.

With the engine hot, each time I goosed the pedal the voltage would jump. Holding at 2k rpms I measured .75 - 69v. -- rich.

I came back two hours later, and ran the test again. This time at idle the voltage began at .47 -- stoic, but soon cycled up to the .70v. Holding at 2.5k rpms .78 to .79v was measured.

I swapped in the stock Air Intake and filter and got the same results. I also tested the ECT output -- 222ohms. And the IAT -- 773ohms. Keep in mind the TPS has been calibrated.

The MAP I wonder about. KOEO, the signal wire sends 2.9v and at idle it measures 1.**v. When I open the throttle it rises slightly but does not return to 2.9v -- I gather the vacuum load does not fall away until the throttle is fully open and held there, which I'm not going to do sitting in the garage.

After seeing no improvement when swapping the air intake I'm not inclined to put back in the stock header. The o2 sensor is heated, and once at temperature it's placement should not matter.

The last time I tested was after a one hour drive and I got right on the dyno. Running the mid-grade gas. I'll put 87 in for next test. As for retarding the ignition, would that not increase HC?

After all this testing it looks like Honda's fuel maps are just on the rich side. My nightmare is spending 4 bills on a metal matrix CAT and have it fail again.

 

Is it a stock ECU? You can compare your ECU to a chipped (google pic) ECU and make sure its not.

Check resistance of injectors. Wouldnt hurt to get them cleaned

Change your oil

How long is your average drive? If the engine doesnt get up to operating temperature it would be running cold rich air/fuel mixture. That extra fuel is then built in your oil

 

ECU is stock.

Injectors test out good. Even burn across plugs.

Since I haven't passed smog, the car is seldom driven. Oil had maybe 200 miles for last test.

 

You really need an oscilloscope to perform this measurement properly. A true RMS meter with the averaging function can be used if you know how to use it and when and only in limited scenarios.

Oil saturated with fuel will cause those high readings. Pull your dipstick and smell for gas. Compare the smell to new oil or another car if you can.

Is this a new build? Or did this happen without any changes?

 

Copy that on the scope. Strictly shade tree here.

Pulled dipstick after a run, ran oil between fingers and smelled -- no scent of fuel.

Not a new build, low mileage swap, jdm version of d16z6. Runs great, just took a drive through the mountains. Nice power.

 

Does it sometimes bog during acceleration?

Whats up with plug 3 in the group? It looks black. Rich perhaps?

This is a stretch but one of your injectors may be sticking open = rich. It might be the same cylinder as plug 3.

Other than that good diagnostics. You rank quite a bit over a shade tree mechanic

 

Never bogs. Smooth, linear acceleration.

I noticed the third plug, too. I'll hop it over to another cylinder, clean the plugs and see if the condition follows.

Thanks for your help.