Oxygen sensor suggestions please... OR...

[Walking-Tall]

... both for reliability as well as inexpensiveness...

OR, read on... ( * WARNING - not for the squeamish of EFI * (EFI guys, please help!) )

I don't have the time or the money right now to YANK all EEC-IV and put carb/pump/duraspark on this '86 CFI 3.8L V6, as has been in the back of my mind since getting this car. Anyway, it's O2's are now no longer "switching" when the car is fully warmed up, both staying at 0.90-0.95 volts, a tad rich, at ~14.45:1 at idle. (Open loop?) I know, a voltage reading at all from them means they're working, and the grounds are good, and that computer may not be adding/subtracting fuel, which is what makes O2's "switch" back and forth just above and below stoich 14.7:1 when in closed loop... my nightmare headache coming true, lol

Before I wander off like a chicken with it's head cut off buying oxygen sensors that maybe aren't needed....

Here are the umpteen checked and verified to this moment facts, if some of you fine folks here have some ideas for me:
- fuel pump working
- fuel filter good
- fuel pressure is at 40psi
- fuel injectors clean, not leaky/leaking, and both reading 2.6 ohms resistance
- TPS voltage is at 1.0 at idle and smooth with no drop-outs to WOT
- no vacuum leaks at/below throttle body or anywhere else (creating a vacuum leak sent O2's lean, normal)
- KOEO/KOER codes only about missing air injection (non-issue)
- cylinder compression tested and all six cylinders 145-165psi (continued head gasket coolant containment remains to be seen)
- replaced 195-degree thermostat recently, due to factory temperature gauge reading at the N
- cooling system burped and full
- temperature sending unit threads grounded (removed some teflon tape that was on threads)
- warmed it up real well and hoses hot, feeling like 195 to me, hotter than before
- thermostat not opening before it should
- factory gauge again reading at the N and lower (maybe a hint that it's remaining in open loop, never entering closed loop? But that's got to do with ECT, not the gauge sending unit... oh, mighta answered my own question there, lol... better check ECT voltages... crap, lol)
- car appears to run really well, with no rich stink or black tail pipe either...

I think that's everything (sorry about this long-winded post. I want to cover all bases, and I'm to get the car out of where it's stored, where I can work on it, tomorrow or very soon this week), but I'm hoping you might see something here (or not here) that I'm not seeing...

Mike

[JACook]

Seems you've pretty much eliminated all the suspects, apart from that ECT sensor. This certainly doesn't
look like an O2 sensor problem, unless there's a problem in the wiring from O2 sensors to ECU. Closed-loop
generally requires the ECT up to operating temperature, and voltage being produced by the O2 sensor(s).
But I would expect KOEO/KOER code 21, if the ECT isn't in the normal operating range. Then again, if the
O2 sensors are always at .9V, you might should be seeing KOER 42.

How are you measuring that the sensors are not switching?

[Walking-Tall]

Seems you've pretty much eliminated all the suspects, apart from that ECT sensor. This certainly doesn't
look like an O2 sensor problem, unless there's a problem in the wiring from O2 sensors to ECU. Closed-loop
generally requires the ECT up to operating temperature, and voltage being produced by the O2 sensor(s).
But I would expect KOEO/KOER code 21, if the ECT isn't in the normal operating range. Then again, if the
O2 sensors are always at .9V, you might should be seeing KOER 42.

How are you measuring that the sensors are not switching?

Thanks, Jeff. Yes, I'll spend later today/tonight checking the ECT sensor voltages. From the Probst book, "How to Understand, Service and Modify - Ford Fuel Injection & Electronic Engine Control" (the "bible" for anybody do-it-yourself-ing with this stuff), ECT's output values for cool/warm/hot are ~3.0/~1.5/~0.5 volts. Good point, scary but good, lol, on the wiring from O2 connector to ECU. Thank you, I'll check that too.

Before, I was getting codes 41 & 91, O2 circuit indicating system lean, left & right. I took the throttle body off and discovered one of the base gaskets was toast. Replaced that. Checked for O2 switching, and they were, up and down normally. Then only got KOER codes of 94 & 44 for air injection, which is gone, and a non-issue. I'll check varying ECT sensor resistances and voltages, the O2's again and their wiring to the ECU, and do the KOER/KOEO again.

I'm measuring running O2 output voltages, with the engine up to temperature, with a digital multi-meter, between the battery negative and the "EGO SIGNAL" plugs inside the 4-pin female connector plug from them, where I last got steady non-changing 0.90+ volts from both of them. The other 2 are the grounds, which are good, both physically at their terminating-to-engine bolts, and continuity to the 4-pin plug.

I'm going to need a LONG length of wire to verify continuity between sensors and ECU behind the kick panel... that'll be funny, like Dr. Frankenstein's laboratory, hopefully without all the sparks (I'll disconnect the battery), I'll take pics, LOL!

[Ethyl Cat]

I skimmed your post so forgive me if I missed it.

I did not see where you made an attempt to drive the sensors lean. Have you done that? In my experience O2 sensors fail lean output.

If you can drive them lean and they read lean it is not the sensors.

[Walking-Tall]

I skimmed your post so forgive me if I missed it.

I did not see where you made an attempt to drive the sensors lean. Have you done that? In my experience O2 sensors fail lean output.

If you can drive them lean and they read lean it is not the sensors.

Thanks, EC, no sweat. Yes, I pulled a vacuum hose and the O2's went immediately lean. I don't think it's them either.

I'll keep checking O2's each step as I go (thank goodness we have them to gauge things) with checking out ECT voltages, resistances, and continuity with ECU, as well as verifying continuity of O2's with the ECU. The fact that it runs at all makes me assume, for now, that the ECU is talking with the two injectors. I think I'm chasing an actual engine low temperature situation, but as I mentioned before, sure seems hot and up to 195-degrees to me (regarding the new thermostat, I learned a very long time ago that "new" does not ever rule anything out of it being a fly in the ointment, though it is not opening prematurely...), or an ECT sensor not triggering closed loop.

[Ethyl Cat]

Thanks, EC, no sweat. Yes, I pulled a vacuum hose and the O2's went immediately lean. I don't think it's them either.

I'll keep checking O2's each step as I go (thank goodness we have them to gauge things) with checking out ECT voltages, resistances, and continuity with ECU, as well as verifying continuity of O2's with the ECU. The fact that it runs at all makes me assume, for now, that the ECU is talking with the two injectors. I think I'm chasing an actual engine low temperature situation, but as I mentioned before, sure seems hot and up to 195-degrees to me (regarding the new thermostat, I learned a very long time ago that "new" does not ever rule anything out of it being a fly in the ointment, though it is not opening prematurely...), or an ECT sensor not triggering closed loop.

What is the Map frequency KOEO and at idle?

What exactly is the manifold vacuum number?

What is the voltage on the sensor ground?

[JACook]

I'm not sure if you've been able to watch the O2 sensors switch in the past using your DVM, but I'm
gonna suggest the sampling rate on a typical DVM is not fast enough to reliably see them switch when
the engine is running in closed loop. A sensitive enough analog meter can see it, but even then you'll
only see small deflections, and you still won't be able to see the slope. I use an oscilloscope for this,
'cause I have one.

If your O2 sensors really are stuck rich, you should be seeing a 42 code.

If the concern is that the engine is running a bit rich, that may well be a symptom of slow O2 sensor
switching, or the result of long term fuel trim from when you were running with the vacuum leak.

[Walking-Tall]

What is the Map frequency KOEO and at idle?

What exactly is the manifold vacuum number?

What is the voltage on the sensor ground?

MAP frequency: I don't have a scope. I could read the signal return voltages, but there's been and is no MAP trouble code set.

Manifold vacuum: Not sure, haven't had the vacuum gauge on it, but it's like a Hoover when I pull a vacuum hose.

ECT: With engine well warmed up, the ECT internal resistance is 2.6 ohms, and it's voltage is 0.65 volts. Good, good.

As well, all continuity between O2 sensors (and their grounds) and ECT sensor through harness to ECU checked fine.

[Walking-Tall]

I'm not sure if you've been able to watch the O2 sensors switch in the past using your DVM, but I'm
gonna suggest the sampling rate on a typical DVM is not fast enough to reliably see them switch when
the engine is running in closed loop. A sensitive enough analog meter can see it, but even then you'll
only see small deflections, and you still won't be able to see the slope. I use an oscilloscope for this,
'cause I have one.

If your O2 sensors really are stuck rich, you should be seeing a 42 code.

If the concern is that the engine is running a bit rich, that may well be a symptom of slow O2 sensor
switching, or the result of long term fuel trim from when you were running with the vacuum leak.

Yes, I've been able to see O2 switching, approximately per second, above and below stoich, in a couple instances, in the past. This digital multi-meter is nothing fancy. So, I dunno... I know there's better ways, but this is what I have.

Yes, no 42 code, but back to 41 & 91 again, lean & lean, which is the mind-blower * ...

With the harness plugs to ECU question, I checked continuity between O2 sensors (and harness connections to grounds good) and ECT sensor through harness to ECU checked fine. ECT itself checked out fine, with engine warmed up good, then shut off, it had 2.6 ohms internal resistance, and momentarily later with the engine running, was putting out 0.65 volts. Interpolating those results with a chart in my Probst book, indicates 180-degrees coolant temperature when I shut it off and measured ECT resistance, and 206-degrees F when running again and checking ECT voltage, which makes all sorts of sense to me, with it's 195 thermostat.

* The mind blower: O2's reading richer yet now, at steady 0.90 & 0.95 volts.

I think this thing needs to be driven, maybe with a bit of injector cleaner through it, and so "adaptive strategy" gets any chance at all to do it's thing, etc. I have no idea how long it's been since it was used regularly, and I disconnect the battery when I leave it. It's coming home so that's it for the work on it (that happens in the daylight anyways). Next is safety inspection and get some miles on it to see how it acts.

Slow O2's, could absolutely be. Thanks very much, Jeff.

[Ethyl Cat]

MAP frequency: I don't have a scope. I could read the signal return voltages, but there's been and is no MAP trouble code set.

Manifold vacuum: Not sure, haven't had the vacuum gauge on it, but it's like a Hoover when I pull a vacuum hose.

ECT: With engine well warmed up, the ECT internal resistance is 2.6 ohms, and it's voltage is 0.65 volts. Good, good.

As well, all continuity between O2 sensors (and their grounds) and ECT sensor through harness to ECU checked fine.

A skewed sensor does not set a code and Ford maps are notoriously prone for that. I have never met an older Ford tech that did not have a known good map in their tool box. Your meter does not have an HZ setting on it?

[Walking-Tall]

A skewed sensor does not set a code and Ford maps are notoriously prone for that. I have never met an older Ford tech that did not have a known good map in their tool box. Your meter does not have an HZ setting on it?

Thanks for bringing MAP up. Here's a little something that I haven't considered: while doing numerous things at the driver's side inner fender, such as a brake line and the master cylinder, I took the coolant overflow tank off when all the coolant was emptied to bring home and clean up since it was filthy. Behind it is the MAP sensor, since they mount them away from engine vibration, and it was hanging loosely by one screw. I found another screw and secured it proper. Do you suppose it might have gotten knocked around enough to be skewed, but still inside range, so as to not set a code 22 or 72, enough to interpret load high enough to cause rich running regardless of O2 signals and ECU fuel metering? I'll grab a replacement for good measure. Thank you, Sir.

[Ethyl Cat]

Thanks for bringing MAP up. Here's a little something that I haven't considered: while doing numerous things at the driver's side inner fender, such as a brake line and the master cylinder, I took the coolant overflow tank off when all the coolant was emptied to bring home and clean up since it was filthy. Behind it is the MAP sensor, since they mount them away from engine vibration, and it was hanging loosely by one screw. I found another screw and secured it proper. Do you suppose it might have gotten knocked around enough to be skewed, but still inside range, so as to not set a code 22 or 72, enough to interpret load high enough to cause rich running regardless of O2 signals and ECU fuel metering? I'll grab a replacement for good measure. Thank you, Sir.

Yes I think it is possible. What does the ecu have to cross reference with? Nothing

It is a real simple computer that has to believe what it sees. Also if the perceived load is high enough it goes into an open loop strategy.

[JACook]

So one thing to keep in mind here, the maximum O2 sensor voltage on a narrow-band sensor is not particularly
relevant. The ECU can't tell between 0.9V and 0.95V, and neither can the sensor. Narrow band sensors are very
non-linear, and are only accurate within a narrow band (hence the name) either side of the stoichiometric point.

If the ECU is setting KOER 41 and 91, then the ECU is not seeing the voltage your DVM is seeing, or you're not
following the correct KOER procedure. KOER requires running the engine at 2000 RPM for 2 minutes, then turn
off the engine, wait 10 seconds, activate self-test, and start the engine. Even with heated sensors, if you don't
run the engine at 2000 RPM for 2 minutes, the sensors will not be hot enough, and you'll get false 41 and 91.

I'll also suggest you should measure the O2 sensor voltage against O2 sensor ground, or better yet, ECU sensor
ground, not the negative battery cable.

One other thing to keep in mind WRT the MAP sensor, the vacuum line to it must NOT be shared with anything
else. It must be a continuous run from the sensor to the port on the intake manifold vacuum tree.

[Walking-Tall]

So one thing to keep in mind here, the maximum O2 sensor voltage on a narrow-band sensor is not particularly
relevant. The ECU can't tell between 0.9V and 0.95V, and neither can the sensor. Narrow band sensors are very
non-linear, and are only accurate within a narrow band (hence the name) either side of the stoichiometric point.

If the ECU is setting KOER 41 and 91, then the ECU is not seeing the voltage your DVM is seeing, or you're not
following the correct KOER procedure. KOER requires running the engine at 2000 RPM for 2 minutes, then turn
off the engine, wait 10 seconds, activate self-test, and start the engine. Even with heated sensors, if you don't
run the engine at 2000 RPM for 2 minutes, the sensors will not be hot enough, and you'll get false 41 and 91.

I'll also suggest you should measure the O2 sensor voltage against O2 sensor ground, or better yet, ECU sensor
ground, not the negative battery cable.

One other thing to keep in mind WRT the MAP sensor, the vacuum line to it must NOT be shared with anything
else. It must be a continuous run from the sensor to the port on the intake manifold vacuum tree.

Points noted, Jeff. I follow that KOER to the letter. They are one-wire non-heated O2's, and in the future I will use their grounds or ECU ground.

I'd be nice if the factory pulled it's head out of it's butt then, because having just been there, I can tell you that the vacuum line connected to the MAP goes into a Tee, which all looks factory to me, and by the looks of things forward of it are the 4 (or 5?) vacuum switches under the coolant reservoir (don't quote me on this, I can't be certain they aren't all separate, because everything's jammed tight into loom), and right behind/beside the MAP on the inner fender is what looks like a teed-in vacuum reservoir "can"... and from there the vacuum line goes straight the intake vacuum tree, but there's a Tee inches prior to the tree, which I plugged a short length of vacuum line upwards with the right size screw to plug it, where a vacuum line WAS going up to the air cleaner.

Thanks to everyone for your help. I'm not able to do too much with it right now until I get driving it around some, hopefully next week.

[JACook]

The factory did not tee anything into the MAP sensor line. MAP is -the- most critical sensor input
to the ECU. You don't want anything corrupting the manifold vacuum signal that it sees.

There should be two distinct lines running from the left fender apron area back to the vacuum tree.
The MAP sensor line is a larger diameter than the line to the reservoir and solenoid manifold. The
line to the reservoir should have a check valve in it. It is normal for the line to the reservoir to have
things teed into it.

There is also a third line that is for EGR vent. That line connects from the EGR solenoid assembly to
the rear of the CFI air horn. Or yours may have a vent filter cap on the EGR solenoid instead, I'm
not completely sure.

[Walking-Tall]

I think I better run a dedicated vacuum line then, because right now MAP is tee'd into a line, so somebody must have done that. I will open up the whole loom and see exactly what is going on.

Yes, the reservoir check valve is there.

It's the second scenario for EGR with this.