Conditions for Closed Loop

Will the ECM still close the loop if the catalytic converter or downstream O2 sensor are bad?

I guess that depends on the severity of the failure. It would have to enter closed loop to determine the catalyst had failed, and I don't recall having ever seen a P0420 cause a vehicle to not go into closed loop. It may suspend other monitors however.
If the downstream 02 were disconnected all together I imagine it could effect loop status, or if it was sending a signal so erronious that the pcm could no longer maintain fuel control it might default to open loop, or open loop fault.
Considering the variables, I'm not 100% sure it's possible, but I don't think it would be norm.

As long as the engine is warm with no misfires and the upstream O2 is working the engine will remain within closed loop provided the fuel trims as are within normal parameters..

I've got a PO420 and the upstream sensor reads a very steady 300mV (backprobed w/scope) with the vehicle fully warmed. When I blip the throttle, the upstream sensor voltage rises and falls, then settles back to 300mV. The downstream sensor sits at around 400mV, rises to near 900mV, drops to 200mV, then back to 400mV when throttle is blipped. The downstream sensor moves rapidly when the throttle is blipped, while the upstream sensor moves fairly slowly when the throttle is blipped.

What I'm really after, is whether the upstream sensor has failed, or is the loop just not closing. I don't have the experience to know what voltage to expect from the upstream sensor under open loop conditions assuming a good sensor.

By the way, the cooling fans come on periodically, and the temp gauge on the dash reads normally - can I safely assume that the ECM is interpreting this as "fully warmed"?

What's the specific vehicle you're looking at?

The upstream O2 sensor signal is unusual. I don't see how it could be in closed loop with that kind of signal.

2003 Subaru Outback, 2.5L

Another data point - I disconnected the upstream sensor, took a spin around the block, no difference in performance. Truthfully, I'm not even certain as to whether it's a narrow band or wide band sensor.

Corrected version ---. It is likely to be a narrow band sensor normally with a [strike]narrow[/strike] wide band sensor there is a very distinctive pattern in response to blipping the throttle followed by a very rapid return to a steady constant voltage very tight control no oscillations. Normal upstream side band voltages vary between car and sensor manufacturer and normall a low voltage is returned for a rich condition and higher voltage for lean. The voltage shown in the PID is an equivalent voltage created by the engine computer as the sensor signal works on current not voltage. All very confusing I have a graphic that illustrates this normal voltage pattern in response to blipping the throttle ( I call it the dippsie doodle test) but I can't access it just now as I am holiday.

So I looked up the wiring diagram for your '03 Outback, and was very surprised to find a wideband air/fuel upstream sensor.



Depending on which wire you were probing, that 300mV reading you saw may be entirely normal. It sounds like you were on the sensor signal wire based on your description. :huh: I'm not super familiar with Subaru A/F sensors. Where was your scope ground connected?

At this point, I think you'll need scan data to make a definitive call on your P0420. If that's not an option, then you can try observing your downstream O2 signal wire during a highway cruise with the engine warmed up. If you see the downstream signal switching rapidly, then it's game over for the cat. That also means that you're very likely to be in closed loop, since the downstream sensor wouldn't be able to switch across stoichiometric if the fuel trims weren't keeping the mix there.

Andy.MacFadyen wrote: . Normal upstream narrowband voltages vary between car and sensor manufacturer and normall a low voltage is returned for a rich condition and higher voltage for lean.

Now that you're confusing me....You didn't happen to mean WIDEBAND voltages in the quoted phrase, did you?

juergen.scholl wrote: Now that you're confusing me....You didn't happen to mean WIDEBAND voltages in the quoted phrase, did you?

:lol: I think you're right. Give him a break, he's on vacation.

Too right guys my excuse is I had a bigly large Aperol Spritz last night and some Bardolino Red wine both of which are local to this part of Italy and although I never take much alcohol when in Italy follow the local customs. This is a crazy place it took me 5 minutes to walk 30 yards yesterday because of a traffic jam composed of 20 Ferraris .

Now back on topic ISTR Matt on Schrödingers Box YouTube channel has a video on known good Scooby Wideband sensor.

paulsomlo wrote: 2003 Subaru Outback, 2.5L

Another data point - I disconnected the upstream sensor, took a spin around the block, no difference in performance. Truthfully, I'm not even certain as to whether it's a narrow band or wide band sensor.

Most cars drive fine with the upstream disconnected they will use a default slightly rich fuel map without closed loop. It might be an issue if you had the car at a high altitude such as Mexico city or Colorado or in exceptional high ambient temperatures. Back before electronic fuel injection all gasoline engines ran very rich by modern standards, so rich that engine life was much shortened by bore washing. Within reason a fuel air mixture that is over rich will produce loads of power at the expense of a lot of pollution.

Andy, enjoy your vacations as much as possible!

Are there still some Vespa scooters from the 60s or 70s around? Boy, they ran rich as well, but on 2 stroke oil (anywhere from 1:50 to1:25).

I'm not sure that wiring diagram is correct - my upstream sensor has four wires. It does appear to be a wideband sensor though. It's a Denso, PN on it is 22641AA140, which corresponds to Denso 2349015.

I used an all purpose digital storage scope and probed the two signal wires (white and blue) taking one of them as ground. I tried it again last night with a DVM, same result, 300mV between the two signal wires. I did measure each signal wire to ground; 1.9V and 2.19 V, using the intake manifold as my ground. The difference voltage across the two signal wires did change when disconnecting a vacuum line from the intake manifold, increasing, then settling back to 300mV with the vacuum line still disconnected.

My previous "driveway" test showed the downstream sensor sitting steady, no switching.

I think you're right - I'm at the point where I need to see scan data. As a DIY'er, I don't have a scan tool. I was thinking maybe a bluetooth dongle and the torque pro app?

PS - I did look at that video, Andy. The voltage displayed on the phone app is right around 2.1V

Nice, thanks paulsomlo. That's what we needed to know. It's definitely an A/F sensor, just not the kind shown in the diagram.

A Bluetooth dongle and an app would be just fine for this. There's a huge selection out there - I have this one from Amazon, works perfectly:

www.amazon.com/BAFX-Products-Bluetooth-D...canner/dp/B005NLQAHS

FWIW, I can't stand Torque. :lol: Setting it up to see what I want takes WAY too long. I'd suggest giving OBD Auto Doctor a try. Super easy to use, and the graphing capabilities are much better. The data logging and file export system is fantastic for sharing data, too.