Sudden drop in fuel economy, why?

[Pandasaurus]

As some of you may have seen in my previous thread, I have a CEL for either a failed/failing catalytic converter, or O2 sensor (could be either, according to people on other forums with the same code). As of my last fillup, I've gone from ~250 miles/tank to... Roughly 180. (quarter of a tank left, I'm at 150)

So, the question is... Would this indicate that it's the O2 sensor, or could it be either? Or something completely different?

Car is '02 Impreza 2.5TS, 207k miles.

 

[phucheneh]

Please don't let anyone convince you that a faulty oxygen sensor is causing horrible fuel economy. They will try.

I would be highly suspicious of a plugged converter.

You have P0420, right? And which O2 code?

The catalyst code would have to come first, I would think, as the ECM isn't going to run the catalyst monitor with a faulty O2.

edit: If you have a scan tool, look for an unusually low voltage from the rear sensor. And probably slow switching from the front.

 

[Throckmorton]

Not sure if this applies to all cars but I believe the post-cat O2 sensor only tells the computer whether the cat is working. The pre-cat one is the one that affects air/fuel mixture etc. It could be an intake leak for all you know. I had a similar code on my BMW E36 and I never diagnosed it.

 

[Zenmervolt]

Please don't let anyone convince you that a faulty oxygen sensor is causing horrible fuel economy. They will try.

They will try because it can. At least, if the bad reading is coming from the pre-cat O2 sensor.

If the ECU is not getting a proper signal from the pre-cat O2 sensor it will default to a "safe" fuel map which is richer than the mixture will be when the O2 sensor is functioning and providing proper trim information.

If the failed sensor is the post-cat sensor, then I agree with you.

ZV

 

[phucheneh]

I'm well aware that open loop is not as efficient as closed loop. But open loop at operating temp is not going to be grossly richer. Disconnect the upstream O2's, for all it matters, and mileage should not take a nosedive. If it does, something is wrong on the intake side.

And if he's just nitpicking minor MPG differences...well, my advice would be 'don't nitpick minor MPG differences,' because you'll drive yourself nuts over nothing. Especially when there are obviously other concerns with the vehicle.

Aside from that, O2's don't generally start outputting higher than normal voltages. A faulty O2 is far more likely to output a low signal, which would indicate a rich condition to the ECM and cause it to dial back fuel trim. Likely until it hit the wall and set a LTFT code. And what causes a high O2 output? A leak. Which is not a bad O2 sensor.

Fact of the matter is, common internet 'knowledge' is: Bad mileage without CEL? Change your O2's! They're totally a maintenance item! To which the proper response is: 🙄

 

[Zenmervolt]

I'm well aware that open loop is not as efficient as closed loop. But open loop at operating temp is not going to be grossly richer. Disconnect the upstream O2's, for all it matters, and mileage should not take a nosedive. If it does, something is wrong on the intake side.

Define "nosedive." It should definitely be a noticeable difference in mileage, though I agree it wouldn't be something like half.

The OP has seen a reduction of about 20%; not catastrophic, but also not merely nit-picky.

Fact of the matter is, common internet 'knowledge' is: Bad mileage without CEL? Change your O2's! They're totally a maintenance item! To which the proper response is: 🙄

And this is relevant to the OP's situation how? The OP has a a CEL which indicates and out-of-range O2 sensor reading. In that situation, you absolutely do replace the O2 sensor. That's the first thing you do. You don't go chasing other possible issues before fixing the known problem and seeing if that brings things back into place.

I agree that the concept of the O2 sensor as a "maintenance item" is outdated and that changing the sensor on an OBDII vehicle without a CEL is unnecessary, but the fact of the matter is that the OP does have an OBDII vehicle and there is a CEL indicating an out of range O2 sensor reading.

The OP needs to replace the O2 sensor, clear the code, and go from there.

ZV

 

[phucheneh]

It's relevant because no code = people recommending O2's, and code = the same recommendation, only now people will be certain that that's the issue because 'the computer said so.' When in reality, the computer is simply saying 'this reading is outside of my specified parameters,' and that's it. Proper diag still needs to be done- that's my only real point there.

What's truly relevant, is that without him specifying the code, any debating we do...is not relevant. 😀

Need to know upstream/downstream, sig low/high, or if it's just a heater code.

...actually...reading his post again, it sounds like he may not even have an O2 code.

I have a CEL for either a failed/failing catalytic converter, or O2 sensor (could be either, according to people on other forums with the same code)

Sounds like a catalyst efficiency code + the typical internet 'wisdom' that neglects actual analysis in favor of the 'it's impossible to diagnose; start replacing things' approach.

I'm still gonna put all my chips on black...the color of his melted catalyst material.:awe:

 

[phucheneh]

OP, in addition to sharing the DTC with us, a link to your other post helps:

http://forums.anandtech.com/showthread.php?t=2313662&highlight=

Did you find a vac leak? Still have the stock cat?

FWIW I've used Walker (makes Dynomax stuff) direct fit cats with no issues. Never used Bosal, but it's probably about the same. Not sure why you're eyeing Magnaflow...is it an OEM-style converter? Not 'high flow'?

edit: Oh, also, your pic...that's a good bit o' oil. Probably coming from multiple places. PCV system may be plugged. May also just have tons of blowby, which is heavily exacerbated by an exhaust restriction.

 

[Pandasaurus]

Starting at the first reply... More or less.

It was a while ago that I pulled the code, but P0420 sounds right. No O2 code. (At least, not when jlee and I pulled them. Might have changed now, I'll try to check Tuesday when I have time off work)

Based on what I've been reading, it's hard to get enough information to properly determine if it's the cats or O2 sensors with the average code reader (IE: the cheap-o Auto Zone type stuff I have access to). If I remember correctly from my previous reading/research, more than one source mentioned that the only reliable way to find out is by taking it to a reputable mechanic with "real" equipment, which typically costs about the same as just replacing the O2 sensors (the more common and much cheaper solution, but not necessarily the correct one in my case). I've seen a couple people who have had the same code for a year or more after replacing both O2 sensors and cats. In some cases, the code has been caused simply by replacing the stock headers with Borla. (As far as I know, mine are stock, as is... Everything else on the vehicle. Even the pseudo-stock stereo...)

If it's relevant (I don't remember if I mentioned this in my other thread that I forgot to link, thanks for that by the way), the first time the code popped up was while driving on the highway at 65, after about 5-10 minutes. Cleared it at jlee's house when we changed the oil, it popped up later (I want to say a week or so, but I don't remember) during normal street driving. To clarify, the light came on while on the highway, not after getting off and stopping/slowing down.

Initially, I thought the cats were the most likely cause, being the high miles on the car and knowing that in all likelihood the cats have never been replaced, but now I'm not so sure. My bank account would like it to be the O2 sensors, I know that much. lol

Regarding your last post, I'm gonna show my car-noobish-ness here and say...

I don't know what a vac leak is, or how to check for one. I can Google it, though! Or just ask jlee. *snicker* >.>

As far as I know, yes, still the factory cats (unknown if any of the PO's replaced them at any point in time, I'm guessing not). Eying Magnaflow because they're OEM style, not high-flow, and I recognize the brand, and the "universal" option is cheap. No idea of their actual reputation, though. (High-flows of any brand have a reputation for making it nigh impossible to get the P0420 code to go away, making it impossible to pass emissions in AZ)

Oil: Yes, it's alot of oil. My confusion comes from the fact that last time I checked it (need to remember to do that in the morning again...) the oil level is still the same as when I changed it... What, 2 months and ~1K miles ago? I cannot explain the large quantities of oil all over the bottom of the engine (and I think transmission, among other things in that general area) and a potentially expanding amount in the driveway when all the various fluid levels haven't changed... (I'm not sure how much of what is in the driveway is oil, I keep forgetting to check when it's daylight out... I think it's mostly water, though. Or washer fluid... Or coolant... It doesn't look like oil when I've checked at night with a flashlight. >.>)


*ahem* On a related note... This car has many little things all over that should be replaced and/or fixed at some point in the near future, but that's a long story. One of which being the front CV boots/axles (and the struts... And the head gaskets... And the... *cough* I digress). My dad proposed a theory that some or all of the "what appears to be oil" all over the bottom may actually be grease from the CV axles. Does this sound like a plausible theory? Also, the first sound I described in my other thread... Could that be caused by potentially damaged CV axles/joints?

 

[Zenmervolt]

It's relevant because no code = people recommending O2's, and code = the same recommendation, only now people will be certain that that's the issue because 'the computer said so.' When in reality, the computer is simply saying 'this reading is outside of my specified parameters,' and that's it. Proper diag still needs to be done- that's my only real point there.

At 205,000 miles, if he's reasonably handy on his own the easiest and cheapest way to diagnose the issue is to replace the O2 sensor and see if the code re appears. If the code comes back with a known-good O2 sensor then it's the cat. The O2 sensors are ~$50 for his car ('02 non-turbo Impreza) for either the upstream or the downstream sensor. If he's reasonably handy it's a fairly simple DIY job and it will cost him less to do it that way than it will cost him to have "proper" diagnosis done.

Although, given the laundry list of things that he's now mentioned are wrong with the car the practical answer may just be keep driving it until it catches fire or something. 😉

ZV

 

[phucheneh]

I still don't like replacing O2's for a catalyst code...can't recall ever having that work. Even on the Volvos that would throw that stupid 'O2 integrator' crap, it was always a leak or a bad cat.

If he does replace an O2, he'd wanna do the upstream, I'd say. Goes back to the sensor being less sensitive/responsive than it should be...since the ECM is comparing front and rear signals to determine catalyst operation, it would be a bad front sensor that would make the waveforms more similar, not the rear, right?

Also, I realized I brainfarted up above and stated the voltage stuff backwards. A sensor that is old would have a low voltage, and the ECM would try and richen the fuel mixture. I said it would lean a low voltage, which is obviously backwards. But I'd still make the same point about fuel economy, depending on how severe we're talking here.

The ECM will, at most, increase fueling by 25% before setting a fuel trim code. So in order to cause what I would consider a noticeable drop in economy, the sensor would have to cause fuel trim to go from around baseline to a butter-zone around 15-20% without continuing to increase. <10-15% and I think you'd have to be REALLY anal to say 'man, my mileage has gotten terrible!'

OP's numbers would be about a 30% decrease in economy, and he's stating it's pretty sudden...which doesn't quite make sense to me, as something that would cause a sudden decrease in economy would likely manifest other symptoms. O2's don't suddenly wear out, and cats don't suddenly get clogged. Usually. The exception would be a heavy misfire condition, or just some really bad luck...as in, cat starts to come apart internally and has a bit of a chain reaction that results in greatly decreased airflow. And that's just not nearly as typical as the usual scenario of 'broken chunk is rattling inside converter but not causing any major impact.'

It would be very helpful if OP graph the O2's. Unfortunately, the equipment to properly do so is not cheap. You can try and graph them with a scan tool, but even a high end (or dealer) tool will not communicate quickly enough to get an accurate graph (gives you a decentish idea of how they're operating, though). I usually just go off of the median voltage and accept that I won't be able to discern anything else until I pony up the cash for the right tool.

In response to some of the other stuff in your last post, OP-

Catalyst monitor typically runs on the interstate. The ECM wants to see you on the throttle (no brakes) and cruising at a brisk speed before it will start and complete it. So that's why your code came at that time.

If it's consistently failing the catalyst monitor, one slightly-ghetto test would be to basically beat the snot out of the car with around-town driving to try and get the cat as hot as possible, then run it on the interstate (...speeding is a plus) and see if the hotter converter causes it to pass. This is both indicative of a bad converter, and a technique I use to 'cheat' emissions with a bad one, heh.

You can also check the cat with an IR thermometer. Outlet should be hotter than inlet. Thermometers that are accurate at those kinds of temps aren't cheap, though. It's worth noting, in case it's not clear, that a bad converter will not necessarily cause a mileage decrease. If the catalyst material is simply 'used up,' you can get P0420 with an engine that runs fine. But I've seen engines that are choking to death also throw the same code. And that can be a bit tricky to diagnose.

'Vac leak' was 'vacuum leak,' i.e. an intake leak between the throttle plate and the cylinder head's intake ports. What's your idle speed like?

 

[XavierMace]

Oil: Yes, it's alot of oil. My confusion comes from the fact that last time I checked it (need to remember to do that in the morning again...) the oil level is still the same as when I changed it... What, 2 months and ~1K miles ago? I cannot explain the large quantities of oil all over the bottom of the engine (and I think transmission, among other things in that general area) and a potentially expanding amount in the driveway when all the various fluid levels haven't changed... (I'm not sure how much of what is in the driveway is oil, I keep forgetting to check when it's daylight out... I think it's mostly water, though. Or washer fluid... Or coolant... It doesn't look like oil when I've checked at night with a flashlight. >.>

It's not water/washer fluid. Otherwise it wouldn't still be there. Definitely looks like oil to me.

 

[phucheneh]

It's oil. The only things that can typically be confused with oil would be- trans fluid; but the red color almost always shows through, unless perhaps the ATF is kinda dark, and probably also mixing (externally) with leaking oil. Or power steering; but it's usually a little thinner (although obviously the system could also be filled with the same fluid as the trans; but from the factory, it's not) and typically not hard to locate the P/S pump (or a hose) as the epicenter of filthiness.

I'm not good with Subie engines, but that appears to all be coming from the passenger side block/head area, and probably the rear main, as well. But oil leaks can be tricky since you've got gravity pulling everything down, airflow at speed pushing everything back, and the cooling fan just generally throwing shit everywhere. That's why I say to try and find fresh oil as high up as you can, and that's either your source, or at least a not-insignificant contributor. Stuff on the bottom like front/rear crank seals and the oil pan seal are hard to be sure about when there is also oil coming from up above them.

The only times I've mistaken something thin as oil (done it with coolant, even) is usually after an oil leak repair...the oil has stopped, but everything's generally still grimy and nasty. So you can get a slow coolant leak wetting the gunk and washing it downwards, and by the time it starts to drip, it looks more like oil.

You can diag the rear main by pulling any inspection covers off of the bellhousing and looking for oil up inside of there.

 

[jlee]

Clearly you should buy me a steam/hot pressure washer so I can clean the bottom of the car so we can accurately pinpoint the leak's location.

😀

 

[Pandasaurus]

I can get you 10% off... Otherwise, my money is currently going into a lab setup and buying a monitor, because Xavier wants his back for some silly reason. 😛

Lab = Real Job = Real Money, so it currently takes priority over... Pretty much everything, at this point. Except possibly doing the minimum to keep the car running until the above happens. >.>