cat converter, emissions test

I can't imagine how you would conduct an emissions test w/o an exhaust probe, but most places that require an emissions test will automatically fail a car with a CEL lit. And having your own ODB-II reader is always a good thing; if nothing else, it will allow you to clear the CEL. Check out the Elmscan products on-line. Aftermarket CATs are frequently better than the OEM parts they replace, but the Subaru CAT(s) is/ are, IIRC, in the "Y" pipe, which complicates things. And after 150K miles, removing an O2 sensor can be challenging. What you might do is hit a salvage yard. Technically, they aren't supposed to re-sell emissions components, but if you ask for a "Y pipe", rather than a "catalytic converter" you might bypass this silly-assed regulation. Good luck! ByeBye! S.

Steve Jernigan KG0MB Laboratory Manager Microelectronics Research University of Colorado (719) 262-3101

So the code was the old good P0420. Is there a chance that this is one of the sensors? Can I read the voltages from the sensors directly , i.e. can I get the raw data from the computer, instead of relying on cryptic codes? After all, it's just voltages. If I must go the cat replacement way, can I replace only one of the converters with an OEM (from what I understand the Impreza has 2)? I am driving this car at most for 2 more years, don't want to give away

1,300.

Hi again RCX!

On Sat, 28 Aug 2010 10:32:23 -0700 (PDT), runcyclexcski wrote:

Well, it _could_ simply be a glitch. Could also be the O2 sensor, or an intermittent in the sensor wiring. I'd try clearing the code and see how persistent the CEL is before getting too excited about it. And while you can certainly measure the O2 sensor output with a DVM, you will find it challenging with the car idling in the driveway, much more so while driving. A decent OBD tool will let you monitor these parameters (and many more) in real time as you tool along. The better ones will log the data for post analysis, which can be very handy. The Elmscan unit plugs into a USB port on a laptop. It comes with basic software, and much more is available as commercial software, shareware, and freeware. I'm sure there are other good packages out there, but I have used the Elmscan stuff with good results for several years, and so can heartily recommend it.

I know the first CAT on the Impreza (on any Soobie I've ever looked at FWIW) is in the Y pipe, right where the headers come together. I think the second one is as well, but it might be in the pipe between the Y and the muffler; you'll have to look. In any event, the code is pointing to the first one, the one in the Y. The one time I had to deal with this, I simply tossed a used Y pipe at it. The O2 sensor (at least the first one) is also in the Y pipe, so if you get a "known good" one, you kill two birds with one stone. I just don't know about aftermarket CATs for this application; presumably they are available. Feel free to contact me directly if you want more info on the scanner. ByeBye! S.

Steve Jernigan KG0MB Laboratory Manager Microelectronics Research University of Colorado (719) 262-3101

Could also be ethanol gasoline. Reset the code and fill up with Shell Ultra. Shell has been on record

- at least in Canada- saying there will NEVER be ethanol in Shell Ultra.

I've had the issue with a Mazda Miata, and switching to Ultra solved the problem. Put regular in and the CEL comes back on within 2 tanks.

Whatever you do, don't drill out a sparkplug anti-fouler and put it between the 2nd O2 sensor and its mounting bung.

This would make the computer happy for about 6 bucks but would be in violation of town, county, state, federal, solar, and galactic law, and would poison the world (assuming the cat is actually bad and the code is not spurious).

Dave

Here in Northern Virginia they skip the dyno test for AWD cars because the test station dynamometers can't handle AWD. I haven't taken my wife's Forester for emissions testing recently, but I think they do run an idle emissions test on it.

If it were me, I would -

1) clear the code and see if it reappears. You need to drive at least 100 miles without the CEL before it will disappear from the computer's memory. 2) If 1 doesn't work, I'd replace that other O2 sensor and see what happens. 3) If neither 1 nor 2 work, I'd definitely go the 'used y-pipe' route as suggested. 4) If no luck on 3, I would locate a different shop. $95 is way too much for a 'computer diagnosis' (BTW, Autozone will pull the codes for free). Call around and get a few estimates. Even try the dealer.

Dan D '99 Impreza 2.5 RS (son's) Central NJ USA

Apparently this all depends on the state.

In IL, I've never had a tailpipe test. If they plug in a code scanner and all's clear, you pass. If you have a CEL (or have teh uncleared long term statuses that won't be clear if you clear the codes just before the test), you fail.

Just out of curiosity... what in the fratz is a sparkplug anti-fouler?

-rick-

its a spacer designed for the spark plugs to lift it up some and prevent fouling. they drill them out and put them between the o2 sensor so it reads sluggish, and fools the computer that the cat is doing its job

Way I understand it is the O2 sensors sees less O2 when out of the exhaust stream, which is the condition it is 'looking for' to confirm converter operation.

Many a sad bloke has replaced both sensors and the cat only to have the code return.

Dave

Notice I said "Northern" Virginia. We're in the DC suburbs and special rules apply here that don't apply to the rest of the state. (Well, maybe Norfolk or someplace also has enough air pollution to earn special rules; I wouldn't know.) The Chicago area may have enough pollution to merit emissions tests that aren't required elsewhere in Illinois.

Has the OP tried running ethanol-free gasoline yet as I suggested last week????? Definitely worth a try , for diagnostic purposes if nothing else.

I've had it happen on Mazdas and Hondas - so nothing saying it can't be the problem on a Soob.

Try Shell Ultra - it has no ethanol in Canada and I suspect that is also true in the USA. (they put 10% in regular, none in premium, so averaged 5% on mid grade.)

Thank you guys for all the great responses. I will get a computer to clear the code, drive for 100 miles, and see what happens.Tthen I will replace the sensor, then I will replace the Y-pipe, etc.

A colleague of mine has suggested that what triggered a CEL could've been a single instantaneous, no-reproducible, event, not achievable under everyday conditions. Coincidentally, when the CEL light came up I was driving on a rough dirt road in WV.The old engine probably was struggling on a random rough uphill and triggered the emissions above the allowed threshold for a few seconds. In that case, the on-board chip will keep that event in the memory as a faulty event, even though it hasn't happened after that instance.

If CELs are indeed designed to go off this way it's quite disappointing. Hence my interest in getting a computer to get the raw data, not a silly binary event (CEL off -> CEL on) that costs $1,300 to take care of. Sounds like a set-up to get the consumer pay $$$ for a minor problem. I will start working on it tomorrow.

I'm reporting experience from Chicagoland.

Not sure if anyone ever has a dyno test here but I've ben here a long time and it's never involved any tailpipe probes or much else other than a plug in to the odbii port.

To the OP -- I had a p0420 catalytic converter efficiency code off and on for a long long time. I recently had a code come up about the knock sensor. Replaced the knock sensor (which is trivially easy to do yourself btw) and haven't seen any code for 3 weeks now.