DIY: Rebuilding rear catalytic converter, bench testing O2 sensor

[proton17]

In 2019 my rear catalytic converter assembly rusted out and got very noisy. The braid around the flex pipes was very degraded, and the flex pipes themselves were in bad shape with small leaks. There were also a few leaks on the hard pipes downstream.

One of the flex pipes:

Flex.png (5.15 MiB) Viewed 1625 times

Support bracket rusted through:

bracket.png (4.79 MiB) Viewed 1625 times

Bracket2.png (2.97 MiB) Viewed 1625 times

Coupling flange (mating with the front muffler downstream) badly rusted:

Coupling.png (5.43 MiB) Viewed 1625 times

Leak/failure behind catalytic canister:

Leak.png (3.14 MiB) Viewed 1625 times

Overall view of assembly after cutting off the flex pipes:

Assembly.png (2.87 MiB) Viewed 1625 times

The canisters holding the catalyst grid are a much heavier gauge of steel and were in good shape, as were the catalyst grids themselves.

View of the rear catalyst grid (removed from car) through borescope:

Catalyst.png (2.98 MiB) Viewed 1625 times

View of front catalyst grid (ie the catalytic converters closest to the engine):

Catalyst2.png (2.96 MiB) Viewed 1625 times

Coupling flanges at front of assembly (mating with front catalytic converter):

Flanges.png (2.56 MiB) Viewed 1589 times

I was not in a position to shell out for a new converter assembly, so I made a "temporary" repair. This involved cutting out the old flex pipes, welding in a new pair (I reused the coupling flanges for the upstream converters), and some further welding to patch the small leaks downstream.

The coupling flanges to the downstream muffler pipes were too rusted to remove (see above), so I had to cut them out and join the pipes with a pair of clamp straps.

This "temporary" repair has held up until recently, but it's time to revisit this for reasons I'll relate in a separate post.

[proton17]

Fast forward to 2022, and my repairs are starting to leak again. I'm also getting the following codes related to my rear O2 sensor in Bank 1:

ECM-P013700 02 Circuit Low Voltage (Bank 1, Sensor 2)
ECM-P227000 02 Sensor Signal Stuck Lean - Bank 1, Sensor 2

I used VIDA to plot the O2 sensor voltages, and it looks like that sensor is shot:

Sensors.png (107.52 KiB) Viewed 1619 times

I am hoping the catalytic elements are still ok - I took some thermal images of the front and rear canisters, with the car warmed up to operating temperature, and they all looked to be around the same temperature. The rear canisters are shown here:

FLIR1691.jpg (269.1 KiB) Viewed 1619 times

I don't know what the "correct" temperature would be, but the fact that all four canisters are similar seems encouraging.

I am planning to remove the rear converter assembly again, to check if the catalysts are still good. If they are, I'll try and do a better job of replacing the plumbing. I did the 2019 repair with the car on the street and snow on the ground - I now have a much better setup for working on the car in my driveway (doing this in June helps too!) I will also replace the faulty O2 sensor (I should make sure it was properly connected first!)

If anyone has suggestions on what to look for to see if the the catalytic elements are still good, or gentle cleaning I might do on the catalyst while the unit is opened up, I'd appreciate it.

[proton17]

I spent the afternoon pulling out the converter assembly. Support bracket had rusted out again, and the thinner sections of exhaust pipe were failing or on the verge of failure:

IMG_1763.jpeg (654.39 KiB) Viewed 1576 times

I managed to get both O2 sensors out (penetrating oil and impact gun):

IMG_1765.jpeg (376.94 KiB) Viewed 1576 times

I was careful not to keep the penetrating oil away from the little air gap of the sensors, although Denso claims they use a filter to protect against contamination in the sampled atmospheric air.

The bottom sensor in the photo is the Bank 1 sensor that is showing zero volts and throwing a code in VIDA. Both sensors are quite sooty on the inside - I was going to replace the Bank 1 sensor but maybe should do both. The info I've found on the web seems generally to discourage attempts to clean the sensors.

My tentative plan is to cut out the catalytic canisters and clean up the exterior metal as much as possible. I'll replace all the other pipes and O2 sensor bungs with new metal.

[proton17]

The front catalytic elements still seem fairly clean:

IMG_1745.jpeg (829 KiB) Viewed 1573 times

Although I'm not completely sure how a degraded converter would present.

[proton17]

Here is a closer view of one of the O2 sensors, showing the part number:

IMG_1768.jpeg (433.27 KiB) Viewed 1554 times

Apparently the gap between the corrugated upper cladding and the lower cylinder is where atmospheric O2 is allowed to enter.

Denso has some good information on troubleshooting O2 sensor codes..

In my case, the rear sensor for bank 1 is basically pegged at zero, with some quantization noise in the least significant bit. This does not look like something upstream causing a lean mixture, but rather a failed sensor.

[proton17]

I did some bench-top testing of the two O2 sensors, and it looks like the heater isn't working on the rear bank 1 sensor (the one flagged in VIDA).

Here's a thermal image of the "good" sensor, with the heater leads hooked up to a power supply at around 12V:

IMG_1778.JPG (159.1 KiB) Viewed 1536 times

After a couple of minutes the interior heats up to well over 200C, and I'm sure the embedded sensor is even hotter. Under these conditions the sensor draws around 700mA.

The "bad" sensor does not draw any current through the heater and of course it does not heat up at all:

IMG_1779.JPG (199.02 KiB) Viewed 1536 times

Although it's not really necessary to show the thermal image for the bad sensor, there it is. The small hot spot under the sensor is residual heat from the "good" one. The "cool" image is larger due to use of a different temperature range on the camera (the higher range combines pixels).

I've ordered a local aftermarket sensor, which is $190CAD with tax. From the local Volvo dealer it's over $300CAD + taxes.

With the heater on in the good sensor, I was even able to get the output voltage to go up by screwing it back into the exhaust pipe and flushing the inside with argon. I don't know the voltage specification is but it seems to be working (and VIDA has been happy with it).

[proton17]

I pulled the whole exhaust out for inspection and cleaning:

IMG_1786.jpeg (596.25 KiB) Viewed 1522 times

These remaining sections are a lot easier to remove than the converter, although I broke a screw off removing the middle support bracket (which was also very rusty):

IMG_1784.jpeg (575.61 KiB) Viewed 1522 times

I will try to replace the support bracket, or make another if it's ridiculously expensive. The mufflers don't look too bad. The long pipe between mufflers has a lot of exterior corrosion but feels a lot more solid than the converter pipes (which are exposed to a lot more heat).

Another advantage to doing this on a hot June day, over a chilly November day, is that the exhaust studs pop right out of the rubber hangers (these are a real bear when cold!)

I really need to find the mythical "auto parts junkyard" I'm always reading about. If anyone knows of a self-serve yard around Montreal or Ottawa, I'd love to hear about it!

[proton17]

If anyone wonders what a 3.2 XC90 sounds like with no muffler, here it is:

[proton17]

So I finally finished all the catalytic converter refreshes this car needed, and posted a brief "cinematic" summary video on YouTube.

Hopefully this will be helpful as this seems to be a common failure point in this vintage of car!

[kallekula]

Nice job! Those aftermarket O2 sensors might not be the best buy. Some cars are very picky about it being the OE sensor. I’m not saying they won’t work but they could be a problem.