+1 on the OE sensor, it is an ND part.
the Mazda Part number is N3A1-18-861A-9U in the USA (drop the 9U if you're anywhere else, the Denso with the plug is 234-1020
234-1000 is the Denso without the connector and the NTK is 24019.

the Bosch screws in, but they aren't very good

 

I tested all of the reused sensors back in 2020 when I was originally installing the new short block. Well I should have rechecked them sooner, but the suggested sensor retesting spree has yielded more potential culprits:

1. Water thermosensor on the back of the water pump was out on the low end with 1.89 kOhms @ 20C instead of 2.2-2.7 kOhms, and also on the high end with .38 kOhms @ 80C instead of .29-.35 kOhms. Since operating temp is ~82C, I suspect this reading colder than expected on the high end could be a potential contributor to the richness across the board.
2. The bigger surprise was that the MAP was also out on all 3 of the FSM tests, though I'm not sure exactly how much I trust the Mityvac's gauge. The -19.7inHg test was way off at .969v instead of 1.25-1.55v, and the 98kpa pressure test had the sensor at 4.00v (instead of 4.35v). The PO had removed the MAP's filter canister, so I suspect this may have affected the sensor. That and almost 200k miles.
3. I didn't bother to pull the AIT, I just ordered another.
4. I also added a denso O2 to the order as well.
5. Fuel thermosensor is new OEM, but if I have to pull the UIM to get at the AIT sensor, I'll pull it and test it as well.
6. TPS checked out still in spec with smooth advance as the pedal was pushed.

One other area that was suggested was checking to see if the evap system/charcoal canister was in some way compromised to allow HCs to get sucked into the testing vacuum placed by the exhaust tip. I had replaced the fuel tank and inspected the lines and separator's plastic shell while I was doing that back in 2020. If the sensor changes don't get me across the line but fix the CO, then I think this is the next place I'm going to look. I don't smell gas from that area, but it's probably still worth a look.

 

Fuel thermosensor passed all of the FSM checks. Based on the sensor testing I did previously, I ordered a new MAP, OEM water temp, and OEM AIT from Mazdatrix and installed. I also put in a new OEM Denso O2. Test #8 had the same warmup prep and stock cat as previous (turbo controls intact) but it managed to fail as bad as before the port air was working. I was worried I may have kinked the ACV switching vacuum line while reinstalling the UIM (or connected the vacuum source incorrectly) like it was in post 27, so I undid the hose, tested it to make sure it had vacuum at idle, and rerouted it. It didn't seem like it was kinked but I may just take it in again on the off change that killed the port air.

I also ended up testing the coils just for peace of mind. Front trailing coil was at .8 ohm, which is getting close to the limit but not quite there yet. The rest were well within spec.

I think this unfortunately brings me to the limit of my knowledge on what could be affecting the mixture sensor or tune-up wise. I guess that leaves mechanical failure in the injection, evap, or catalyst systems as the cause. As always, I welcome and appreciate other thoughts or suggestions.

 

Not that I'm an expert on emissions systems or catalytic converters, but...
In my past readings and experience, excess HC's can cause an overheating cat, but a cat in good condition will normally oxidize excess HC and CO. Oxidizing excess HC and CO is, IIRC, a cat's primary function.

So, IMO, all these failed tests are a clue that the cat you're currently using is no longer functioning as it should. There are numerous causes of a failed cat. IIRC, the primary one is using leaded gas which coats the catalyst and renders it useless. Another is using an oil which has too much anti-wear additives such as ZDDP.

 

Cats just naturally degrade due to heat cycling. When they do emission certification at the OEM level they have to take stock cats and put them on a cat aging bench that basically heat cycles it up to 150k miles. It can be done with a burner type thing or a misfiring engine for example. There's a reason why stock tunes are so rich; limiting heat cycles has a huge impact on the emission conversion down the line.

So in this case it's still possible your cat is just worn out. That being said, if I compare one of your best reports (1st one below) to your most recent one (2nd one below), you've now got higher CO and HC, and lower NOx overall. That tells me you're running rich as a result of whatever work was done recently. It's entirely possible that your airpump or ACV isn't working right for example, for whatever root cause.

The sledgehammer approach to solving this problem is to make sure your airpump and ACV are working and to install low mileage stock pre cat and low mileage stock main cat. I'm sure you can source those off this forum. It'd be a nice science experiment to just install a pre cat on top of the cat you have now for example. Then change the main cat.

 

How much have you spent on repairs? One year I wasn't able to pass emissions after replacing the cat converter and was able to get a "repair exemption" waver for a year.

I found the page. You need two failed tests, one before and one after repairs exceeding $715.

https://dmv.colorado.gov/emissions-waivers

 

Total since I decided to put a 13b back in it? Probably about $10k in just the short block, fuel system, solenoids, and engine electrical. I explored this route a few times, but in typical government fashion, their view on requirements for the repair waiver miss the point. I'd say about $9k of that was spent prior to the car being drivable, much less testable. All of that spending, regardless of the intent to restore the emissions systems, doesn't count since it wasn't between failures. Also, non-Colorado vehicles are not eligible for waivers at all. You must register at least once in CO somewhere (including passing emissions) prior to applying for a waiver. As this was originally a TX car, they told me I couldn't do it regardless of how much I spend.

It doesn't matter though, since I finally got a pass. I only changed two things mechanically with the car. First, I installed another air pump that I had laying around from earlier. This one was damaged on the intake side during shipping so I hadn't installed it back when I was fighting the crumbling connector. Further, since I could feel air coming out of the split air during that troubleshooting and my old one didn't have any other signs of failing (like squealing or the clutch failing) I thought the old one should be good. I suppose the flip side of the 'too much fuel' coin is 'not enough air' which this pump may have provided. I also pinched off the evap vacuum line between the catch can and the TB, on the off chance that my charcoal canister was full of fuel or something. In order to reduce vapor pressure above the fuel in the tank, I stopped for 10gal of fresh cold 91 octane right next door to the emissions place after warming up the car. Lastly, today was also chilly in the 50s, conditions which were also similar to my other near passing test. Upon passing, I did beg for the passing trace, but the program they use doesn't allow it to be printed.


On the topic of testing air pumps, the FSM testing procedure for the air pump was something that is still a bit unclear. Are they corking the outlet and just measuring the pressure? Is there some kind of outlet in the SST that is calibrated? If there was a good test procedure for this then I think I could have saved myself a lot of headache by testing it back when it was discussed near post #14. The clutches on both air pumps worked on and off the car, and I could see both engaged at idle. That leads me to believe that the issue was with pump output, which this FSM procedure should have been able to tell me if it didn't require an unknown SST.

 

0.7 PSI is not a lot of air. Roughly a pound of pressure, given the diameter of the house.
I would hazard to guess that this a fancy way of saying "stick your thumb in the hose and see if air is coming out."
I don't think you would be able to accurately measure air pump output using a pressure gauge like that without plugging it.
Go the hardware / auto parts store and get some fittings to try it.

Wouldn't cold fuel and air actually make the car slightly less clean? Are you certain the cat was actually getting air with the previous pump?

 

I'm certain the previous pump was pushing at least some air. I tested at the pump outlet with my thumb, at the split air (about low-med garden hose pressure there) before the switching vacuum was fixed, and at port air based on the drop in emissions from the 2020 test to test 4+.

 

Garden hose pressure sounds about right. Feel like doing an experiment with the old and current pump? Because if you don't nail the exact cause know, you are likely to have the same problem next time.
If you don't have a pressure gauge, you could zip-tie a balloon on the outlet and see how big it gets...

 

Well... this is indeed the experiment thread. I can probably rig a plug with a port on it and see what pressure each puts out when fully capped.

Regarding the cold air and cold fuel, I did that to see if I could mitigate any non-combustion related HC detection coming from a leaking evap connection or component. I thought that the exhaust heat might be heating a mostly empty tank, raising the vapor pressure, and potentially leaking out of one of the vapor seals or hose connections on top of the tank or near the exhaust tip. Those leaks might get sucked into the testing hose and reported as HC on my test. In filling the tank and cooling everything, I figured I could mitigate this some. I'll be able to rule this out conclusively next weekend with a smoke generator once it gets in.

In a similar vein, I need to also double check the purge solenoid and its vacuum connections as potential contributor to the richness. The solenoid is new OEM, but I never actually hooked it up to a battery to verify it switched, nor verified that I couldn't blow through when closed. If my charcoal canister was fuel soaked or something and the solenoid was stuck open or leaking, then my engine huffing all those vapors could have been a potential contributor. I think this is less likely than the air pump being the cure since my CO was definitely correlated with the acceleration zones rather than being rich at cruise.

 

Come to think of it, if excessive acceleration by the tech is part of the problem, is there a reason you can't just rig the accelerator pedal or the throttle to only open to 50% or so? hehe

 

Congratulations on getting it to finally pass. This will be a really good thread for anyone trying to pass emissions tailpipe tests. I'm glad it was only the airpump in the sense that that is not an expensive fix (used good condition air pump) compared to the cost of a cat. As for testing it, I do know these types of pumps are basically more about flowing air than achieving high pressure.

The cold fuel and colder day was a good call. For those who are fuzzy on evaporative emissions, basically you get HC emission from vapors developing in the gas tank, and hotter fuel means more vapor pressure. Think vapor lock when hot starting an old carb'd engine. The evaporative purge solenoid is supposed to put some of that vapor back into the intake to burn up. So on an actual EPA development certification test, they fill the charcoal canister to a specific spec, the fuel is a specific spec, and the ambient temperature is controlled to 75F/23C. An inspection station doesn't have that kind of controlled condition.

Other things that can affect evaporative emission HC are the fuel blend. Winter fuel blends are designed to vaporize better, so it's easier to start, while summer fuels are supposed to be harder to vaporize for emissions purposes. Considering it's late October now, it's possible they switched the fuel blend for the year. Or just having cold fuel in there was just enough to eliminate that aspect. The air pump was probably the bigger factor however because both the CO and HC are way down vs your best test up to that point.

You went from 1.1 HC and 9.8 CO grams/mile on your previous best test to 0.5 HC and 4.7 CO. You basically cut emissions in half by better oxidation and are way under the test limits to pass.