Failing emissions: high CO but the ACV checked out
 

So I'm failing emissions with high carbon monoxide, and it keeps coming up as being air control valve related. I was able to get an emissions exemption from the state 2 years ago and they thought it was the ACV as well, so that winter I tore into the engine, pulled out the ACV and bench tested the unit. All the valves and solenoids on the ACV checked out, and I also retested the secondary air switching and bypass solenoids, and they all operate within spec, triple checked my vacuum hose diagram, and put it back together.

I can see the clutch engaging on the airpump, but the only thing I haven't been able to verify is that the airpump is putting out pressure. I've asked my mechanic to check that first before disassembling the intake system. Is there anything else I could have missed?

Could your cat be getting clogged? Do you pre-mix? There could also be a small leak. What about the bypass air tube coming off your cat?

* Clogged cat - The main cat was replaced two years ago, and the emissions numbers are near identical now as they were then.
* pre-mix - I run a very small amount of pre-mix (~4oz per tank), but that was something I started doing last year, and the numbers haven't changed for better or worse. Good point though, I will top it off to dilute as much of the oil as possible.
* air leak - After I got home I did find one of my silcone lines had split, so i'll replace and retest.
* split air - it's hooked up.

First of all, this is an IM240 test. It's stricter than constant mph tests, although it is still done on a warm engine.
https://www.rx7club.com/attachments/...t-im240dds.gif

Since it is passing NOX and HC, you should get a new O2 sensor. They degrade to the rich or lean side over the years. An OBD I car isn't smart enough to detect a bad one. It's probably reading leaner than the mixture really is, so you get CO emission breaking through the catalyst. On an OBD 2 car the front and rear sensors cross che like each other and diagnose the cat as well.

Make sure they thoroughly warm up the engine before the test.

Ok, let me continue with this line of thinking. You've checked the ACV many times, so I'm going to assume it's ok. You haven't replaced the o2 sensor or you replaced it and forgot to mention it.

Let's review each emission and the factors that contribute to it.

HC emission: HC emission is related to unburned fuel, besides just having a dead cat. It's especially high during cold start where the fuel doesn't vaporize in the intake ports very well. It also occurs during misfire from unstable combustion. This test is a hot test, so assuming the engine was warmed up correctly, the wall wetting is minimized (at least for purposes of this test) and you don't have to worry about HC from bad atomization.

Misfire from unstable combustion can be caused by bad tip in fuel, like a tip in hesitation. That's TPS related, or a bad tip in tuning if it's not a stock ECU. A pre-Rx-8 rotary needs the air pump to stabilize the combustion and reduce misfire at idle and low load because the port timing creates a strong overlap in addition to the gas leakage through the older sealing system. So if you don't have a working air pump you're going to have higher HC.

NOx emission: NOx emission is caused by high combustion temperatures and lean mixtures, or just a cold/dead cat. It's usually highest when oxygen (due to lean mixture) passes through the cat without being reduced into less harmful compounds. It can also be produced at medium loads, when the throttle is open enough to accelerate but not enough for the mixture to richen up. That's what the 150 to 230 second portion of the IM 240 test is supposed to test. It's at a higher vehicle speed than a typical steady mph test in order to load the engine down into the high NOx producing operation area.

CO emission: the number 1 contributor to high CO is rich mixture. For an Rx-7, it's either a non functioning air pump or a bad O2 sensor, OR on this type of emission test the driver being very aggressive on the tip in and the ECU starts to richen up due to boost. Typically a non functioning air pump will also result in increased misfire, so you get high HC with your high CO, but low NOx because the mixture is rich. In your case, the HC is higher than it should be, but it still passes. On a steady mph test the driver doesn't matter, but on a transient test like this see if the driver will be kind enough to drive easy on the acceleration portion of the test. See if he will be easy on the gas pedal and lag behind the target mph trace a little bit. The EPA doesn't like that on official tests but this is just some guy at an inspection station driving your car.

this chart from the 2nd generation FAQ can help guide you:

https://www.rx7club.com/attachments/...onsfailure.jpg

TL;DR replace the o2 sensor and ask the driver to accelerate as gradually as he can get away with so it doesn't boost much and therefore doesn't richen up.

i haven't gotten an Rx7 to fail with high CO, so congrats! um the bad o2 sensors i've seen have so far all had high NOX, but i agree with Arghx, try an O2 sensor.

Denso 234-1040

does that Denso part number plug right in?

all your tests are idle and steady mph aren't they? In that case there's no risk of the driver giving it too much throttle. Experienced emission certification technicians can drive the mph trace on a transient cycle almost perfectly. The EPA actually has a formula to calculate how aggressive the driver was on the test because it is so important for CO emissions on acceleration. Expensive emission labs can give you a datalog showing the emission on a second by second basis. That's how you find out where in the test you failed.

Good to know!

If you are looking for an exact replacement O2 sensor, consider this website: Oxygen Sensor Depot
They have the sensor with OEM connector at reasonable prices. I also understand that Rock Auto has it available.
A recent thread of mine that discussed O2 sensor replacement: https://www.rx7club.com/3rd-generati...ement-1104560/

Side note: what the hell did IB do to the reply box? Damn it, why fix something that wasn't broken!

Arghx is correct, Colorado uses a IM240 rolling test. Also, I did forget to mention, when I had the cat replaced two years ago, the O2 sensor was also replaced, and the spark plugs and air filter were replaced this spring. Now that doesn't mean the O2 sensor was put in correctly, and my shop is checking that out too. They did buy a "universal" unit, so I don't know if the stock connector is still in there, I'll check that out as well.

The tech who drove the car during the test was a little too aggressive with the accelerator and was repeatedly out pacing the dyno, and I do try and go during low wait times so the engine is good and warm when the test begins, but that's not always in my control.

If it helps, I've included the measurements taken during the test.

EDIT: As I'm looking at these graphs more, as Arghx mentions, the HC seem to spike under deceleration, like the fuel injectors are still firing (stock ECU). Is that normal, or also a byproduct of a misbehaving O2 sensor?

yep, the universal one is a different number (234-1000).

in CA we have three different tests depending on where you live, either none, the IM240 test, or third the stationary idle and stationary 2500rpm tests.

by far the hardest one to pass is the third one. the idle test is really really hard, stock FDs won't pass this pretty much no matter what.

the IM240 test is really easy, its a high enough rpm ~1800, where its in closed loop, but its still in port air so its right were all the emissions equipment is working.

just as a note the split air pipe isn't used during the test, and it can be removed, its kind of a red herring anyways.

wow thats cool, we don't get anything helpful like that here....

so far O2 sensors do two things, they get lazy, and/or just completely fail.

my friends FC had a failed one, and it was too lean, NOX was really high, everything else was low
my miata had a lazy one, and it was also lean, NOX was too high, but just by a bit.

i haven't seen one that failed in a way where the car was too rich, but i'm sure its been done

also it could be a bad/failing water temp sensor, it has a rather large impact on the mixture, and i have seen it where its broken and alternates between, fully rich, fully lean, and reading correctly.

if its open or short it will throw a code, but if its just slow/lazy the ecu may not figure it out.

Just throwing this out there, but along the lines of the ECT failing if the coolant temp is too low due to an aftermarket or stuck open thermostat, the OE Mazda ECM may keep the engine in a partial cold start enrichment mode causing elevated CO. Not sure if this applies to a hot running car like the FD, but I found this exact thing on a customers car we had procured for IUVP testing - ran the car on the FTP test and it failed miserably for CO. Did some checking and found that the owner had installed a 160° thermostat and the car never fully exited cold start enrichment.

You said that you replaced the catalytic converter and the O2 sensor two years ago and you got a two year test exemption. So, two years ago the car did not pass with a new cat and new O2 sensor?

Stock cat?

I went through YEARS of this, gave up, but came back and passed in Colorado. With a Bonez high flow!

I "heard" a bit of denatured alcohol in the mix doesn't hurt.

if you go new O2, you might consider wideband with both 0-1 and 0-5 v outputs, you can kill 2 birds that way... good luck. It sucks.

Keep in mind that there is a delay in the second by second graphs due to the time it takes for the gases to travel.

I think a lot of this is coming down to the driver. How does the gear shifting work on this test? Does the driver follow a shift schedule or do they decide on their own when to shift?

see if you can get a different driver or go to a different inspection station.

Has the air pump been ruled out? And do you have a down pipe. Usually you can pass w/o the pre-cat, but not sure in CO

The fuel purge system sometimes messes with emissions also

I would start with a stock 02 sensor then try a different station

OP can you give us a list of every engine mod you have? I have been assuming it's stock everything except a downpipe. Do you have a stock ecu, stock boost control, all that?

Ok, a lot of things to reply to here, I'll try and catch them all.

j9fd3s & jza80 - water temp sensor is new, but I decided to go with a lower operating one. I've also been dealing with an overheating issue. The stock temp gauge never hits the upper white hash mark, but it is rising from it's 9-o'clock position under spirited driving. A pressure test revealed a small pinhole leak, which is getting fixed as well. I just don't know if a pinhole leak would be enough pressure lost to effect the cooling system that much or if it's the fact that the Silla radiator I have just doesn't have the cooling capacity. Anyway, I don't want to sidetrack the thread from the emissions failure, unless it's a symptom.

mdp - correct, the car didn't pass two years ago either, the numbers where almost identical to two years ago

johnchabin - yeah, I've heard those things about alcohol too. I'd rather not have to go that route, but we'll see. I'd love to go wideband right away, but I need to read up more about it.

arghx - hard to tell, but he seemed to follow the prompts to shift pretty accurately. With the 3" exhaust, I have been boosting to about 12-8-12 psi

TomU - haven't heard back about the airpump yet. I clamp off the purge lines. I did try a different station last time and readings were higher.

My car is pretty stock, I've mostly focused on reliability mods:
* 3" exhaust with both pre-cat and main cat (long story short, the dealership I bought the car from put a pre-cat in to try and help with emissions)
* Rotary Performance SMIC
* 180 degree thermostat and 185 degree fan switch from Pettit
* aluminum AST

So the only thing not stock is the SMIC, AST, T-stat, fan switch, and O2 sensor

It should pass (the prior fail rules out a testing anomaly).

Since you've ruled out the control system, my bet is on the O2 sensor and/or the air pump

Oh. There's your problem. turn the boost down! try turning off your boost controller so it goes back to base pressure at the precontrol actuator. Or go back to stock exhaust and stock boost control system. The stock boost control and stock fuel maps are specifically tuned together for passing emissions tests.

when you add more boost you add more CO, but only during acceleration. It won't ever show up in these constant speed tests that most cars go through, because the engine isn't under enough load in those cases.

also, these transient emissions dynos simulate road load more accurately than a dynojet. The road load calculation, at least in certification tests, is for worst case vehicle test weight. That loads the engine down. So if you have a base model it will actually load the engine down as if you had the slight extra weight of a touring model.

The guys at County Line & Colorado were RX-7 fans and kept it out of boost. It helped.

Guess that’s a bit of a haul for you.

One other general tip for this type of test is to minimize rolling resistance. Make sure your rear tires are at least at the recommended pressure. It never hurts to pump them up to the maximum psi rated on the side wall for the test. It can make the difference if the car is right on the edge of passing under load.

Forgive my ignorance on boost controlling with an essentially stock setup, I'm currently reading up on it from Robinette's page, it sounds like for me boost is limited based on the restrictor pills in the wastegate control and turbo precontrol lines. When I did the vacuum lines I did reinstall the pills, but they did not fit as snugly as the stock rubber lines. Could they be rotating in the line, effectively limiting or stopping the actuator from doing it's job? My first thought is a couple of zip ties could help hold them in place, sound reasonable? Or if I decide to go with MBCs, what are some decent, reasonably priced options?

Thanks, I'll check them out, I heard Precision Imports at W. Hampton and Federal is supposed to be good too.

EDIT: oh you mean the air care station, might be worth a shot. It's definitely out of my way but if I keep having trouble with the golden station I might make the trip.

Can you name every mod you have that affects boost? Intake system, exhaust system, boost control, vacuum lines, ECU. The restricter pills raise the boost by limiting the amount of air that can vent from the wastegate actuator. More restricter pill restriction = higher boost, no restricter pills = lower boost. I doubt the engine is getting into secondary turbo operation on the test cycle btw. I do this for a living, so let's look at the datalog they supplied you with:

https://www.rx7club.com/attachment.p...1&d=1506359406

The little spiky ones aren't a big deal, so you want to look at all the points that look like a square in the pattern . I have circled them in red. In those you are clipping the range of the CO analyzer (or at least maxing the range on the chart) which is why it's a square: you're blowing tons of CO through there. At about 155 seconds there is an acceleration which is probably in 3rd or 4th gear, loading the engine down. This is showing up at 160 seconds due to the delay in the analyzer. It looks like the 160 second one might be what made you fail the test more than anything else.

Here is the target AFR "zone" map from the service highlights document:

https://www.rx7club.com/attachment.p...1&d=1506360664

And here is the Secondary Air (air pump) "zone" map:

https://www.rx7club.com/attachment.p...1&d=1506360664

On a steady mph test and in portions of this IM240 test (like the 50-80 second mark), you are going to mostly be in the above "Feedback zone." This is where the condition of the o2 sensor is important for CO and NOx. If it's biasing one way or the other you will see CO emission break through the cat. Look back at the 60 second mark from your datalog. Do you see significant NOx or CO spikes? No, you don't. This is a good indication that the O2 sensor is ok, and possibly the secondary air. Every major acceleration though you are blowing CO through the cat. This is because the engine is going rich, and the catalyst runs out of oxygen to convert it.

Any mod that changes the throttle angle vs boost relationship to make it more responsive or more torquey is going to increase transient CO emissions. That includes anything that affects boost. The reason for this is that aftermarket ECUs and non modern stock ECUs just have dumb look up tables of speed & boost (or some other load calculation) vs target AFR. Instead of a simple "zone" map like the FD has, on most modern ECUs they calculate the temperature of the catalytic converter and delay going rich in order to drop the CO emission and improve the fuel economy. On a stricter emission test (for a modern light duty vehicle) you pretty much can only have one or two CO spike, in a much longer test cycle, or it will fail. Here all we need to do is reduce your worst CO spikes and you'll pass.

Using the stock air box, a 3" exhaust with pre and main cats (as I mentioned above it's a bit of a long story but the dealership I bought it from installed a precat to help with emissions at the time), stock ECU, stock boost control (I transferred the restrictor pills from the old rubber lines to the silicone lines), silicone vacuum lines.

Agreed, watching the guy drive it, it definitely didn't sound like he got into the upper RPM. But he was out accelerating the program which I plan on pointing out next time I'm over there.

Understood, and agreed.

I need to fix the over-boosting issue first and foremost. But I'm not there yet, as I don't believe the tech was hitting WOT and thus full boost.

https://www.rx7club.com/attachment.p...1&d=1506423357

so only the catback (A and B) is aftermarket? What brand? Do you have a stock catback? put a regular paper air filter in if you have a K&N drop in. More restriction will make it spool slower. You just want to reduce any spiking on spool up. Most importantly though you should talk to the tech and ask him to drive it easy, as easy as he can get away with without getting flagged. Maybe you can make it worth his while.

quickest thing you can try is remove the restricter pills to drop the boost, then take it back to the station and ask them to drive it easy. I would do that before messing with the exhaust. A throttle, even an old cable throttle, is a non linear thing. Once you get to about 30 percent pedal angle you are almost WOT.

the test is usually done in 2nd for the 15mph part, and 3rd for the 25mph part, most cars run ~1800-2200rpm, so its right in that sweet spot where its in Feedback AND Port Air.

that being said, in a stock twin car a big throttle application will see 1-2psi of boost, which is enough to take it out of closed loop. there is no way its seeing more boost than that, the test is only 25mph, and the car is really fast.

maybe a more gentle tester is in order?

Yes, for the steady mph tests, boost isn't a problem. For OP's test, it gets up to ~55mph, and there's at least one significant acceleration in what looks like 3rd or 4th gear. I don't know how much boost he's getting into, but it's definitely dropping out of closed loop, and the CO goes off the charts for a few seconds, and then he fails.

https://www.rx7club.com/attachment.p...1&d=1506423065

Yes, I'm talking about the air care station. I've had the car for almost 20 years, and I've never found a good rotary shop. I do everything myself.

arghx - man you know a lot about emissions! All I did was made sure all my emissions components were functioning, adjusted the TPS, got it hot, and brought it in. I honestly didn't think it would pass.

That said, the only big adjustment I ended up making was the TPS... it was pretty far out of spec. Have you measured the voltage output of the TPS on a warm engine? Mine was easy to check with the Power FC. I don't know if it has any effect on emissions... maybe arghx can chime in?

TPS should be checked for sure. If it's not right it could affect air pump operation (narrow range TPS signal) and deceleration fuel cut (narrow range TPS signal). It can also affect boost control (full range TPS signal), which can affect CO emission.

There is some confusion by many folks here on the type of testing being conducted in Colorado. In actuality it has become the wild west in testing procedures since the EPA has given the states leeway in choosing the appropriate tests in locations that have failed to meet clean air standards. Here is an article explaining the history of regular emissions testing and the evolution of the IM240 test from a emissions certification test conducted by vehicle manufacturers under laboratory conditions, to a ubiquitous tail pipe tests conducted thousands a times a day at hundreds of testing stations.

I can chime in on that. From a policy perspective, it's clearly a mess, with all the differing requirements. I'm not sure - do they require an IM240 test if you have an OBD II car, or do they just plug in and check for codes/emission monitor readiness? Maybe that varies by the locality. OBD II has been out for over 20 years now so how many old cars are you really going to test? Then you've got all those repair costs which tend to hit lower income people, and a minimum amount of money you have to pay before they give you financial relief--a sacrifice to the low emissions gods you could say.

From a technical perspective, the IM240 is similar to certification tests (like the US FTP75, US06, and SC03 tests, the Euro NEDC or WLTP tests, or the Japanese JC08) but is actually much much easier. It's a short test, it doesn't require the engine to sit overnight and cold start (either at room temperature or in a cold chamber), it doesn't require you to shut the engine down and heatsoak for 10 minutes. The test limits are not that difficult to pass for any 100% stock car with all the equipment working. It doesn't run at very high speeds and doesn't accelerate that aggressively. An IM240 test has a better chance of detecting a tuned car that has all the emissions check engine lights disabled in the software. It also has a better chance of catching an engine that is wearing out and burning a lot of oil. So if all you cared about were emissions, really every car should run an IM240. Now I'm not advocating that as a public policy, I'm just pointing out that from a technical perspective you are going to catch a lot more polluting cars with a transient test. People have to decide if catching those extra number of cars is worth the cost and the hassle.

A roller dyno test that includes a cold start isn't feasible for a high volume of cars, so a short hot test like an IM240 makes sense. However, on modern cars CARB requires a "cold start strategy diagnostic monitor." This calculates whether the emissions controls used during cold start are working correctly. It will make sure that the spark and injection are functioning in such a way as to heat up the converter, by doing things like calculating the heat transfer from the exhaust into the cat. Then it will throw a code if that's not working. The whole idea is to make the stock ECU's so smart that you don't need to do any dyno tests to detect emissions. As I mentioned though, modern cars that allow the stock ECU to be reflashed will have the capability of disabling all that emission control stuff and tricking the inspection computer into thinking everything is fine. That's where visual inspections and dyno tests fit in, again from the perspective of maximizing the number of polluting vehicles caught rather than keeping money and hassle minimized.

from the perspective of CO emissions, OP's IM240 test is nowhere near as challenging as the EPA US06 test used in certification (which wasn't around when the FD was a new vehicle):

https://www.rx7club.com/attachment.p...1&d=1506443269

Even the current Euro cycle (NEDC cycle, note km/h on Y axis instead of mph) would be tough:

https://www.rx7club.com/attachment.p...1&d=1506443269

This is the test the FD was designed for, US FTP75 test (cold starting the engine after sitting overnight at 75F temperature):

https://www.rx7club.com/attachment.p...1&d=1506443530

That was developed when some guy in the 1960s drove around the LA area and said "ok, this is a typical American commute." The Japanese cycle was very easy, because it is basically the equivalent of stop and go traffic in Tokyo, with no major acceleration required.

Sorry for the confusion, it's a franken-exhaust. The previous owner had a 3" downpipe on it, and the dealer chopped & welded in a pre-cat. I then put in a RB cat-back. Then two years ago, I asked my mechanic to get a 3" main cat so I would have a full 3" exhaust. I should have just got a 3" Bonez DP+main cat.

Good deal, I will make sure to talk with the tech before he drives it.

Yeah, if I was outside of the Denver metro area, I wouldn't have to put up with this crap. I know the guys in Colorado Springs don't have to worry about this.

wow thats crazy, what happens when someones car falls off the dyno?

our test is just to hold 15mph for up to 120 seconds, then hold 25 mph for up to 90 seconds. if the car is really clean or really dirty the test is shorter. ours are also HC as a PPM, FD's are generally allowed about 80, CO is a percent, FD's are allowed about 0, and nox is some random number, we're usually allowed 450 of them, it makes big swings.

Yup. The steady speed tests are a lot easier to run, and don't require as expensive of equipment. Basically you stick a tailpipe sniffer in there and look at the concentration of the emissions at the tail pipe. PPM (parts per million) HC & NOx, and % of CO (so parts per 100 basically) are easier to measure than actual mass of emission. With a transient test like the IM240, you trap the air and measure the mass, which requires equipment that costs a lot more money. You also need technicians who can drive the mph trace, which requires more skill and experience than just going to a speed and holding it.

What happens if cars fall off the dyno? Well, that's what insurance is for. I bet insurance for IM240 testing stations is more expensive. That's why I'm saying that from a purely technical/emissions perspective, IM240 all the things. From a policy standpoint there's a lot more to consider.

So it's got an aftermarket pre cat, an aftermarket 3 inch main cat, and an RB cat back. Aftermarket cats have way less precious metal loading than OEM, which is bad for emissions. You won't see it so much at idle or steady mph (when presumably the O2 sensor and secondary air injection pump are working), but in transients you will. You will also see it on cold starts, which are not a concern here. Cats use cerium for oxygen storage. The oxygen storage helps oxidize CO that breaks through the cat when the gases are richer than stoichiometric (Lambda=1). I don't know the precious metal loadings of the stock cat vs the cats you have, but directionally, two aftermarket cats is probably worse than just having only the stock main cat. There's a reason why stock cats cost so much. The RB catback isn't a big deal, although less restriction helps spool a bit, which means it could be helping you flow more CO when the driver puts the throttle down.

I didn't see where Phu5ion mentioned which brand of cat he replaced the stock cat with. But, I do recall, over the last 20 years, on the big list and this forum, that there were three members with similar problems. We all went round and round trying to figure out what was causing the repeated failure in emissions testing on these stock cars. I believe one or two of them had new aftermarket catalytic converters. As I recall, after eliminating all other possibilities, two people finally coughed up the bucks and installed a new Mazda catalytic converter and the other borrowed a know good OEM converter. All three passed after installation of the OEM converters. I would say that after making sure everything else is working to OEM specifications that you are left with trying an new, or known good, OEM converter if you don't already have one.

That is consistent with my experience here in Colorado. The stock cat is much cleaner. I did pass with the Bonez, but the numbers are higher than with the stocker, so less room for error. The air pump is a necessity. You verified it's connected to the cat and functioning?

I'll get a picture next time I'm over there, the dyno is built into the floor so it won't "fall," and then there are some side rollers which keep the tires from tracking off of the rollers.

I'll ask about where they got the main cat from, but it's most likely some generic supplier. The car's airpump and O2 sensor is finally getting tested today. With any luck I might be able to get the car back, remove the restrictor pills, and test again on Saturday.

Bro, just drive by the mobile emissions testers a few times while coasting and you will get a letter in the mail saying you passed.

I passed!
 

Took a lot longer to get back there than I expected, life happens, but here's the update. I pulled out the restricter pills, pinched off the catch can purge line, tires were already up at 37PSI, and filled the tank without any more premix. When I got there, I talked the manager and explained that some of his techs were out accelerate the program and blowing my CO through the roof, so he drove my car, did a great job and brought my CO from 25GPM down to 9GPM with a 15GPM limit.

Thank you to everyone for helping me out.

As for my overboosting... Well, there's my problem.
Attachment 730223

And a few were interested in the facility so here are a couple of shots at their rolling road.
Attachment 730224
Attachment 730225

Congratulations on passing! Is that a orange silicone hose? It looks similar to the failure mode in this test.

Nice job. Which station did you go to?

It's actually red silicone, ordered it from boostcontrollers.com.

Thanks, I went to the one in Golden.