Found this excellent article and thought I'd share it with you'll.

Real-World Repairs for Real-World NOX Failures

Noel

A couple years ago a fellow list member had one heck of a time making NOx with a B23F - a non-egr vehicle in a warm climate. Though we had tried suggesting several diagnosis steps centered around the tendency for the LH2.0 AMM to degrade in the lean mode, and service bulletin (check valve) inspired Chrysler ignition timing lean-out, the results of various parts changing led to the conclusion the block passages were too full of rust to permit proper cylinder wall cooling - a good reason to dump the motor. Hard to swallow then; supported by Phil's tests now.

I especially liked the repair effected by tightening the exhaust connection. It is hard to picture a leak drawing in fresh air without doing the experiment.
--
Art Benstein near Baltimore Crrrrazy Ray's!

Art;

I think Phil is saying the opposite:

There was lots of speculation about plugged radiators and bad fan clutches or electric cooling fans and their effects on NOX. Perhaps someone will prove an exception to the rule, but here is one tech that has never achieved a NOX reduction by performing cooling system work.

Regarding intro of air into exhaust, I also find that hard to believe. On a 4 cyl, for ex, wouldn't there always be pressurized exhaust in the pipe?


--
Norm Cook Vancouver BC; 1989 745T 202,000KM

Norm, I apologize for my poor and confusing sentence structure, but I agree with you (and Phil) and did so at the time.
--
Art Benstein near Baltimore Crrrrazy Ray's!

Volvo's pulse-air injection works on the fact that there is a vacuum in the exhaust system. Pressure pulses followed by low pressure pulses. If Volvo uses them, then a hole will pull in air, in between pushing exhaust out. O2 sensors also see oxygen from bad manifold gaskets.

It is kinda intuitive there are pulses created by the exhaust stroke, but that same intuition (as I perceived it) has the net pressure at one atmosphere or above.

The idea that the pressure in the exhaust system falls below atmospheric is maybe not so easily taken in my simplified view of the process. I'd be grateful for someone's sensible explanation.
--
Art Benstein near Baltimore Crrrrazy Ray's!

I won't say that it is the case with the case with this exhaust system, but have you heard of Bernoulli's Law? To sum it up, "For a non-viscous, incompressible fluid in steady flow, the sum of pressure, potential and kinetic energies per unit volume is constant at any point." The short version means, as the velocity of the flow increases, the pressure drops. That is how an airplane wing develops lift.

The same thing happens in the flow of a gas through a tube. Where the tube becomes narrower, the pressure drops. If the rest of the exhaust sytem (past the leak) has little back pressure, you could --
john

John, you most eloquently made a point for me. As anyone who has not as a pilot taken an aircraft to flight would instinctively know, a couple tons of airplane could not possibly remain in the air. The gap between instinct and science must be immediately closed on your first take-off with you at the controls, regardless of how many hours you have flying your hand out the car window.

More down-to-earth(!) it still seems a bit odd people will evacuate their AC systems by connecting a passive device to an air compressor. Or modulate the EGR valve in my early k-jet according to the pressure differential between two points along the megaphone shaped intake duct. It is one of those things you (I) have to see and feel to become completely real. The donut gasket story is an example of that.

Have you ever connected a manometer to the test port on the downpipe? I think that is done routinely as a check for clogged cats, but I'd like to try it to get a baseline. I think the port is an M12-1.5.
--
Art Benstein near Baltimore

I hooked up a pressure gauge once when I was checking for a bad cat and a very peculiar thing happened around 2100 rpm. At that speed the gauge needle vibrated like crazy above and below zero. Maybe something to do with standing waves at that point, or the resonant frequency of the exhaust, I don't know. But I do know if I kept it there for very long the gauge would have self-destructed.

The exhaust 'vacuum' phenonema is what headers rely on to help scavenge the cylinders and produce more power. Imagine the exhaust 'pulse' as a ball bearing, and the exhaust as just a straight tube. You open a gate (the exhaust valve) and shoot the ball out. Shortly afterwards you close the gate, and the ball continues travelling under its own momentum. As the ball travels it creates a pressure wave in front, and also a vacuum behind it. If everything goes as planned, you will have a nice fat vacuum waiting at the exhaust port to help pull the next ball that shoots out.

As a side note, I read a long time ago something about tuning the length of an open exhaust system for maximum power. It had to do with cutting the pipe just before or just after to point where the pipe turned blue. If any of you tuners out there know the finer details please let me know.

The ball bearing analogy is pretty clear. I can picture a bolus of hot gases acting as a solid with inertia long enough to move a few inches under its own power, but then I have a hard time seeing the variation of the analogy that would predict a vacuum left behind in a cannon barrel or that of an automatic weapon. Throwing a concentration of gas down a pipe seems like a mighty feat with it leaving a vacuum behind to pull it back. Somehow I just picture the expansion slowing as the valve closes. But no doubt the pressure drops when it does.

The concept of standing waves in a tuned cavity at resonance explains the amplitude (pressure variation) peaks in a way I like, meaning it fits the electrical and acoustical behavior I'm aware of. But there are a lot more variables with the speed of hot gases, cam profiles and header pipe diameters to where I'd think each RPM would have its own set of peculiar combinations.

Do you recall reading about an electronic muffler project using a theory where an opposite phase acoustic wave canceled the exhaust note? I vaguely remember one of the problems or patentable solutions being able to measure and predict the exhaust behavior throughout a wide range of operation. Not something Harley riders would go for.

Yes, vacuum in an exhaust pipe reminds me of that old Bob and Ray Mr. Science skit on gas refrigeration. My belly jiggles just to remember the sound of it. warning 2MB wav file

--
Art Benstein near Baltimore

Well, I'll be the first to admit I don't know all the theory behind exhaust systems, nor would I pretend to even understand it all. No doubt there are countless volumes dedicated to the subject. So much to know, so little time.

As you mentioned the idea of a vacuum in a cannon barrel is not easy to see, unless you picture a very long cannon barrel. At some point the ball is no longer being pushed but instead travelling under its own momentum, and as long as the ball is still within the barrel it will leave a negative pressure behind it.

Continuing this analogy to its absurd conclusion, at some point the cannon ball would come to a stop and then reverse(!). For a tuned exhaust, my guess is the point where the "exhaust ball" stops is where the pipe should be cut for maximum performance.

Something else that wasn't touched on was the effect of the cooling of the gases and how they contract as they travel down the pipe.

The electronic muffler sounds like a variation of the anti-noise generator or active noise control. I wouldn't scoff at the electronic muffler just yet. I'm sure someone somewhere is working on a DSP for that at this very moment, probably the guy with a Harley rider for a neighbor. We'll just have to wait a bit longer.

Besides, who knew we would be using scan tools and DSOs to diagnose our cars today?

Although I don't owne one of these vacuum pumps, one of my friends has used one for years evacuating auto AC systems. No moving parts, and it runs on shop air.

I see those in the HF catalog and wonder. Lots more affordable than the Robinair pump I got in dutch with my wife for buying 10 years ago.
--
Art Benstein near Baltimore

I have on of those pumps and yes it works. In case anyone is interested and wants to do some research these worked based on the Venturi Effect.
--
1989 Volvo 244 DL 218k Miles, the proper color brick red.

it was a good read.
Some Volvo 240s have a 4.0 CO top end specification, which I suspect fails with distinction in any jurisdiction.

--
Stef -- 1981 wagon B21A SU M46 317000km

A lot of technicians replace 02 sensors thinking that they fixed the problem since the vehicle passes smog. What happens a lot of times is that the weak (I say weak because an 02 sensor is really a battery)can ruin the cat because of a rich condition that burns the cat internally and can plug it overtime. It is a good idea on 240's to replace yout 02 sensor from time to time since you can wire in a 1 wire Bosch inexpensively. Don't forget about decarbonization and injection cleaning. One 240 I heard about went to a few shops and still didn't pass smog for CO failure. The several shops installed a new AMM, 02 sensor and Cat. 240's clog with carbon and after this aforementioned 240 was decarbonized and the injectors were cleaned it passed the emissions test CO portion easily.

4% CO is high.

I notice the "good reading for your vehicle" spec on my '89 745 is 24ppm NOx (can't recall what CO was). This was from local emissions test, done about 7yrs ago so they might be cutting me more slack since it's older now.
--
Norm Cook Vancouver BC; 1989 745T 202,000KM

That was an excellent article. I read it all and sent the author an email discussing a few of his points.

Interesting to note how the ailing O2 sensor (example 2; Honda) caused NOx emissions failures. I would've liked the know more of the mechanism behind that.

--
Norm Cook Vancouver BC; 1989 745T 202,000KM

Well, he did say it only fixed it "just enough" to pass the e-test. What I suspect was happening there is the O2 sensor never (or infrequently) dropped below the ECU's lean decision point. Since the ECU kept thinking the mixture was rich, it continued leaning it out.

The ideal rich/lean decision point (or threshold) is 500mv. Let's say for argument that a certain ECU needs to see 400mv or below to detect lean, 600mV or above for rich. If a O2 sensor is at 800mV, the ECU detects rich, and will not detect a lean condition until the sensor has dropped below 400mV. (And similar holds true for the lean-to-rich direction.)

A certain amount of hysteresis is necessary to overcome electrical noise and prevent false sensing. Imagine if a sensor were, for some reason, hovering around 500mV. Throw in a bit of noise from the vehicle's hostile electrical system and you now have 495mV-505mV, crossing 500mV hundreds or thousands of times (or more) per second. If the rich/lean threshold was at exactly 500mV, the ECU would go nuts trying to figure out what the truth really is. So by adding a small 'deadband' the ECU can safely ignore most transients without the need for any elaborate signal processing or filtering.

The price that is paid is a slight delay in response, but since the rate of change during the actual switching across decision points is fairly quick (say, from 600mV to 400mV in my example above), that delay is minimal.

This is speculation on my part of course, but quite plausible in my opinion. But since I haven't taken the time to prove my theory and I might just be blowing a lot of smoke, take it with a grain of salt.

Your speculation is right on if we happened to be talking about Volvo's LH2.0 stuff. The front end is made of two comparators (Bosch chip has same pinout as an LM339 quad comparator) where the lower of the two is set at 0.49V, the common input is biased with 5 microamps sourced from a precision divider set at 0.52V, and the upper comparator set for .58V on the way up. Hysteresis from its output reduces its setting by 10mV on the way down (~500K resistor) to make it clean.

The sensor never really spends any time in the .49-.58 range even when "lazy" or slow rise time. But how quickly the sensor can react to the engine's correction makes a big difference in overall mixture. Here's a correlation between a lazy sensor's output and the CO test point on my LH2.0 car.

Great pic, Art. It puts it all into perspective.

Don't forget to change that sensor ;)

Cheers.