Are Arizona emissions tests bogus?, test history makes me wonder... Options

[914GT]

I was bored with work and I was curious how the emissions test results compared going back over the years on one of my cars. I kept almost all the inspection results for an Acura Legend that I bought new in 1987. My daughter has been drivng this car the past few years, and I still maintain it. I picked this car as I know its history and I always maintained it well. Basically the 2.5L engine has had little done to it over the years, plug changes a couple times. Replaced and/or cleaned the injectors a couple of times. The usual routine filters, etc. It's never failed emissions. I replaced an O2 sensor once because the engine service light came on.

I don't see any rhyme or reason to the numbers. We'll see how it does next month when it's due. There are some screwy variations over the years that I can't correlate to any particular repair or maintenance item. Makes me suspicious of the whole thing. I'd expect some variation due to the car and testers (all at the same location) but there are some years that are drastically different.

[lapuwali]

Engine, and in particular CATALYST temperatures can make a huge difference to measured HC. If the engine is fully warmed up, the numbers will be lower. If the cat is allowed to cool much before the reading, it will be higher. If you pull into a station and measure the the exhaust just after a lengthy high rpm run on the freeway, you'll get much lower numbers than if you test a minute or two after letting the car sit for 4-6 hours.

Differences in the gas analyzer used, or how it's used, will also show variations. The standards are actually pretty high compared to what the "average" car produces, largely to account for variations in testing from equipment calibration, operator error, and test conditions. The standard for '87 is 220ppm, but you're getting under 10ppm most of the time. I've typically seen 10-30ppm on cars with a cat, and 150-200 on well-tuned cars w/o a cat.

Ask any experimental scientist (or better yet, one of their grad students) how difficult it is to get accurate measurements even with lots of care and training. The variations are really only 2-3x, for the most part. Getting results that GOOD in the usual environment of a smog testing station is pretty remarkable. Note that you could still pass if the test was no more accurate than a 10-20x variation.