Combustion analyzer calibration & accuracy

Boon

Hey. I sent my analyzer, a UEI C85, in for its annual calibration. The 'certificate of compliance' that I received with the unit states the following:

CO2 Signal Input: 9.01%
Instrument Reading: 9.0%
Accuracy Specification: +/- .3%

That 'Accuracy Specification' of point 3 percent seems like a big variance when I&O manuals spec in the tenths of %.

So when my boiler is groaning now & then at low fire but my analyzer says 9.1% CO2, should I push the CO2 up to 9.3% or 9.4% or 9.5%, or are there additional subtleties to professional judgement in cases like this?

NY_Rob

The install manual for my mod-con instructed you adjust the combustion to factory specs, but if "foghorning" occurred at low-fire it instructed you to enrich the mixture slightly to remedy the condition.
Make your boiler happy!

Boon

Yeah. There is a range. I’m over thinking it. I was going for 9.2, though, based on recommnedation from tech support

STEVEusaPA

That's plus or minus. I wouldn't even worry about it. What were the rest of your combustion test result numbers?

BobC

I have someone in every couple of years to check my Smith G8 with it's Carlin EZ-Gas burner. The guy did his thing and also checked my heatpump to make sure it was ok. A couple of days later I looked at the tag on the boiler and the printout said the CO was 221ppm.

I put my old bacharach analyzer on it (needs a o2 sensor) and it said the CO was 700ppm AND CLIMBING. I killed the power to the boiler and went up[stairs to order a new analyzer.

A week later the new Testo 310 came in and that said the CO was 670ppm. I opened the airband and got the CO under 25ppm with O2 and CO2 well within specs. i sent the service company a letter on the service they provided, it will be interesting to see what they say. Lesson learned, from now on I'll do my own analyzing and hire people only when I need heavy work done that I can't do any more.

As to the accuracy of these analyzers, I'd be surprised if they are any better than 5%. I worked in calibration for a decade and i know how it's all done. These are not precise instruments, they are ballpark instruments, if you want to see precision instruments I'll show you my Tektronix storage scope or my Fluke 8060A DVMs that are rated at 0.05%.

Those Fluke's cost $400 back in the early 80's and was well worth the price. Having worked on those a lot I can tell you the electronics inside a combustion analyzer is worth about $50 retail, the rest is the two sensors and profit margin. There isn't enough competition out there, I'm really surprised nobody hasn't bought one out for a couple of hundred dollars - they would make a killing.

Bob

Solid_Fuel_Man

Bob, you should build them and sell them to us. I'll be the first to buy one. Funny you would talk about calibration, I just calibrated my Richie inches of water diagram gauge this morning, slack tube manometer (home made) never lies. Used to carry that thing to jobsites years ago.

I think there must have to be a paper trail for calibration for compliance which I'd wager has significant cost associated with tool price.

captainco

The btus in the gas can be 900 to 1100 per cu.ft. That makes the CO2 from 10.2 to 12.2. CO2 is more like guessing. We always makes 99.99% of the CO2. It just gets diluted with O2. Therefore the lower the O2 the higher the CO2 readings will be. O2 is always more accurate. Equipment has a range not a fixed single point. We set within that range and the equipment will decide where it likes it.

BobC

@Solid_Fuel_Man traceable calibration requires you to have instrumentation that is traceable back to NIST. I used to have some instruments I used as "standards" that had to be sent to a calibration lab where they checked those instruments against their standards that were traceable back to NIST.

If something did not meet spec you were told exactly how far off it was. The company I worked for had about 500 pieces of test equipment and I had a folder on each one stating when it was checked and exactly how it compared to my standards. There was a lot of paperwork.

I can see why the sensors in combustion analyzers are costly because they are a specialty item and there aren't enough made to get any real economy of scale. i did find automotive sensors cost about $20 each so why do we have to pay $100-150 each for the ones in our analyzers?

Someday someone will use one and that might help bring the costs of the instruments down.

Bob

CLamb

Here is a list of gas sensors available for the Arduino microprocessor platform: http://playground.arduino.cc/Main/MQGasSensors The ones which would be of most interest to the folks here would be the CO and CO2 sensors. The sensors themselves just produce an analog voltage proportional to the concentration of the gas sensed but microcomputer platforms usually include a ADC anyway. You could probably put one together for circa $50 for those gases which communicate with a cell telephone to use as a display. The expensive one seems to be the CO2 sensor at $30. You can see the prices for them direct from China by searching on the sensor id at aliexpress.com. I'd been tempted to throw one together and see how the accuracy compares to a real meter but I have enough to do already.