Combustion analyzer tolerances

Robert O'Brien

Robert O'Brien

I was quite surprised by the plus minus allowable error on O2 and every number derived from it.

mattmia2

That means you want to aim for the center or at least that distance from the edge of the range to allow for the accuracy of the instrument. Welcome to the world of measurement.

ChrisJ

Robert O'Brien said:

I was quite surprised by the plus minus allowable error on O2 and every number derived from it.

It can be within 0.6%?
Doesn't seem that bad.

Most thermometers are +- 2 degrees, so a 4 degree tolerance and yet no one questions it and they give you tenths so it must be accurate.

mattmia2

Wait until you start looking at the accuracy of thermocouple thermometers built in to DMMs.

Jamie Hall

Shouldn't be, really -- these things are remarkably hard to measure accurately, even under optimum conditions. Then when you go out in the field, in some dark basement somewhere, you throw in a host of other variables which complicate matters.

Robert O'Brien

Jamie Hall said:

Shouldn't be, really -- these things are remarkably hard to measure accurately, even under optimum conditions. Then when you go out in the field, in some dark basement somewhere, you throw in a host of other variables which complicate matters.

Not very difficult at all!

Robert O'Brien

ChrisJ said:

Robert O'Brien said:

I was quite surprised by the plus minus allowable error on O2 and every number derived from it.

It can be within 0.6%?
Doesn't seem that bad.

Most thermometers are +- 2 degrees, so a 4 degree tolerance and yet no one questions it and they give you tenths so it must be accurate.

ChrisJ said:

Robert O'Brien said:

I was quite surprised by the plus minus allowable error on O2 and every number derived from it.

It can be within 0.6%?
Doesn't seem that bad.

Most thermometers are +- 2 degrees, so a 4 degree tolerance and yet no one questions it and they give you tenths so it must be accurate.

Chris, the proper O2 readings are in a narrow range, about 2.5%, 6/10ths is huge!

ChrisJ

Robert O'Brien said:

ChrisJ said:

Robert O'Brien said:

I was quite surprised by the plus minus allowable error on O2 and every number derived from it.

It can be within 0.6%?
Doesn't seem that bad.

Most thermometers are +- 2 degrees, so a 4 degree tolerance and yet no one questions it and they give you tenths so it must be accurate.

ChrisJ said:

Robert O'Brien said:

I was quite surprised by the plus minus allowable error on O2 and every number derived from it.

It can be within 0.6%?
Doesn't seem that bad.

Most thermometers are +- 2 degrees, so a 4 degree tolerance and yet no one questions it and they give you tenths so it must be accurate.

Chris, the proper O2 readings are in a narrow range, about 2.5%, 6/10ths is huge!

Except whoever published those specifications knows the equipment that is used to test. They would include that tolerance in their spec, no? You would hope, anyway.

Kind of like air conditioning spec'ing a SH of 10-15 degrees when in a perfect world 1 degree is more than enough. They know temp probes will be wrong etc so they give a huge safety factor.

mattmia2

The additional superheat (or subcooling) in a charging spec accounts for more extreme operating conditions. Usually there is a chart with outdoor dry bulb and indoor wet bulb temps that tells you what part of the curve to aim for when charging a system.

ChrisJ

mattmia2 said:

The additional superheat (or subcooling) in a charging spec accounts for more extreme operating conditions. Usually there is a chart with outdoor dry bulb and indoor wet bulb temps that tells you what part of the curve to aim for when charging a system.

There's still a very big safety factor included for user error and instrument tolerance.

Jamie Hall

All true. Fortunately, in many cases, what one is looking for is not an exact number, but a peak or a valley -- or a trend. I admit to being not particularly familiar with combustion measurement for boilers and furnaces -- but I am for internal combustion engines or gas turbines. There one is almost always going for a particular peak (or valley) in some measurement, or a situation where some value is stable or close to it, and then starts to change more quickly. The exact value of the measurement isn't particularly helpful (provided it's at least reasonable!) -- but the change is. Thus one wants precision and repeatability (that is, you get the same measurement every time you measure the same thing), with accuracy (conformance to a standardized value) being somewhat less important.

STEVEusaPA

+-.03 for draft is pretty severe. Wonder if I should go back to the old school Bacharach for draft, and how accurate is it?
https://www.positive-energy.com/wp-content/uploads/2016/03/MZF-1S.jpg

Robert O'Brien

STEVEusaPA said:

+-.03 for draft is pretty severe. Wonder if I should go back to the old school Bacharach for draft, and how accurate is it?
https://www.positive-energy.com/wp-content/uploads/2016/03/MZF-1S.jpg

It is too much wriggle room for my comfort also

ChrisJ

Magnehelic model 2000-0 is +-1% with the HA option.

EBEBRATT-Ed

Combustion is not that stable. It changes with air temperature, humidity, fuel temperature etc. You looking to keep the equipment weather it is combustion, refrigeration etc operating within a safe range. Spending too much time on it is a waste of time

CLamb

I don't know much about combustion analyzers but looking at the figures for O2 I see two different ways the tolerance can be read--either as 0.3 percentage points or 0.3 percent of the reading. From that table alone I would favor the latter interpretation.

Zman

CLamb said:

I don't know much about combustion analyzers but looking at the figures for O2 I see two different ways the tolerance can be read--either as 0.3 percentage points or 0.3 percent of the reading. From that table alone I would favor the latter interpretation.

Judging by the units called out in the other rows, I believe they are referring to %O2

Zman

Anytime I have used 2 analyzers on the same system to check accuracy, they almost always read at least 0.2 percent different.
I think it is just the nature of it.
Outdoor air temp on sealed units will change the reading far more than 0.3 percent, not to mention random fuel quality and draft changes.
It is truly an imperfect world. All we can do is our best with the limitations we are given.