Aquarium Advice - Aquarium Forum Community - View Single Post - How to make a reference KH solution for CO2 measurement

Zezmo · 01-07-2007, 04:11 AM

Interested in how accurately you can use a test?
I posted most of this info as a response to a question about hobby tests. Since this sticky thread is about making a reference solution, I though I would add some info about how to determine if you are measuring your reference solution in a repeatable fashion.

There is a simple study called a Gage R&R (repeatability, and reproducibility). It will take some work, but you only go through this series once and it will tell you a lot.

In it's simple form a GR&R basically consists of taking 5 consecutive measurements on 2-9 separate occasions, and then plugging the results into a formula. (the link below has a spreadsheet that does all the math for you).

Taking Ph as an example because it is the most common test most folks use (though can do a Gage R&R for Kh or any other testable parameter). Ideally, you would buy a gallon of bottles spring water from the store and use it for the entire series of tests (or make a reference solution as described in this thread). This should give a reasonable constant sample to work with so that you are only focusing on accuracy of the test and the tester.

You would using this water do 5x Ph tests in a row and record the results. Do this at least 2 days in a row to get a total of 10 measurements. (minimum)
To really get good data, you should do this 3 days in a row, on 3 consecutive weeks. Sounds like a lot, but again you only ever need to do this once.
Plug these results into the spreadsheet. The results in yellow at the bottom of the spreadsheet will give you an idea of how closely you can rely on your test results.

http://webpages.charter.net/zezmo/Ga...quarium_v1.xls

Note: A gage R&R looks at variations in the test and tester, not the actual test accuracy. Accuracy can only be verified against a reference solution (calibration). However, without repeatability, and reproducibility the actual accuracy of the test does not even come into play.

01-07-2007, 04:11 AM	#15
Zezmo Aquarium Advice Addict Community Mentor Join Date: Mar 2005 Location: Northern California Posts: 1,113 Images: 14	Interested in how accurately you can use a test? I posted most of this info as a response to a question about hobby tests. Since this sticky thread is about making a reference solution, I though I would add some info about how to determine if you are measuring your reference solution in a repeatable fashion. There is a simple study called a Gage R&R (repeatability, and reproducibility). It will take some work, but you only go through this series once and it will tell you a lot. In it's simple form a GR&R basically consists of taking 5 consecutive measurements on 2-9 separate occasions, and then plugging the results into a formula. (the link below has a spreadsheet that does all the math for you). Taking Ph as an example because it is the most common test most folks use (though can do a Gage R&R for Kh or any other testable parameter). Ideally, you would buy a gallon of bottles spring water from the store and use it for the entire series of tests (or make a reference solution as described in this thread). This should give a reasonable constant sample to work with so that you are only focusing on accuracy of the test and the tester. You would using this water do 5x Ph tests in a row and record the results. Do this at least 2 days in a row to get a total of 10 measurements. (minimum) To really get good data, you should do this 3 days in a row, on 3 consecutive weeks. Sounds like a lot, but again you only ever need to do this once. Plug these results into the spreadsheet. The results in yellow at the bottom of the spreadsheet will give you an idea of how closely you can rely on your test results. http://webpages.charter.net/zezmo/Ga...quarium_v1.xls Note: A gage R&R looks at variations in the test and tester, not the actual test accuracy. Accuracy can only be verified against a reference solution (calibration). However, without repeatability, and reproducibility the actual accuracy of the test does not even come into play.