Taylor R-0870 test powder (crystals) - hard to measure for anyone?

[waterbear]

Someone on another forum did a test using different scoops and they did get slightly different readings but generally only differing by 1 drop or sometimes 2 if the transition was on the edge between drops. The reason is mostly that with less dye the transition from light pink to clear is a little more subtle.

And since the precision of this test is 1 drop this is not significant at all! People seem to loose sight (or don't know or understand) that ALL testing methods have an inherent margin of error. Drop counting tests are normally +/- 1 drop. With the Taylor K-2006 that would be +/- .5 or .2 ppm for the FAS DPD test, depending on whether the sample is 10 ml or 25 ml. and for the TA and CH tests it would be +/- 25 ppm or 10 ppm once again depending if the sample is 10 ml or 25 ml. On the pH test you can interpolate between 2 color blocks so you can get an accuracy of .1 unit. However, the CYA scale is not linear so you really cannot interpolate between 2 markings so your accuracy is +/- 10 ppm.

Many people confuse accuracy with precision. Accuracy means that the test is producing results that can be trusted and that they are correct. Precision is the amount of variation of results you get when you repeat the test on the same sample.

For example, let us say we have a standard solution for testing CH that contains 100 ppm. If we test it and always get 80 ppm from a particular test kit on 4 separate tests on our standard then that test kit precise (it give us the same result every time we test on the same sample) but it is not accurate (it is giving us an erroneous result).
Now let's test our standard with a different test kit. We get results of 70 ppm, 90 ppm, 100 ppm, and 120 ppm upon testing 4 times. Since we cannot get results that repeat at all this test kit is not precise. It might be accurate but have a very large "window" or resolution so we are not really getting any meaningful info (such as test strips that have as much as 40 ppm between color blocks). Actually, if we average our 4 tests we come up with 95 ppm so, in reality, this test kit is is fairly accurate but not very precise so the results from one individual test are not going to give us useful info but if we average many tests we can get useful info. This is time consuming so it is not a good solution.
Now, let's use a kit that has both good accuracy AND precision (such as a Taylor K-2006). Using a 25 ml sample (which should give us a precision of +/- 10 ppm because the test has an accuracy of +/- 1 drop) we get results of 90, 100, 100, and 110 ppm. If we average these results we get 100 ppm. In other words, we should always get 100 ppm but a result of 90 or 110 ppm are considered to be acceptable and in range and in the grand scheme of things with a pool or spa are not going to matter very much at all since the accuracy and precision of our water volume measurement usually have a much larger error when it comes to how much of a chemical to put in to adjust a water parameter. This is why we usually say to add less than what you think you need of a chemical and retest to "creep up to your goal" so you don't overdose. The only exception I can think of is when shocking a pool, where it is generally better to slightly overdose with chlorine than to slightly underdose.

To summarize, we just need to remember that this is a swimming pool, not a science experiment. We need a test kit that will give us reliable, repeatable results with small enough margin of error on any one test that we know that we are about where we need to be but we don't need to become obsessive that when you do a test one time you need 6 drops and when you repeat it you need 7 (or even 8!) so something must be wrong with the testkit. You are within the accepted margin of error for that test! If you consistently have an error greater than that then there is a problem somewhere and the first thing I would look at is your testing technique.

Also, for those of you using 'digital testers' be they strip readers or colorimeters with liquid or tablet reagents be aware that this does not change. Just because you get test result reading with 2 decimal places does not mean that it is accurate (repeatable) nor does it mean that you have two decimal points of precision. A perfect example is a respected company's pool store testing lab that uses little vials of dry reagent that have water added to them with a calibrated pipette and are then read in a colorimeter that costs over $1000. Lets use the TA test as an example It has a precision of + 10 ppm and - 20 ppm even though the readout gives you a number like 98 or 71, so one would assume the precision is 1 ppm, which is totally false. If one saw these two numbers on two readoutsand wanted a target TA of 80 ppm they would think that they would need to lower the TA in the first case since it is almost 20 ppm to high or bump it up a bit in the second case since it is almost 10 ppm too low However, when we look at the precision of this test we realize that if our TA is, in reality, 80 ppm then both test test results are within the precision of this testing system. The problem is that when you get your printout from the pool store it just says that your TA is 98 ppm or 71 ppm with instructions on how much "Akalinity UP" or "pH Down" that you need to bring the reading to the "ideal number". Just because a testing method is giving you a digital readout does not mean that is the precision of the test.

[JimK]

And since the precision of this test is 1 drop this is not significant at all! People seem to loose sight (or don't know or understand) that ALL testing methods have an inherent margin of error. Quote trimmed by mod

Great post! Thanks.

[AnnaK]

Does this make sense?
I'm probably wrong on the chemistry but I think the analogy works.

Loved your post!