A few days ago, someone asked me about a recommendation for a high end Refractometer recommendation. Theirs wasn't acting in a way they thought was accurate.

Quite frankly, mine is a fairly cheap model, so I don't have a recommendation. But, it got me to thinking about testing the accuracy. I've trusted mine for years, now and have taken measurements by hydrometers to back it up, too. I'm only worried about accuracy to a degree, since there is an acceptable range of variation to brewing.

So, I created a series of steps that are repeatable and can help you determine the accuracy of either a refractometer, or hydrometer.

Refractometers tend to read in Brix, which is close, but not precisely like Plato. For added accuracy, you'll have to make the conversion. Brix is convenient that it references almost 70F (68F), which is easy to achieve. Plato refrences 64F.

The Brix conversion to Specific Gravity is not linear. At high gravity (25) it is x4.23, and at low gravity (3) it is x3.91; both of which are close enough to 4, for me. For purposes of this post, I'll use Brix and the full conversion to Specific Gravity.

Two calibrations were performed. First, zero brix was set with distilled water. Simply, the steam from a pan of boiling water was condensed and read. This is easily done with a clean plate or pot lid, as only a drop or so is needed. The second was to read my tap water, which came to 0.3, and all of the readings were adjusted down by this amount.

I used four hydrometers. Two were the common cheap hydrometer sold in most homebrew stores and two were a "professional" long stem model with an integrated thermometer. They varied on a tap water reading at 1.002 to 1.004. Their average was used to compute the readings I note, later. Half mark readings were rounded up.

To start the experiment, the refrence was 100 grams, as 25 grams of sugar in 75 grams of water. After that, it was a simple linear dilution rate to see if there was a linear reduction in readings. This 100 gm reference should yield a reading of 25 Brix. I heated the water in the microwave to speed up dissolving the sugar. Just warm, not hot or boiling. Temperature was irrelevant since the refractometer auto corrects and I took temperatures of hydrometer readings, making the correction needed.

The margin of error for making the samples was wider than I’d like. My most refined scale measures whole grams, so care had to be taken to carefully get just to the weight specified. Without better resolution, it is very easy to be up to ¾ gram heavy. I measured the sugar weight twice in each instance, but that alone may not reduce the margin of error, much.

My first reading was 25.6 brix. The 100 gram sample is too small for a hydrometer reading.

100 grams of cold water were added. The refractometer read 12.8 brix.

From that sample, I took 100 grams of it and added another 100 grams of water. The reading was 6.5 brix. At this point, I had enough sample to use a hydrometer. The prediction would be 1.028. The actual reading was 1.029, when temperature corrected.

Again, 100 gm was diluted with 100 gm of water. This brings us to an accumulated 8x dilution rate. The mathematical prediction would be 3.2 brix. The hydrometer reading was 3.3 brix. The hydrometer prediction would be 1.013. The temperature corrected hydrometer reading was 1.014.

Since the question was about lower end readings, I did a second set of tests, using 2 grams of sugar and 98 grams of water. This should represent the lowest runnings of a sparge at 1.008, or 2 brix. I did this 3 times. The results were pretty consistent at 2.2 to 2.3 brix. This would calculate out to 1.009 specific gravity.

My conclusion is that my cheapo refractometer is reliably accurate to within 0.4 brix, a margin of less than 1.002 specific gravity. My cheapo refractometer appears to be both linear and accurate enough to create good and reproducible brewing records.

Very interesting. Thanks Kevin. Maybe I have a bad refractor. Mine has errors at the low end, plus the reading is fuzzy down there. I get a good line when calibrating with distilled water. I also get good reading and no fuzzy lines at the higher end when I start my must. I'm going to do some more tests.