Hydrometer and refractometer errors

[dunbruha]

So I brewed my first batch yesterday, and took measurements of my gravities at various times throughout the process. I used a hydrometer and a refractometer (an ATC model with specific gravity markings directly on the readout). I noticed that the readings were very different between the two instruments. So I decided to make a set of standard solutions of DME at various gravities (0, 10, 20, 30, 40, 50, 60, & 70) and see what values each method gave me. The graphs and the data are attached below. (The hydrometer readings are corrected for temperature—everything was at 72F.)

You can see that the hydrometer was 5 points too high from plain water to 1.030, and then the error started to shrink, getting to pretty much spot on at 1.070.

In contrast, the refractometer started out correct in plain water, but the error increased in an increasing rate up to 1.070.

Is there any way that I can apply a correction factor to these graphs? Neither one has a constant error, and my math skills aren’t good enough to figure out what to do. I’d appreciate any ideas!

 

[BlueHouseBrewhaus]

I don't use a refractometer but I see two possible reasons for your hydrometer errors. First, the hydrometer should read 1.000 for distilled water at 60F. That is pretty easy to test. Try that with actual 60F water (i.e. no need for temp correction) and see what you get. If it's still off, the hydrometer is bad. The other possibility is in how you made up your standard solutions. If the water measurement or DME weight were not totally accurate, this could lead to an error in your readings. In other words, your standard solutions may have actually been not quite "standard". However, I would still go back to your plain water reading first and double check it.

 

[oljimmy]

Wow. That's terrible. If you really were careful with those measurements, then you have a bad refractometer, and you've motivated me to test mine in the same way. Also, while BlueHouse is technically right that you should test the hydrometer at its calibration temperature, 12 degrees F is not nearly enough to produce a 5 point difference (1 or possibly 2 at most), so you seem ti have a bad refractometer.

Are you using a scale to weigh the DME, or going by some kind of teaspoon/tablespoon method? Obviously the latter is insufficient.

One thing: these numbers could partly be explained if your value for gravity points added by DME is off. Actually, now that I think of it, the fact that the hydrometer error *changes* as you go up makes me think that the DME numbers have to be off. A hydrometer is simply responding to density changes, and I can't see what would make a little glass thing respond differently to different changes in density. It would make much more sense if you both (1) have a bad hydrometer, and (2) aren't using the right density numbers for your DME.

 

[Gavin C]

What's the meaurement error for the scales used to measure the DME. How was the sacles calibrated

What's the meaurement error for the volumetric container used to measure the water volume. Was the water at the correct temperature in the flask to match it's calibration temperatures.

Did you add DME and then water to reach the correct volume and gravity or take the water volume and add the DME second? The latter will introduce error, increased volume and reduced gravity.

In short, this data does not mean the hydrometer or refractometer are unacceptable without exacting calibration of the other measuring tools to measure mass and volume.

I would recheck the hydrometer in distilled water with the sample and hydrometer at the corret calibration temperature. (usually 60 or 68F)

 

[justenoughforme]

i know those errors look linear, but they are actually logarithmic, and as such correcting for them will not be possible entirely. extend it out to 1.1 SG, and see what the linear regression tools give you ( it wont be linear, might read as quadratic due to rounding limitations in Excel.)


not unless your correction factor has a variable for input S.G., and thus could curve fit this reading.


but seriously, just adjust the slope of the graph, know your error. you're going to be REALLY close eyeballing it.

to correct for the logarithmic nature of the error, you'd have to go out to something like twelve decimal places.

 

[dunbruha]

What's the meaurement error for the scales used to measure the DME. How was the sacles calibrated

What's the meaurement error for the volumetric container used to measure the water volume. Was the water at the correct temperature in the flask to match it's calibration temperatures.

Did you add DME and then water to reach the correct volume and gravity or take the water volume and add the DME second? The latter will introduce error, increased volume and reduced gravity.

In short, this data does not mean the hydrometer or refractometer are unacceptable without exacting calibration of the other measuring tools to measure mass and volume.

I would recheck the hydrometer in distilled water with the sample and hydrometer at the corret calibration temperature. (usually 60 or 68F)

The hydrometer reads 1.004 in DI water at 60F (the calibration temp for the hydrometer). I knew it was off, but I thought it would be linear. Maybe my measurements of the "known" solutions are what is causing the non-linear error for the hydrometer. My scale reads to 0.1g, and I was careful in the weighing. But I do not have a volumetric flask, so I was adding the weighed DME to a constant volume of water (150ml for each solution). Perhaps the larger volumes of DME at the higher gravities was causing the non-linear behavior.

I guess I could make up the solutions by weight (add the appropriate amount of DME, then bring up the weight with water). ex: to make 150ml of a 1.040 solution, weigh out 17.1g DME, then bring the total weight up to 150g with water. Would that work?

 

[Gavin C]

The hydrometer reads 1.004 in DI water at 60F (the calibration temp for the hydrometer).


....weigh out 17.1g DME, then bring the total weight up to 150g with water. Would that work?

Time for a new hydrometer I reckon.

I use a set of 3 narrow guage ones. Really very useful.

1.000-1.070
1.060+
0.980-1.020

I have a refractometer. It's Ok for pre-ferment but after alcohol is present you are relying on the accuracy of a an algorithm not the device. I concluded there was no benefit to me using the refractometer. I no longer use it at all.

I don't know if what you describe will work. You are still subject to measurement errors (unknown) in the scales.

I too would have thought the measurement error in a hydrometer would be a constant. (i.e. correct by subtracting 0.004). I guess I'm more stupid than I thought.

 

[justenoughforme]

IF your scale is accurate enough, weighing the DME and adding weighed ::EDIT:: WATER is the best way for consistency

 

[Gavin C]

yea i think so too

 

[stpug]

So I brewed my first batch yesterday, and took measurements of my gravities at various times throughout the process. I used a hydrometer and a refractometer (an ATC model with specific gravity markings directly on the readout). I noticed that the readings were very different between the two instruments. So I decided to make a set of standard solutions of DME at various gravities (0, 10, 20, 30, 40, 50, 60, & 70) and see what values each method gave me. The graphs and the data are attached below. (The hydrometer readings are corrected for temperature—everything was at 72F.)

You can see that the hydrometer was 5 points too high from plain water to 1.030, and then the error started to shrink, getting to pretty much spot on at 1.070.

In contrast, the refractometer started out correct in plain water, but the error increased in an increasing rate up to 1.070.

Is there any way that I can apply a correction factor to these graphs? Neither one has a constant error, and my math skills aren’t good enough to figure out what to do. I’d appreciate any ideas!

Did you happen to record the brix readings for those several solutions as well? I always feel more comfortable using the brix scale and then plugging into a conversion calculator that takes into account the maltose correction factor.

As for the hydrometer, if you are certain about the error, you can either remember to account for it each time you take a measurement with it, or attempt to knock the paper in the tube the direction it needs to go in order to fix it (using gloved hands and knocking gently on a soft surface - protect your hands because that glass is thin and can easily break).

 

[justenoughforme]

knocking the paper won't help as there is at least one point (and in actuality ONLY one point) where the reading is actually spot on.

moving the paper will be a linear shift and will only change WHICH point reads spot on

 

[dunbruha]

Time for a new hydrometer I reckon.

Perhaps. Although I am not encouraged by this sentence from the MoreBeer web page for a "Final Gravity Hydrometer": "It is common for the economical hydrometers made for the brewing or winemaking hobbyist to be off up 3 points." Mine is more than that, but not by much. I tried knocking the paper down, but it wouldn't budge.

I use a set of 3 narrow guage ones. Really very useful.

1.000-1.070
1.060+
0.980-1.020

Where did you find these?

 

[dunbruha]

IF your scale is accurate enough, weighing the DME and adding weighed ::EDIT:: WATER is the best way for consistency

I will give this a try.

 

[Gavin C]

Perhaps. Although I am not encouraged by this sentence from the MoreBeer web page for a "Final Gravity Hydrometer": "It is common for the economical hydrometers made for the brewing or winemaking hobbyist to be off up 3 points." Mine is more than that, but not by much. I tried knocking the paper down, but it wouldn't budge.



Where did you find these?

Northernbrewer.

Was getting some grains and decided to add them to the order.

 

[bernardsmith]

Ok - so there is some error here but in reality what is the impact of the error is 5 points when you use your hydrometer. An error of 5 points is about 1/2 % ABV. Unless you are being taxed at a level where .5% ABV in any direction will make a difference in the amount you are paying or unless we need to be so anal that we really think that an hydrometer reading in fact does not accumulate but neutralizes every single other error we have made in weighing solids and measuring liquids then we know it is an inherently inaccurate tool. All it needs to be is reliably inaccurate - so if you measured the same sample 100 times in 5 hours and you got the same reading each and every time then the tool is doing its job.. The weakness and the error may be located in you and I and not in it.
I imagine similar issues apply to the refractometer..

 

[justenoughforme]

bernard makes a good point.

precision is far more important than accuracy

 

[phug]

I did a similar thing with my hydrometer and refractomer. I found my refractometer read about 3% low and my Hydrometer read about 6% high. I used my scaled with .1 gram precision to weigh out my sugar, and then to add my water, since 1 ml of water is equal to 1g of water. And besides, the brix scale is define as a percentage by weight, so use your scale all the way.

When you're making your tables to graph your results, keep in mind that your hydrometer likely only has about a 1 or 2 point resolution, so anything less than that is your interpretation of the reading, same with the refractometer (damn fuzzy lines) so even if your tables don't fit PERFECTLY, I think that's probably due to small amounts of measurement errors in ingredients, and or reading errors of the instruments.

So if you can come up with something that closely fits your observed lines, that's good enough.

 

[phug]

if you take your refractometer readings, and plug them into excel usign the following formula to correct them, I think you'll find you get pretty close. =SUM(((A2-1)*0.824)+1) with A2 being your cell containing the reading from the refractometer. I have 82.4% being roughly the deviation.

That being said, I would suggest making some additional solutions, and measuring all the way up to the top of your refractometer and hydrometer's range. You'll get better resolution on the errors at the top of the range, and any measurement errors, or reading errors will make up a smaller portion of the results.

 

[phug]

I don't have anything for your hydrometer readings. but I say we need more data points. And if one of your results is significantly off from the amount of variation of the others, I say redo that solution and measurement.

 

[dunbruha]

if you take your refractometer readings, and plug them into excel usign the following formula to correct them, I think you'll find you get pretty close. =SUM(((A2-1)*0.824)+1) with A2 being your cell containing the reading from the refractometer. I have 82.4% being roughly the deviation.

That being said, I would suggest making some additional solutions, and measuring all the way up to the top of your refractometer and hydrometer's range. You'll get better resolution on the errors at the top of the range, and any measurement errors, or reading errors will make up a smaller portion of the results.

Thanks. I'll re-do the solutions with a weight basis for the water, and see how it goes.