SG and "the elephant in the room"

[PeeBee]

This is all heady stuff! But there is a reason behind it. I ran into it mixing a solution of Calcium Chloride and trying to use SG (measured with a hydrometer) and density (weight/volume) to figure out its concentration (for water treatment of brewing water). I'd also been toying with the use of specific gravity bottles (pyknometers) instead of hydrometers (needing huge samples) and refractometers (needing calculators to interpret results).

But it all started to get overly involved:

On another thread I was getting a bit muddled with density (in g/ml) and relative density (aka specific gravity). We think of them as the same, but they are not. "Density" is easy, it's mass/volume. SG (specific gravity) is what we use in brewing, it is the density in relation to a reference. The reference is usually water at its densest (a tiny smidge lower than 4C) when it rather conveniently has a density of very nearly exactly 1 g/ml. Specific gravity (relative density) has no units, it is a ratio and therefore dimensionless; probably the reason we sometimes give it an imaginary dimension like "g/ml".

SG is not quite so straight-forward (substance being tested may be required at 20C, but the "reference" used, water, is assumed to be at 4C). So this is perhaps the "elephant in the room". It is likely a reason behind:

ASTM and ISO both agree that specific gravity = density/0.9976

https://www.intertek.com/polymers/testl ... astm-d792/

Density, kg/m3 = (specific gravity) x (997.6

Note the correction has 4 units after the decimal point, tiny but SG for brewing is measured to 3 decimal places. The correction will make a difference of two or three points of SG. But:

Who expects their hydrometer to indicate anything other than "1.000" when in pure water (at 20C, or in some cases 16C)? Who expects to apply a small correction to the 1.000 reading to make it perfectly correct? Who doesn't expect their SG to be the same as density in grams per millilitre (cm3 if you like)? And would it makes much difference if we did all those things?

I don't believe so. Who's going to support the contrary?

(Additional reading: https://en.wikipedia.org/wiki/Relative_density).

[Kev888]

Personally I find it helps to view this kind of thing within the context and purpose that it is being applied to. In many home-brewing cases such subtleties or high accuracy is out the window for a number of reasons, and often unimportant in any case. A lot of uses are also relative rather than absolute - e.g. is it still dropping?, how much has it dropped?, how does this compare to the last brew? And there are also issues such as apparent attenuation rather than real attenuation etc etc.

With temperature, device accuracy, reading error and confounding effects of alcohol it is all fairly rough to begin with. There are so many factors that I think there would need to be a special reason (or perhaps special person!) to quantify and compensate for the myriad of subtle inaccuracies, simplifications and inconsistencies involved with gravity readings in a homebrew setting. That said, a refractometer can be several points out in very dark wort, so I'm not suggesting the issue should be entirely ignored either.

[PeeBee]

Personally I find it helps to view this kind of thing within the context and purpose that it is being applied to. …

I'm already using "density" (g/ml) to replace SG for preparing CaCl2 solutions, and correcting "density" to be more like SG (g/ml / 0.9976 = SG-ish) resulted in adding 0.5ml less 8% solution to 25L water. Clearly not correcting it is no issue.

But using "density" (g/ml) in place of hydrometers and refractometers is different (reading, or calculated to read, "SG)". Does anyone else do it? Are there pitfalls? Will 25ml be a big enough test sample, or would it need larger (50ml? 100ml?)?

Using "density" (instead of SG) has only become a practical option recently because of the influx of affordable scales that will weigh 0.1g increments (even 0.01g). Before hydrometers brewers used density, but would then be weighing the entire batch or perhaps a barrel (36 gallons), which would have been slightly inconvenient!

"Density bottles" or "pyknometers" don't need careful aligning of scales by eye (despite the liquid's surface meniscus). I'm obviously attracted by this suffering as I do from double vision plus I can't fix images sharply for several seconds ("saccadic intrusions"; don't ask!). If we don't need to use hydrometers, why do we bother?

Density in g/ml (or scaled up to Kg/L) looks very much like SG, so, for example, a liquid with density 1.050Kg/L will have an SG very close to 1.050. And if "very close" isn't "close enough", shouldn't we expect our hydrometers to read 1.002 in water at 20C?

[PeeBee]

Answering my own questions? Well it was a bit of a long-shot hoping that there's already someone reading this doing the same thing? But an important question was "will 25ml <sample> be big enough". I scribbled it out:

Seems unlikely. The best I could glean was "probably not". Those 0.01g resolution cheapo Chinese scales probably do not have good repeatability (i.e. weigh something one moment, then repeat a moment later and the weight has changed! That's not the same as tolerance).

But you can track down scales that offer a bit more confidence. They do cost 2 or 3 times more but still under £30.


I'm beginning to question whether there is a need to convert SG measured with a hydrometer to density (g/ml). I'm getting the impression the "correction" is due to the hydrometer being calibrated at one temperature and the "reference" sample being at another temperature (4C for water - when it is at its densest)? This is depicted like 20C/4C. But any change from this should be marked. But on hydrometers from a brew shop? So you can get 20C/20C hydrometers, although having the reference at the same temperature as the sample must induce its own errors. They will read 1.000 in pure water.

But I think this stuff is well outside the requirements of home-brewing hydrometers! Does that also mean there is no point adding corrections to have g/ml changer to SG and visa-versa?

[McMullan]

I think you'd need a salometer to measure a CaCl2 solution. Hydrometers work fine for our purposes. They are very consistent. It's just a case of finding one that offers a little confidence. They don't really add any improved quality to the final product so I wouldn't lose any sleep over it.

[PeeBee]

I think you'd need a salometer to measure a CaCl2 solution. Hydrometers work fine for our purposes. They are very consistent. It's just a case of finding one that offers a little confidence. They don't really add any improved quality to the final product so I wouldn't lose any sleep over it.

Seems I have to lose sleep over something. So it may as well be this (I do have plenty of opportunity to catch up on sleep). I lost confidence in hydrometers some time ago anyway. Wasn't helped by digging out my old hydrometer a few days ago and finding it hopelessly inaccurate (I was making the CaCl2 solution, using a beer/wine making hydrometer is just in range if keeping the saline strength down to 9-10%).

My eyesight is cocked up (alignment; focus is fine) so fine graduations and a meniscus is a nightmare. I've suitable weighing scales turning up Friday, but they cost a bit (£27.50) but does include, and uses, two 200g calibration weights (and bells and whistles of course). I've still got to order my pyknometer, not helped by my analytical chemist partner saying she threw dozens out a few weeks ago, and they can be tricky to use. She keeps pointing out I should have used a "salometer" too.

But the big thing I want to avoid by exploring the possibilities of using pyknometers, is fishing out 200-250ml of sample for a fragile hydrometer. I'm sure others might agree, but currently the cost of my (uncertain) alternative is mounting. But my motivation isn't to improve quality, it's to improve my life!

Cheers

[PeeBee]

… But I think this stuff is well outside the requirements of home-brewing hydrometers! Does that also mean there is no point adding corrections to have g/ml changer to SG and visa-versa?

Take a look at this table:

It came from here: https://www.simetric.co.uk/si_water.htm. The "alternative" reference for SG here is water at 16C (60F) although all but old hydrometers for brewing are calibrated for 20C and probably (my guess) use a 20C "reference" (so SG of water is 1.000 at 20C).

Stare at the table long enough and you too will be convinced that (for brewing hydrometers) SG = density (g/ml) without any need to apply a "correction" factor.

[Silver_Is_Money]

You are correct specifically (and only) for the case whereby you read density via your scale method at 4 degrees C, and then correlate that scale measured density directly to specific gravity as "beer hydrometer" measured at 20 degrees C.

The density of deionized water at 20 degrees C is 0.9982, and at 23 degrees C (a fairly typical room temperature), it is 0.9976.

If you are scale measuring for density at nominal "room temperature" you can not make the claim that this density computation is equivalent to the SG reading of a hydrometer.

[PeeBee]

… The density of deionized water at 20 degrees C is 0.9982, and at 23 degrees C (a fairly typical room temperature), it is 0.9976. …

Just to make that clearer for other folk that might read this: Those measures are with a "proper" scientific hydrometer that has been calibrated against a 4C reference sample of pure water. A "brewing hydrometer" will be calibrated to a nominally (often more like "roughly") 20C reference where water at 20C will have a "SG" of 1.000.

And the elephant leaves the room (And don't tread on the daffodils as you go). <EDIT: The elephant never left the room! I just fell in with everyone else and tried to ignore it. See later).

"Fairly typical room temperature" at "23 degrees"? I appear to be an Eskimo! 20 degrees would be nice!

[Silver_Is_Money]

Just to make that clearer for other folk that might read this: Those measures are with a "proper" scientific hydrometer that has been calibrated against a 4C reference sample of pure water. A "brewing hydrometer" will be calibrated to a nominally (often more like "roughly") 20C reference where water at 20C will have a "SG" of 1.000.

And the elephant leaves the room (And don't tread on the daffodils as you go).

"Fairly typical room temperature" at "23 degrees"? I appear to be an Eskimo! 20 degrees would be nice!

Density is not measured via a hydrometer. It is a weight per unit volume measure. Put the elephant back into the room.

[PeeBee]

Aye, you're right, I'm drawing up parallels not tying to change a fact. Have you got any buns to coax the elephant back? It prefers the iced ones with a cherry on top, I can PM an address to send them to.

[PeeBee]

So what was I doing wrong? Go back to the OP and you can see I have TWO scenarios I'm trying to solve. I get a grip of one - using density (g/ml) to replace SG (as measured by hydrometer) - and I think I've answered both. On go the rose-tinted glasses

"Silver" is hinting I've only papered over one issue, but I can't see much in these glasses.

So the glasses are off

That don't help! Let me try Flip, that's where the elephant has gone … nowhere, it's still in the room!

The first scenario works okay 'cos I'm trying to emulate a poxy "home-brewing" hydrometer that reads in "accordance" (very roughly!) to a 20C reference (i.e. water reads 1.000 at 20C). So I use this same method (density) on my CaCl2 solution. It comes out as (26.62g / 25ml = 1.0648, I've got my new swanky weighing scales now; sort of swanky, they are still cheap, just not "as cheap"). So, look up 1.065 against the tables and I have a 7.2% solution …

Except it ISN'T! The "tables" would be drawn up using a proper scientific measuring tool for getting SG, and will be using a 4C water reference (by default, if using anything else it should say so, i.e. on a "real" hydrometer pure water reads 1.002 at 20C). So that "fix" needs applying to the second scenario: "specific gravity = density(g/ml)/0.9976". Look up 1.067 and I have a 7.4% solution.


Fortunately this doesn't derail my attempt to replace poxy homebrew hydrometers with weighing scales and volumetric flask (or "pyknometer"), because as long as precision SG tools don't get a look in there's no need for the "fix".

<EDIT: As well as the "0.9976" fix, I also see references that suggest 0.9982 and I know not why? But the difference it makes isn't worth the trouble of sorting it out.>