Hydrometer using a load cell

[The-Engineer-That-Brews]

I've recently been making a hydrometer using the tilt principle. It's a clever idea but in practice it seems hard to get repeatable accuracy; possibly due to the effect of krausen build up on the lid. Also from my perspective it's a bit of a pain having to use batteries.

So I'm going to try a more direct measurement of the buoyancy of a submerged object - a bit more like the principle of a traditional glass hydrometer.

I'm planning to use a small float held under the surface by a thin stiff vertical wire from a sensitive (100 gram) load cell mounted on the side of the FV:

If I use a float that's about 100ml volume then the buoyancy force should be around 1N (100g), changing about 0.1g for every degree of specific gravity. That corresponds to about 0.1% of measurement accuracy from the sensor which should be fairly easily achievable (10 bit accuracy)

Sources of potential inaccuracy I can think of include the buoyancy force on the wire (negligible) and sensor creep - unknown, but hopefully less than 0.1%.

Thoughts, comments, ideas?

 

[The-Engineer-That-Brews]

I was initially thinking of having the float as a small bottle full of air, but it probably makes more sense for it to have neutral or slightly negative buoyancy because then (a) I can replace the stiff wire with a thin nylon filament; and (b) the effect of sensor creep will be much less.

 

[Cwrw666]

Wouldn't the depth under the surface affect the reading?

 

[The-Engineer-That-Brews]

Wouldn't the depth under the surface affect the reading?

Good question... I don't think so, so long as the float is completely submerged
There's a reasonable explanation of why here: What is buoyant force? (article) | Fluids | Khan Academy

"Surprisingly the buoyant force doesn't depend on the overall depth of the object submerged. In other words, as long as the can of beans is fully submerged, bringing it to a deeper and deeper depth will not change the buoyant force. This might seem strange since the pressure gets larger as you descend to deeper depths. But the key idea is that the pressures at the top and bottom of the can will both increase by the same amount and therefore cancel, leaving the total buoyant force the same."​

 

[The-Engineer-That-Brews]

Useful info on sensor creep here: Does this loadcell's creep mean that I can not let a constant load on the loadcell?

 

[The-Engineer-That-Brews]

Might be wise to make the float cone-shaped to prevent yeast sediment accumulating on the top...

 

[moto748]

As an engineer myself,finding this interesting. I have done a bit of work on weighing systems in the past; I remember load cells first coming onstream, and they used to be crazy expensive, nowadays they are pretty cheap.

Daft thought, but if all we are trying to do is determine the density of a liquid, can that not be achieved by simply weighing a fixed volume, zeroing the tare on a measuring jug or similar? Maybe using a pipette or similar for the best accuracy?

 

[The-Engineer-That-Brews]

Daft thought, but if all we are trying to do is determine the density of a liquid, can that not be achieved by simply weighing a fixed volume, zeroing the tare on a measuring jug or similar? Maybe using a pipette or similar for the best accuracy?

Yep, that should work (or for more accuracy you could use a pycnometer)
However I'm really looking for a continuous measurement :-)

 

[Buffers brewery]

I've recently been making a hydrometer using the tilt principle. It's a clever idea but in practice it seems hard to get repeatable accuracy; possibly due to the effect of krausen build up on the lid. Also from my perspective it's a bit of a pain having to use batteries.

So I'm going to try a more direct measurement of the buoyancy of a submerged object - a bit more like the principle of a traditional glass hydrometer.

I'm planning to use a small float held under the surface by a thin stiff vertical wire from a sensitive (100 gram) load cell mounted on the side of the FV:

View attachment 53814

If I use a float that's about 100ml volume then the buoyancy force should be around 1N (100g), changing about 0.1g for every degree of specific gravity. That corresponds to about 0.1% of measurement accuracy from the sensor which should be fairly easily achievable (10 bit accuracy)

Sources of potential inaccuracy I can think of include the buoyancy force on the wire (negligible) and sensor creep - unknown, but hopefully less than 0.1%.

Thoughts, comments, ideas?

Neat! So if the “float” is 100 ml and weighs 110 grm and is immersed in a brew of OG 1.100 the tension in “string” would be zero. As the brew ferments out the FG ends up at 1.010 creating an upthrust of 100*1.010=101 grm on the float creating a tension of 9 grm in the string. Is that right?

 

[The-Engineer-That-Brews]

Neat! So if the “float” is 100 ml and weighs 110 grm and is immersed in a brew of OG 1.100 the tension in “string” would be zero. As the brew ferments out the FG ends up at 1.010 creating an upthrust of 100*1.010=101 grm on the float creating a tension of 9 grm in the string. Is that right?

Yep that’s my line of thinking - yet to be proven!

 

[Buffers brewery]

Yep that’s my line of thinking - yet to be proven!

Can't see why it shouldn't work. The challenge will be calibration and repeatability.

 

[The-Engineer-That-Brews]

Can't see why it shouldn't work. The challenge will be calibration and repeatability.

Calibration shouldn't be too hard... measure the volume of the float by Archimedes' method, and then weigh it. The calibration of the load cell should be ok - it's the same chip as is used in my kitchen scales.
We shall see!

 

[johncrobinson]

The only problem so far not discussed I can see is CO2 sticking to the float thereby affecting the boyancy.

 

[The-Engineer-That-Brews]

The only problem so far not discussed I can see is CO2 sticking to the float thereby affecting the boyancy.

Very good point… might need a float-spinning motor

 

[peebee]

Rather than end up with an over-complicated hydrometer, wouldn't it be better looking at what a hydrometer tries to tell you?

Before hydrometers that's what they did: Weigh a barrel of beer and calculate the sugar it contains. But it wasn't very accurate with the instruments they had.

But now? If you can determine the volume of beer, determine the weight of beer, and determine the temperature of the beer, then you've got all you need to make measurements far better than a hydrometer can tell you.

You'll be arriving at "density", but converting to a beer-hydrometer's "relative density" ("specific gravity") is easy-peasy ... divide by the density of water at 20C (0.9982).


(And I've written that post without once repeating what I keep blagging on about ... a pyknometer. Err ... damn, I just have mentioned it haven't I.).

 

[The-Engineer-That-Brews]

Rather than end up with an over-complicated hydrometer, wouldn't it be better looking at what a hydrometer tries to tell you?

Before hydrometers that's what they did: Weigh a barrel of beer and calculate the sugar it contains. But it wasn't very accurate with the instruments they had.

But now? If you can determine the volume of beer, determine the weight of beer, and determine the temperature of the beer, then you've got all you need to make measurements far better than a hydrometer can tell you.

You'll be arriving at "density", but converting to a beer-hydrometer's "relative density" ("specific gravity") is easy-peasy ... divide by the density of water at 20C (0.9982).


(And I've written that post without once repeating what I keep blagging on about ... a pyknometer. Err ... damn, I just have mentioned it haven't I.).

Because, as I said above, I want a continuous measurement :-)

 

[peebee]

Because, as I said above, I want a continuous measurement :-)

But then that's the problem to solve. And it would be really handy to have a continuous measuring method that could be reapplied to figuring how much beer is left in a SS keg.

 

[The-Engineer-That-Brews]

continuous measuring method that could be reapplied to figuring how much beer is left in a SS keg

That is in fact a project in progress :-)

 

[ChrisT]

I was initially thinking of having the float as a small bottle full of air, but it probably makes more sense for it to have neutral or slightly negative buoyancy because then (a) I can replace the stiff wire with a thin nylon filament; and (b) the effect of sensor creep will be much less.

I think negative bouyancy is the way to go, the force is then always downwards and as you say, you can use a wire.

I like the idea! I'm now going to do a bit of shopping for parts. I make wine so I'm not what the weight I'd use is going to be yet, time to experiment

The only problem so far not discussed I can see is CO2 sticking to the float thereby affecting the boyancy.

I'm thinking I'll start with an oval shaped "sink" if I can find one? Hopefully that would reduce bubbles and gunk sticking to it? For example for the same reasons, I know the iSpindels I use aren't wonderfully accurate in the middle of their range but I'm very happy with them at the begining and end of fermentation.

Chris

 

[peebee]

The problem with the density approach, and the reason we use hydrometers, is you do need to be fairly accurate with measured weights and volumes. After all, the figure you're after (SG) is to three decimal places (4 for the new Tilt Pros, or so they claim). Still, I manage with a sample about 25ml in size (which needs weighing to within 0.0125g), when the "sample" is over 20 litres (and 20kg+ in weight) in size, how hard can it be?