Digital hydrometer?

Read about beer bug...not for sale yet. Anyone know of a digital hydrometer?

I can't wait for the collective brewing geek community to jump on figuring out how it works.

Kai

hog2up said:

Read about beer bug...not for sale yet. Anyone know of a digital hydrometer?

Click to expand...

Yes, and they are widely used in the brewing, distilling, soft drink and other industries i.e. anywhere that a quick means of determining the composition of a binary or near binary mixture is needed. Do a search on Anton Paar. They seem to be the major manufacturers of these instruments.

They work by measuring the natural (resonance) frequency of a glass U shaped tube into which the substance to be measured is injected. The resonance frequency depends on the mass of the tube which, of course, includes the mass of the material it holds. Given that the tube holds a fixed volume its mass and thus resonant frequency depends on the density of the fluid in the tube. Resonance is determined by exciting the tube by means of an electromagnet near a permanent magnet fixed to the tube itself and then removing the excitation so that the tube oscillates at its natural frequency. The density is prortional to the square of the period (reciprocal of the frequency) of oscillation normalized by a reference frequency. Calibrataion constants are determined by measuring dried air (barometer reading required) and de ionized water. Temperature is critical. To read to 0.000005 (as the best instruments do) temperature must be controlled to a couple of millidegrees) C.

I should note that I'm guessing as to exactly how the natural frequency is determined. Sweeping the excitation frequency while looking for a dramatic phase change between the excitation frequency and the motion (sensed by a second magnet/electromagnet) should work and in the older ones you could hear the frequency varying but in the newer ones you cant. And the salesmen are kind of vague on the details. I put what I did above because the better instruments correct for viscosity which they could do by measuring the decrease in amplitude over time after removing excitation.

A.J, the beer bug doesn't seem to have this U-tube and I also wonder how one can determine gravity through resonance while there are CO2 bubbles in the solution. The latter is the case during active fermentation.

Kai

FWIW, we had a few discussions on realtime gravity measurements in the past. Here is one that has a lot of information and may be connected to the BeerBug: https://www.homebrewtalk.com/f11/real-time-fermentation-monitor-85661/

Here is some more info: http://www.kickstarter.com/projects/parasitx/the-beerbugtm-digital-hydrometer

There is a sensor that is submerged into the beer.

Kai

Kaiser said:

A.J, the beer bug doesn't seem to have this U-tube and I also wonder how one can determine gravity through resonance while there are CO2 bubbles in the solution. The latter is the case during active fermentation.

Click to expand...

Ah. I didn't realized that the 'beer bug' was an existing product (or quasi existing product) and so responded to the 'ever heard of a digital hydrometer' question.

Yes, CO2 bubbles are a PITA when measuring fermenting wort or finished beer. Even degassing the beer by the usual means isn't sufficient. What I do is draw the beer into a horse syringe, pull back the plunger to put a partial vacuum on it and then shake like mad. Repeating this a few times seems to do the job. What Anton Paar recommends uses the horse syringe too but they suggest putting pressure on the plunger to compress the bubbles in the U-tube. That works too. This is not workable for a 'in line' process however.

For something that goes into a carboy and is intended to read out continuously my first thought would be to measure the differential pressure between the ambient and a point say 10" immersed in the wort. The differential pressure will be 10*SG WC for 10" immersion. Suitable sensors are available from Ashcroft and others. There will be plenty of engineering challenges such as finding a sensor at a cost less than the example Ashcroft unit I looked up ($179), coming up with a configuration for the sense diaphragm, orienting the diaphragm (presumably upwards so that any bubbles that do form on it get shed), compensating for temperature etc.

I note that the Ashcroft sensor has an accuracy of ±0.4% of span. For a 10" immersion and 10" WC sensor that would correspond to an error of ±0.004SG. I also note that the bug people seem to be pretty carefully avoiding any mention of accuracy (or at least I wasn't able to find anything on it based on a cursory look at the linked site).

Further thinking on this has gotten me all excited about what I should be able to do with my cylindroconical fermentors. By attaching a differential pressure gauge one port to the cone and another to the zwickle I would see a pressure difference of

g*(density_wort - density_air)*distance_from_zwickle_to_cone_port

This assumes that there is air, and not wort, in either of the tubes connecting the ports to the pressure gauge. So that leads to a question:
does anyone know where I can get low pressure isolators? These would be made up of a piston or diaphragm with the liquid on one side and air on the other. I know they make them for, for example, isolating fuel in injected engines from the fuel pressure gauge but can't find anything that works at a psi or two.

Note that the scheme I am proposing is not effected by bubbles nor is it affected by barometric pressure or the CO2 blanket, even if the tank has been spunded (i.e. is under positive CO2 pressure).

Why couldn't you just fill both tubes connecting to the pressure gauge with wort? Hard to clean and/or find a pressure gauge that would be wort compatible?

I'm not that familiar with conical fermentor terms. Where's the zwickle port? Is that below the top of the wort or up in the headspace?

So you mean the drawing below.

But I think the escaping bubbles still affect the pressure difference by reducing the density of the wort. What seems to matter is the weight of the wort column between both ports and that will be less when there are bubbles present.

However, I care most about precise gravity towards the end of fermentation since this is when I want to know how far the beer has to go.

I like the idea. How precise can we get with this and reasonably priced parts?

For a bucket or Carboy we could use something like this. For simplicity the ends of the tubes could be left open while pointing down. The only problem with that would be that over time the air in the tubes may get replaced with CO2. I wonder how much that will make a difference, but we could figure that out. The tubes could be thin glass or even better SS.

and if we can fill the tubes with water, at least the distance between the two ports, we could measure the wort to water difference more precisely.

and if we can fill the tubes with water, at least the distance between the two ports, we could measure the wort to water difference more precisely

Click to expand...

Yeah, this was why I was asking AJ about why not just filling the tubes with the wort. If you couldn't use wort, I don't know that you'd want to use water, since the wort would diffuse/dissolve into the water. An alternative option would be to use something like food grade mineral oil which will stay separated from the wort.

But I think the escaping bubbles still affect the pressure difference by reducing the density of the wort. What seems to matter is the weight of the wort column between both ports and that will be less when there are bubbles present.

Click to expand...

It's not clear to me why bubbles in the wort would affect this measurement. I'd have to think about it a while to be sure, but it seems like if any bubbles form they'll displace the rest of the wort, increasing the absolute pressure at each port, but keeping the pressure difference between the two ports the same.

Bubbles lower the overall density of the beer. Just like a hydrometer sinks deeper into beer when bubbles are still rising in the hydrometer jar. They lower the absolute pressure at the lower port more than at the higher port since there are more bubbles above the lower port than there are above the higher port.

According to my math, the pressure difference between 0.25 m beer column and a 0.25 m water column is 0.001 inches H20 at a density of 1.001. It's ~0.5 in H20 at 1.050. So the 2001 from Dwyer may work: http://www.dwyer-inst.com/Product/Pressure/DifferentialPressure/Gages/Series2000/ModelChart

I think it is important to fill the other tube with water compared to wort. That way we can messure the difference between water and beer/wort which is much smaller than the difference between wort/beer and air.

Ad for the diaphragm, that separates water and wort/beer, a think rubber membrane should work since it doesn't not have to move much anyway. There should not be any noticeable compression of the air above the water.

Kai

The diagram shows how I conceive doing it with the conical. Your last diagram shows how it could be done in a carboy and how I suspect these guys may be doing it.

The math is on the diagram. The pressure difference depends on the differences in liquid level in the two lines, the density of the liquid and the density of whatever gas is in the lines. Dry air has density of about 0.00112 g/L which is about 0.1% of the density of water and moist air is less dense than that. CO2 is, of course, appreciably denser than air.

CO2 shouldn't move into the tubes nor should wort because there is no flow through the sensor. Yes, there would be a little diffusion and if temperature were crashed in the fermentor (your drawing) the gas in the tubes would contract and draw some in. It is l2 - l1 (see drawing) that is the most important parameter and it should be possible to calculate the change in (l2 - l1) from PV = nRT (i.e. air/CO2 is near enough to ideal). Note that in the cylindroconical the gas in the tubes would not be subject to the effects of cooling (unless the brewery got cold). In your drawing this effect could be minimized by making both tubes the same length (coiling one) so they held the same gas volume, both shrink by the same amount, draw in the same amount of liquid and leave l2 - l1 the same.

I think I see what you are saying about water but I'm not sure how you would get and keep it in there without these isolators which I can't find. I was thinking that we would just do calibration with water as we need to find (implicitly) the local value for g anyway.

I've looked briefly at the Ashcroft differential sensors which are reasonably priced at about $180 and at the Honeywells which are not (don't know but their similar altimeters are about $1K). Both seem to have an accuracy of about 0.1% of FS which is, of course, about 0.001 SG. The Honeywells are really nice because you just hook them up to the RS-232 port on your computer an you are in business.

Why can't you fill the tubes with wort? Because if you do the pressure across the meter will be 0. Look at the last equation on my sketch. The second term would be equal to the first and cancel it out: Pz = Pc.

As to the need for water: look at the first term - it is g times the density of the wort. The density of the wort is the specific gravity times the density of water. If you fill the system with water and adjust to read 1.000 then your adjustment has compensated for local variation in g and you are set. Plus, as clearly a computer is going to be involved here, you have the ability to compute the density of water at any temperature and thus to correct observed SG for temperature. The second term, if the tubes are filled with air, might even move your computation closer to the apparent specific gravity on which the ASBC tables are based but I'd need to check that thought out.

Why can't you fill the tubes with wort? Because if you do the pressure across the meter will be 0. Look at the last equation on my sketch. The second term would be equal to the first and cancel it out: Pz = Pc.

Click to expand...

Haha, yes, very obvious when you actually write out the equations. One of these days I'll learn and do that before I post.

I was doing some searching on Ebay regarding devices to measure differential pressure. I'm finding things like this: http://www.ebay.com/itm/Digital-Air...466&pid=100016&prg=1006&rk=2&sd=271121305903&

The price is acceptable but I'm not sure about the precision. The goal would be a precision as good or better than a standard hydrometer.

As A.J. mentioned, the pro of filling at least the long tube with water is that you'll get instant temp correction, compensation for g and your output will be a small pressure difference with no large offset. The con is that you need a flexible membrane to keep the water from mixing with the beer.

W/o filling the tube with water the device could be as simple as I have shown above. But you need to use a differential manometer that still gives you sufficient precision to detect the slight change in pressure while there is significant offset. That pressure delta should be 0.1% of the pressure offset. The pressure offset can be changed though the length difference of the tubes. For 25 cm the pressure offset (pressure for using water) would be 0.025 bar and every gravity point would increase it by 0.000025. The listed resolution for the cheap Ebay mannometer is 0.001 bar. Looks like that this is not all that suitable after all.

Another thought I had is, why not simply use a water filled U tube to measure the differential pressure? But with this design beer is expected to enter the longer tube, which has to be taken into account since this changes delta H. But I'm concerned about the ability to read the scale. With a delta H of 200 mm each change by one gravity point is only 0.2 mm. That's a bit difficult to read.

I think there is some misunderstanding about my thoughts on the use of water as the other fluid. I do not think the tubes should contain water. They should contain air (or hydrogen of helium). I updated the image in #13 to show the rest of the math. The bottom line shows how one would calculate the apparent specific gravity (what brewers use) from the differential pressure reading. Note that it depends directly on the differential pressure reading divided by the distance between wort levels in the tubes conveying pressure to the pressure sensor (carboy arrangement or conical). Thus if the pressure instrument is in error by 0.1% the SG reading will be in error by 0.1%. The nominal pressure difference depends, of course, on the delta L - the separation between the liquid levels. If that is about 27" the pressure difference will be about one psi (for water) and thus your instrument needs to be able to measure to an accuracy of 0.001 psi if you want accuracy of 0.1% in SG which, as SG is about 1 is 0.001 SG. For 13.5" separation you would have to measure to 0.0005 psi.

If you use a water filled manometer with 27" separation then, unsurprisingly enough, the height of the water column in the manometer will be about 27" and you have to be able to read that to 0.027" to achieve the 0.1% accuracy goal.

Now the other factor in apparent SG is 1 plus the ratios of the density of the gas in the tubes divided by the product of g, the density of water at the reference temperature and the separation. What I am suggesting is that this factor be set to the pressure difference when the setup is filled with water as water has apparent SG = 1 (to 3 decimal places). There is no need to measure or obtain a value for g and no need to know the density of what is in the tubes or indeed the density of water at a particular temperature.

In the conical case there is one potential problem with doing things this way if the tank is spunded and appreciable pressure applied. If, for example, 1 atmosphere CO2 is allowed to build in the head space the air in the tubes will be compressed to approximately twice its density. At atmospheric pressure the 1 + density ratio terms is approximately 1.001. With an additional atmosphere's pressure in the headspace that would increase to 1.002. That's really only 0.1% change so perhaps not a problem after all but the idea of using hydrogen or helium as the gas in the tubes was prompted by consideration of this.

A.J, my thinking behind filling the longer tube with water up to the port for the shorter tube is that the resulting pressure differential will only depend on the density difference between the beer and the water. So you should end up with an equation that should have the following structure

sg = 1 + A * delta l * delta P

I don't have the time to do the math right now. But the factor A can be calculated. My point is that you can measure delta P with a more sensitive manometer since it only has to deal with the pressure that comes from the density difference between beer and water and not the larger pressure difference that you would have if you put beer against air or some other gas.

This also assumes that there is no actual movement of air such that the water column in the long tube is pushed up a little.

Kai

I didn't mean to imply that I don't think using water will work - but rather that it isn't necessary IMO and therefore not, given the necessity to keep the fluids apart at the boundary, something I'd consider (though my mind is not closed on this by any means).

If you look at the pressure difference formula on the image in #13 it is

D_P = g*D_L*(dens_wort - dens_fluid)

where D_L is the level differnce and dens_fluid is the other fluid (air in my calculations). If the other fluid is water then

D_P = g*D_L*(Dens_wort - 1)

approximately and the developed pressure differential is much smaller. At first look you might say that's bad because the 'signal' is less but if you solve the D_P equation for dens_wort

dens_wort = D_P/(g*D_l*dens_water) + dens_fluid/dens_water

and take the partials with respect to pressure and the density of the other fluid you get

partial WRT pressure = 1/(g*D_l*dens_water)

partial WRT density =1/dens_water

it's clear that selection of a fluid other than air does not improve accuracy for the same absolute accuracy in pressure measurement.

Now if the absolute accuracy of the available pressure gauge improves as it's span gets smaller then using water could be a viable strategy. With a 27" separation and air the pressure differential will be about a pound (27" WC). To measure 0.001 SG you'd need accuracy of 0.027" WC. With water as the fluid you'd still need 0.027" accuracy but the pressure differential for SG 1.040 would only be 27*0.04/1.04 = 1.038 " WC. Now 0.027/1.038 is 2.6% which is a more relaxed accuracy requirement than 0.1% for sure. The problem is that whereas the accuracy seems to be a function of span at higher pressures it seems to become an absolute error at lower pressures. For example, the Honeywell PPT series has a 0.0375% FS typical error spec for 2psig and above but it doubles (on a percentage basis) to 0.07% percent for the 1 psig unit (which is the most sensitive in the line). There may be other instrumentation with better performance but given the cost of these Honeywell things I'm not interested.

Using a manometer you will need to be able to read 0.027 " (0.7 mm) whichever fluid you use but the height of the column will be less.

ajdelange said:

Now if the absolute accuracy of the available pressure gauge improves as it's span gets smaller then using water could be a viable strategy.

Click to expand...

That's the point I was trying to make. In addition to the fact that it is easier to read a 0.027'' change on a gage that goes from 0-0.25'' compared to a gage that goes from 0-30''.

But finding a diaphragm that could be used to hold the water in the pipe doesn't seem to be easy.

Kai

Check out this site
stm-instrument.com

Less accurate by a factor of 0.0015/0.000005 = 300 but less expensive by a factor of 22000/500= 44. Hmmm.

Really 0.000005, for home brewing, I've used the anton par densitometers with that accuracy its designed for the lab and its not easy to use. Why not compare the anton par hand held... 0.0015/0.001 = 1.5 but 2500/500= 5 times more expensive...lets compare apples to apples.

As it was my little joke (which apparently fell flat) I get to pick the model.

If you are having trouble with a DMA 5000 it is probably technique as they are simpler to use than a hydrometer (except for the cleaning). Unstable readings are usually caused by outgassing of bubbles (guess the vibration promotes that). Anton Paar recommends filling with a syringe and keeping pressure on it while the measurement is being made. That works.

What about having a glass hydrometer touching/pressing a pressure sensor then translate its output in something readable/processable?

An initial calibration(reset button or something) would be easy to do but i dont know how ambient pressure would affect the readings.

what are your thoughts on this

perhaps a digital refractometer coupled directly to a fermenter?