No-contact Device For Measuring Fermentation Progress

[chovain]

so u really want a digital gas flow meter for $50 & some software ?

Kind of, but my understanding is that gas flow meters are designed to not affect the system they're measuring, so use ultrasonics, thermal variation and other passive tricks. That makes it hard for them to accurately measure low rates, which make up a significant proportion of fermentation.

You'd need one hell of an expensive flow meter to keep a good tally of the CO2 released during fermentation, but we have the advantage that we know we can get away with letting the pressure build up, giving us a more accurate measurement.

Edit: Correction. Flow meters seem to directly measure kinetic energy in a pipe, but that's still not much use in this application.

 

[redbeard]

so would a 30kg digital postal scale with 1g accuracy be suitable for measuing the changing weight ? (cheap on ebay)

 

[chovain]

so would a 30kg digital postal scale with 1g accuracy be suitable for measuing the changing weight ? (cheap on ebay)

Unfortunately not. By the way, that's 1g precision, not accuracy (the eBay ads are wrong). The scale will vary by many grams over the time it takes to ferment. Having them in a temperature controlled environment will help a little, but they're going to vary too much.

 

[barry2]

Very interesting discussion.
I would settle for something that would count the bubbles through my three piece airlock.I have an ale in my fermenter at present with US-05 yeast.I count the bubbles over one minute about the same time each morning when I am changing the ice bottles in my cooling system..Since counting from the day following putting the brew in the fermenter the readings have been 15,30,6,4,3,3,2,1,1,1.No break in the krausen as yet.At this stage it would be interesting to have a count of the bubbles over a 30 minute period over the last three days.I haven't taken a hydrometer reading since the one for OG.

 

[Hutch]

Unfortunately not. By the way, that's 1g precision, not accuracy (the eBay ads are wrong). The scale will vary by many grams over the time it takes to ferment. Having them in a temperature controlled environment will help a little, but they're going to vary too much.

Hey Mark,
Interesting discussion. It would be a simple task to learn just how much variation one of those cheap scales gives over time -
measure 20Ltrs of water in a sealed cube on the scale over successive days, with the scales permanently ON, in a controlled temperature environment. The absolute error on those scales might be a bit off, but the relative error (as the fermenter gets lighter) might not be so bad.
Certainly worth a bit of experimenting.

 

[chovain]

I would settle for something that would count the bubbles through my three piece airlock.

I reckon I could probably take a crack at that (and bound to be cheaper than my idea). We can call it the "mesa" method, in true AHB-style. The biggest problem is that it would be hard to detect stuck fermentations without taking a hydro reading.

The trick is building a circuit that triggers on the air-lock vibrations, but nothing else. I'm going to go away and have a think about it.

 

[jendres]

I reckon I could probably take a crack at that (and bound to be cheaper than my idea). We can call it the "mesa" method, in true AHB-style. The biggest problem is that it would be hard to detect stuck fermentations without taking a hydro reading.

The trick is building a circuit that triggers on the air-lock vibrations, but nothing else. I'm going to go away and have a think about it.

how about an optical sensor across the lowest part of the airlock? You would need something in the water to block the light/IR though.

Maybe a small styrofoam ball? The problem would be if it gets sucked back into the fermenter during cooling/emptying.

 

[chovain]

I am sure you meant 2 Moles of Alcohol and 2 Moles of CO2, or the beer will be very hot indeed.

I don't get the heat reference: "mol" is the standard abbreviation for Mole (although I did incorrectly pluralise it in a couple of spots). What else can "mol" mean?

How about a sight tube, fit a pressure transducer at the bottom

From P=Ro.g.h

We know g ~9.81
The hight we can measure
Rearrange to give Ro or the density.

Sorry, I missed this post last night, but I like the idea!

Assuming I understand what you're saying, you'd need a waterproof sensor. I suppose you could get away with a non-water-proof one if you put it in an airtight chassis with a flexible membrane to transmit the pressure from the fluid.

If you want to be able to gauge SG to within 2 points (to be as accurate as a hydrometer), and you assume ρ will be pretty close to 1 for finished beers, we can say you're allowed a margin of error of +/- 0.1% (any more than that, and you won't be as accurate as a hydrometer).

With a marked sight glass, you expect h to be around about 0.4, I suppose, with mm precision (0.001). Error in the height measurement is going to be around +/- 0.125% - we've already blown our error budget.

You could use an ultrasonic transducer to get a more accurate height measurement than a marked sight-glass. A transducer will be able to measure to within its wavelength without too much trouble (from above, so the changes in the wort don't affect the measurement). At 40kHz, you can measure the liquid level to within 0.025mm, giving us an introduced error of +/- 0.003% - well within the error budget. Assuming the transducer is being run from a crystal (via MCU), the error in the waveform should be negligible.

Now for the sensor: We're looking at pressures around the 38-45hPa above atmospheric. So we should be considering a range in around 1000-1100 hPa (I can't for the life of me remember how much atmospheric pressure changes day-to-day, but that range should be safe near sea level). The device I was looking at, while not water proof, will measure an accuracy of 0.5hPa, and meets the required range. 0.5hPa at 1000hPa pressure gives you an error margin of +/-0.025%. Even at 8000m, where atmospheric pressure is somewhere around the 300hPa mark (the lower limit of the HC03D), you're error is still only +/-0.08%.

So: at sea level, with a transducer, the "MHBometer" D) is accurate to about half a point of SG. For those brewing beer on top of everest, they'll have to settle for similar accuracy to their hydrometer.

Anyone see any problems with this my logic here?

 

[chovain]

how about an optical sensor across the lowest part of the airlock? You would need something in the water to block the light/IR though.

That's going to be tricky, especially in a 3-piece. The problem with water is that it's really hard to see. In a 3 piece, you don't know where the bubble is going to come out. Maybe a little light-weight flap on top of the airlock (or replacing the cap on a 3-piece) would do the trick: You might be able to have a microswitch or something to detect it jumping.

 

[jendres]

That's going to be tricky, especially in a 3-piece. The problem with water is that it's really hard to see. In a 3 piece, you don't know where the bubble is going to come out. Maybe a little light-weight flap on top of the airlock (or replacing the cap on a 3-piece) would do the trick: You might be able to have a microswitch or something to detect it jumping.

I was thinking about an S-type airlock. And yes water is hard to see, that is why I suggested a styrofoam ball.

The flap might work with a IR sensor as well.

 

[chovain]

I was thinking about an S-type airlock. And yes water is hard to see, that is why I suggested a styrofoam ball.

Ahh, that's a point. I was still thinking that the sensor would be in the bottom bend. If the ball is bigger than the diameter of the smaller diameter of the airlock, you might have just invented a one way air lock, too!

 

[drsmurto]

Synthetic chemist hat on.

I did a similar experiment to measure CO evolution during a synthesis where 1 CO ligand was displaced from a molecule that contained 6 CO ligands. To do this i did what someone has mentioned and collected the evolved CO and used it to displace a liquid in a measuring cylinder. The height the liquid is displaced is proportional to the amount of CO evolved and thus once the desired height was obtained i stopped the chemical reaction.

Since many people use blow off tubes this is a very simple experiment.

 

[jendres]

Ahh, that's a point. I was still thinking that the sensor would be in the bottom bend. If the ball is bigger than the diameter of the smaller diameter of the airlock, you might have just invented a one way air lock, too!

Yeah, i've thought about that, but it would probably move too quickly and randomly to be easily measured, plus the seal might not be so good, in which case the ball may not move at all. (Same with the flap idea, it might just crack open a little and let out a fairly continuous stream of CO2 rather than letting out puffs.

I was thinking about using water to get the seal and a float (styrofoam ball) to measure the water level. As the water and float gets pushed up into the second chamber, it would pass the sensor allowing you to measure it. Without the flaot, you would have to have some pretty precise tolerances to determine whether water is there or not.

You might have some success with measuring the varying refraction shining a small IR laser through the tube. This would be refracted to different angles if it passes through water vs simply air, but that seems a bit expensive and difficult to calibrate.

I've attached a couple of quick pics to show what I mean. One with the water pressure low, the next with the airlock about to release the bubble.

 

[mesa]

Anyone see any problems with this my logic here?

Hi Mark,

I think you may have messed up the wavelength for the 40kHz signal (seem to have set the speed of sound to 1m/s rather than 340m/s) Admittedly you should be able to measure to sub-wavelength accuracy, but I don't think you'll get that accurate without some expensive components. A laser range finder might be more appropriate, but thats also expensive. I also wouldn't want to think about how the meniscus on the beer changes as the sugar content changes or the sight tube gets dirty.

Come to think of it, maybe thats another option, measure how well the beer climbs a capillary tube? (I've never worked with capillary tubes so I wouldn't know how to keep them clean)

As for measuring bubbles through the air lock, I was thinking of just using a microphone inside a fermentation fridge (so it should be relatively quite to begin). If you are using a computer for data logging you get to do a frequency filter for the sound a bubble makes, then gate the whole apparatus so that more than 2 bubbles in half a second will not register. I'm sure there are many tricks to filter out background noise as well.

If you want to do the IR trick, the first thing I'd try would be IR LED and sensor on either side of the bottom bend in a S bend air lock, so that just before a bubble is released there is only air in the way, as the bubble is released water fills the gap between the sensor and LED and then returns to air gap over a short period of time. Then look for variations in the signal from the sensor filtered for the correct sort of frequency (around 1-5Hz I imagine). If the water is still too hard to see, I'd probably play with various additives rather than the styrofoam ball. I imagine either idophor/betadine or some sort of food dye is largely IR opaque.

 

[mesa]

Since many people use blow off tubes this is a very simple experiment.

Only problem is the estimated 200L (at STP) to capture and measure. You'd have to arrange to regularly bleed off a known quantity (or perhaps displace a VERY heavy liquid ;-)) Then again if you use the valves that Mark proposed right at the start of this thread, perhaps thats not a problem.

 

[chovain]

I think you may have messed up the wavelength for the 40kHz signal.

Ahh - I knew I'd mess up something somewhere. . I might have to do some experiments to see if I can work out some trickery on a high speed MCU. It should be fairly easy to get ~1mm precision using a comparator to convert the incoming signal to a square wave, then xor with the output generator. With a bit of tuning and filtering, I should be able to get an even more accurate measurement. I'd need to test how changes in the viscosity of the reflective media affected the shape of the reflected wave though.

I wonder if it'd be possible to surround the transducer with an acoustically reflective material, and let the signal bounce back and forth a few times to get a better measure. You'd need to filter out other strong reflections (like the one from the bottom of the fermenter), but I expect the signal will still be quite good after a few bounces if it's driven with full power.

Perhaps combining two (or more) non-harmonic frequencies would help too. The device is starting to get a bit expensive with too many transducers though.

Alternatively, Farnell has 235kHz transducers: They'll measure to 1.4mm without phase detection, but they're $168 :blink:.

The other problem of course is ensuring the sample in the tube is representative of the rest of the wort. I suspect the sensor would actually need to be in the bottom of the fermenter itself (perhaps with a plate mounted above it to protect it from trub). It wouldn't be too big a problem as long as it were sanitisable.

 

[mesa]

Just stumbed on this this post and thought it was relevant to this old thread. I hadn't even considered the change in speed of sound would be anywhere near large enough to enable measuring anything. Maybe you could just place transducers on either side of the fermenter and measure the resonant frequency (not that I'm sure how to design an analogue circuit that will lock to the resonant frequency but it can't be all that hard ;-))

--
Chris.

 

[chovain]

Just stumbed on this this post and thought it was relevant to this old thread. I hadn't even considered the change in speed of sound would be anywhere near large enough to enable measuring anything. Maybe you could just place transducers on either side of the fermenter and measure the resonant frequency (not that I'm sure how to design an analogue circuit that will lock to the resonant frequency but it can't be all that hard ;-))

--
Chris.

I can't seem to find the relevant post in there. Any hints? There is a message in there about measuring CO2 with a temperature sensor, which is interesting. We have used a similar technique where I work for detecting extremely small liquid flows (I mean *really* small - as in less than a nanolitre per second.) You can do it even simpler than they suggest in that message, too: If you take two thermistors, one in the airlock, the other in the fermentor headspace, and put the highest current across them that the specs allow, the airlock thermistor will be cooler than the head-space one. It unfortunately takes a bit of empirical measurement and modelling to convert that to an accurate flow rate, but it can certainly be done.

Based on the words you've used, I assume the point is to wire a receiver to a transmitter through a decent amplification circuit, then send a signal through the transmitter. The signal will then be fed back, providing a new, low frequency signal. By measuring the frequency of that lower frequency signal, you can calculate the speed of sound through the wort, allowing you to calculate the density. I see a problem with this approach: My fermenters are 33cm wide. At standard speed of sound, propagation will take around a millisecond, giving a frequency well within audible range . Not sure if the power needed to get the transducers to trigger would be large enough to make the signal audible, but I know I wouldn't want my beer beeping in the middle of the night .

Then again, you might be talking about something completely different.

 

[mesa]

I can't seem to find the relevant post in there. Any hints?

How Odd, when I tested from the computer I posted from it was a direct link to the post (Subject Quantative Clinitest) but not from this computer. Oh well there's your hint ;-)

As for the resonant frequency of the fermenter, I was thinking of a scenario where a standing wave was setup like ripples on a pond (l=0 solution for a standing wave on a drum head? or have a completely forgotten the notation for waves on a drum head?). I hadn't even considered it coupling strongly to the air but I suppose it would. I just thought it was funny that in all this discussion of measuring weights and pressure to determine density by the barometer method we hadn't even though of the effect of density on the speed of sound.

 

[MHB]

How about one of these

View attachment 18038

Refractive Index Sensor (Link)

MHB