StillDragon® Community Forum


Be part of our community & join our international next generation forum now!

In this Discussion

Electric Parrot Measurements For Data Logging

This will take a while as still waiting on parts but....This gets talked about a lot and I am not so knowledgeable in it all but would like to give it a whirl anyway. As one idea from crazy idea thread which is magnet inside the alcometer. Maybe together we can all get something that kinda works at least.

Basic idea is hall sensors to measure distance of the magnet. There is already an expensive electric hydrometer on market that used this approach so at least it is possible. Just different scale for software to calculate as not linear. The question will be accuracy. If I go into this under the assumption of data logging runs rather than pure accuracy I think its an experiment worth doing. Unless someone here says Im wasting time?

Example someone made gausemeter

experiment 1
is to have a magnet on my fake alcometer stick inside a metal tube. Position a hall sensor on outside and measure how accurate the movement of my magnet is and range. This part will be a mind screw for absolute beginner like me. We want vertical accuracy but magnet strength will shift if it moves horizontally in the parrot as well maybe? Or with poles facing up down will this not matter much. will the sensor work in that orientation etc
For this I just need to order some cheap hall sensors as I dont think I have an analog one.

I want this to output a reading to lcd. So have to learn using hall sensors and lcd. Plenty experiments to copy the code for this on the net.

experiment 2
play with multiple hall sensors and accuracy distance increases.

experiment 3
add a scale for non linear alcohol conversion
To do this need to know plot points along the curve if anyone here has that data.

experiment 4
add temperature probe and make adjustments to final figure based off liquid temp

Experiment 5
Usb data logging to pc over time.

experiment 6
add wifi

experiment 7
options perhaps like adding a button that will cycle what data is displayed on screen avg temp avg % over time. Is there anything that people may need during a run displayed? If two line lcd we can have temp and adjusted % displayed already.



  • edited August 2014

    We have a discussion around here somewhere, it been put in the "too hard basket" mostly.

    I reckon you have to measure the distance from the top of the alc meter to a setpoint above it. As the meter moves up and down the actual distance changes, but will need to fairly precise to work. ie add a flat disc on top of the alc meter to allow a decent surface area to be able to measure, and a stand or attachment for the thing the "reads" the distance. I'd look at mm or less type scale, ultrasonic or optical measuring device.

    I'll look for the thread for you, have a read here


  • If you are going down the path of adding a magnet for experiment 1, and you had access to a glassblower, replace the bottom weight in the alcometer with a magnet.

    Alternatively position a magnetic stirrer bead somewhere in the centre of the alcometer, and remove the corresponding weight (lead/steel shot or whatever is used) from the bottom of the alcometer and seal it. Just some thoughts, but this option could be expensive. Also add a silicone base to the alcometer so that I don't keep breaking them!!!

  • edited August 2014

    This is more for fun but I would assume if we actually got something worth producing someone would figure out how to get a magnet in place. Problem with distance measurement is non are accurate that I have seen.

  • Anytime you'd like to experiment with a prototype alcoholmeter just let me know but please get it right the first time because the minimum order is 1000 pieces.

  • I'd send you a prototype of what I have built if things were to progress that far let you decide for yourself if it were worth doing. While I'm waiting on parts im looking into optical solutions as well that can read what's on side of alcoholmeter. A thought I had wondered was how accurate colour sensors are if a rainbow stripe was on meter could it tell how far along the rainbow it was. Or if barcode style system is possible where it can calculate how far along pattern it is reading

  • I have read a bunch of threads about this around the web and most of them seem to be trying to measure the distance between the alcho and a set given point. Im wondering what would happen if you used a presser censor and completely submerged you alcho meter. Then as your alcohol content rises the presser on the censor would increase.
    The idea of trying to have a alcho meter free floating in a running liquid leves you with to much room to play with if you ask me. At least with it submerged its locked in so the running liquid will have minimal affect on it. This just leaving you with needing to figer out code or algorithm for it. I think both ways are going to be equally as hard to put together.

  • It's totally doable Colour sensors that can easily be calibrated to any existing alcometer. Checkout 'cubestormer' for inspiration. :)

  • why not the way the industry does it, with a vibrating u-tube 4-20mA sensor...

  • edited August 2014

    Suspect for that a given still in a given configuration, if you knew the wash temperature, the vapor temperature post dephlegmator, and the vapor temp at the base of the column, you could approximate the abv output somewhat accurately using a set of look-up and correction tables. Am I wrong here? You have essentially a fixed system with numerous known variables, that follow a fairly well understood process. You'd also be able to calculate the approximate theoretical plates of rectification being provided by the column, and estimate the remaining alcohol in the wash. If you wanted to get even more accurate, you could take into account the current atmospheric pressure, and the pressure in the boiler (column backpressure).

  • the most single important thing is you put beer or low wine in and spirit comes out :))

  • Damn, knew I was doing something wrong...

  • @grim said: ... you could approximate the abv output somewhat accurately using a set of look-up and correction tables.

    I've seen the graph but has anyone come across a digital version of that? I'd think that the vapor temp above the dephlegmator would be the most critical input information and the others would help to fine tune the calculation.
    This line of thought could completely eliminate the need for a parrot.

  • If it were just one temp then an arduino could make conversions based off a plotted curve. However programing to take into account other temps what calculations need to be made?

  • if the vapor temp is between 170 and 212F in the 180 at the top of the still, you have a very accurate measurement of the percent alcohol with that one temp, don't you?

  • Temperature would be the easiest. If you need a automatic parrot, a 10" or so tall column with top and bottom taps going to a Dwyer MS2-W101 ∆P transmitter or something similar would cover the 60-95% abv range easily but you'd still need a temperature input.

  • edited August 2014

    It's more complex than just one temp. I mean, if you sat down and data logged a few runs, you would be able to create a table of temperature to ABV. I suspect if you were able to keep the reflux ratio the same, heat input the same, and wash abv the same, you would be able to nail it.

    How about this one, even easier.

    Weigh a known volume of distillate.

    Create a "parrot" that is not connected to the condenser, but sits alone on a high precision load cell that has 1mg sensitivity.

    Make it overflow in a manner that leaves it physically disconnected from the receiving vessel.

    Zero the weight of the "parrot" so that you only measure the distillate weight.

    You would need to adjust for temperature, but that's just math.

    At 60 degrees F, 1 gallon volume weighs:

    200 Proof - 6.60970 pounds

    100 Proof - 7.78007 pounds

    50 Proof - 8.08606 Pounds

    1 Proof - 8.32198 Pounds

  • edited August 2014

    So converting into a smaller volume, say 200ml?

    100% - 200 ml - 158.403g

    90% - 200ml - 166.381g

    50% - 200 ml - 186.452g

    25% - 200 ml - 193.785g

    .5% - 200 ml - 199.439g

    0% - 200ml - 200.000g

    Someone check my math, did this fast.

  • edited August 2014

    Or step down to something smaller like this load cell, which would allow you to use a smaller 100ml weighing cup

    100g max weight with .05g precision.

    I think you'd be able to measure to a precision of .1-.2% abv.

    90.5% - 200ml - 166.050g 90% - 200ml - 166.381g Difference is 0.331g

    This would give you damn good accuracy and repeatability, and doesn't need anything too hokey or non-standard..

    You could probably do this pretty easily in an Arduino or similar. It would still probably cost $250 plus fabrication to put together. Suspect that the level of custom fabrication to do this would be very small compared to some of the other approaches.

  • edited August 2014

    Something like this:


    Only example I could find of what a good receiving vessel would look like:

    Stainless Steel Straight 200ml jigger, Measuring Cup, Cup @ AliExpress

    I swear I don't spent my time on bar sites researching jiggers.

    521 x 497 - 12K
  • That looks like a fun thing to try and usefull in kitchen if it doesnt work

  • edited August 2014

    Ok this configuration would probably be more easily utilized. Biggest issue is you have no parrot drain. Would still need to figure out how to get a probe type temperature sensor into the liquid from above, with no contact with the cup.

    Looking through load cells, Vishay 1004 looks ideal, about $75 usd. The included photo is just a random load cell, not the Vishay. With that Vishay I suspect you could easily get .1 level precision with solid repeatability. Assuming there is some way to easily calibrate the volume, and zero the balance each run, I suspect this might be more accurate than a simple hydrometer, especially considering you'd have high precision automatic temperature correction.


    529 x 508 - 24K
  • Ok here is version 3 - this one is the winner.


    561 x 506 - 22K
  • A good chunk of expense goes into processing by the look of it. The arduino is only 10bit so needs external hardware for that level of precision.

  • edited August 2014

    Also we have been discusing this on a NZ distilling forum and a member suggested that we cut an angled slit in card that floats then use light. So if its pointy end is up then less light will be recorded passing through at high alc volume. Downside is overcoming ambient light so would be enclosed

  • edited August 2014

    Ah damn I didn't realize it was so coarse, I just assumed at least 16 bit. I've got an old Labjack pro floating around here that I've used on numerous projects before that I know has 16 bit precision. Also remember seeing something from NI a year or so back about a new USB daq they had for under $200. Going to require at least some basic programming skills, regardless of the approach. There isn't any way to do this that isn't going to require roughly the same level of instrumentation. Doppler, temperatures and pressures, IR distances, etc etc. You'd probably want to go the DAQ route anyway, and just be able to data log everything.

    This is a way past the typical DIY projects you normally see around here, what's the option? An Anton Parr for 3 grand? Or, well, I guess stick with reading the side of the floaty thing.

  • It always seems to come back to reading the side of the $5 floaty thing.

  • edited September 2014

    USB Load Cell 5kg With PC Software @ eBay

    So we are after this but with our own software.

Sign In or Register to comment.