Any interest, or suggestions on, a bolt-on still automation kit?

Lloyd

@CothermanDistilling said: how about making a device that looks like PID, but is just a temperature monitor with a lookup table from temp to %alcohol.. you place it above the dephleg, and you don't even need a parrot....

That would be AWESOME! Any idea where to get the digital version of the lookup table?

punkin

You need to factor in altitiude as well, maybe even barometric pressure on the day?

grim

Problem is that almost everything of interest is taking place in a 4 degree range, and the curve flattens out significantly as you approach azetrope.

You'll need to correct for pressure, and have extremely accurate calibration.

Simple measuring gear will not provide the necessary precision, you are talking about measuring to .01.

Difference between 85 and 90 is about a degree F, difference between 90 and 95 is half a degree.

It's doable though.

CothermanDistilling

well, you do have a known point at which you are at 99.9% reflux, sounds possible to this engineer... actually, the fact that it only has to be super-accurate in a couple degree range makes it way easier than making it accurate from 0-100C!

grim

Would be a little bit easier to pull this off on a pot still, where your output ABV range is in the taller part of the curve. You could get reliable readings with less overall accuracy, a little bit of error wouldn't be a big deal.

Up towards azetrope, a 1 degree F/.5 degree c error is massive. It's a matter of you thinking you are running 95, but really, you are running somewhere around 85. You wouldn't miss that on a hydrometer, that's way past user error. Even worse than the accuracy problem, is you need repeatability, if every run the error is different, maybe +.5 degree today, spot on tomorrow, -.3 on Saturday, this would be worthless, you'd need to calibrate it every time.

So it would requires some very expensive kit to pull off. You need to use 4-wire RTDs to get this kind of accuracy, and they are expensive. Then you need a meter that can deal with 4-wire RTD, which will also be expensive because this tends to be specialized gear. The temp meters I posted previously have 0.005% accuracy, and support 4 wire, as well as ethernet connectivity to grab the measurement for post processing. But you are talking near $700 just for the meter, probably another $200 for a high precision RTD.

TheMechWarrior

This thread has been a great read. I'll be keeping an eye on it.

@grim, those prices wouldn't scare a commercial distiller ;)

Cheers,

Mech.

grim

Hmm, even ultra-precision RTD sensors might not be good enough. The best ones I can find do +-0.15 F, which means a total of .3 degrees error. Factoring in the error rates of the measuring equipment, this means that near azetrope, you'll only know that you are within a range of 90-95, this is at best, somewhere between 90 and 95. No 90, 91, 92, 93, 94, 95, just "somewhere" between 90 and 95. On the hydrometer this is a big fat 1/2 inch wide strip, very easy.

Little glass thing wins again.

Lloyd

Little glass thing always seems to win. No one can come up with a digital look up table. We need to solve this but I am not smart enough to do it alone.

I think the first step is the digital look up table. Any takers to help concur that?

One step at a time is how we get there and the digital look up table, I think, is the first step. So far no one that I know has done it.

CothermanDistilling

I disagree, I have seen people pulling 97 with glass bobber... which is impossible... you need temperature correction with density measurement, more room for error... the +-.15 accuracy goes to near zero if you calibrate it at azeotrope, the best your multi-plated still can do at 99% reflux after getting rid of fores... highly electronic accurate temperature measurement is nearly free when compared to highly accurate electronic measurement by any other means... no, it is not TTB approved, but neither is any hydrometer that fits in an SD parrot, and not without a calibrated probe by it's side...

grim

From over at HD: Re: Temperature & ABV Relationship Charts

V=1.00234870142762+0.00448539174680633(T+AF)^2/((T+AF)^2-76.8192824761981^2)+0.167549870972697(T+AF)^2/((T+AF)^2-107.95147196715^2)

T=Temp in °C
V=%ABV of vapour
A=Altitude in metres
F=fudge factor (%0.34 for water or %0.31 for azeotrope, depends which end of the scale to you want more accurate) Just leave at 0m for no effect.

grim

Using that formula, at 1 meter altitude, here is a rough chart:

95C - 43.99
94C - 49.03 (Aprox. 0.504abv per .1 degree rise) ** REALLY TOUGH HERE
93C - 53.40
92C - 57.24
91C - 60.64
90C - 63.69
89C - 66.43
88C - 68.92
87C - 71.22
86C - 73.34
85C - 75.34 
84C - 77.26 
83C - 79.13 (Approx 0.187 per .1 degree rise) **REALLY EASY HERE
82C - 81.04
81C - 83.10
80C - 85.57
79C - 89.14
78.1C - 95.59 (Approx. 0.645abv per .1 degree rise) ** REALLY TOUGH HERE

grim

+0.15 error at azeotrope will translate into roughly 1.935abv of total error, and I have doubts that most of us can get to those levels of accuracy.

So you could be pulling azeotrope with the meter telling you anywhere between 94.66% and 96.60% (almost as bad as pulling 97% with the bobber). Like I said, this is approaching theoretical best case. A more realistic +-0.30 error, the meter will read anywhere from 93.69% to 97.56% (pulling way more than 97%).

Look at the sweet spot of the curve around 75%abv to 83%abv, that's where things are going to be much easier.

At 83C, +-0.15 of error is only 0.561abv total, 4 times less than at the other end of the curve, way more accurate.

We still haven't even added the error associated with atmospheric pressure.

Would be much more appropriate for a whiskey distiller than a vodka distiller, and even then a wide body parrot with a gauging hydrometer would stomp all over it in terms of accuracy.

grim

Ok, so I'll throw in $150 for this experiment, just ordered a type PL ultra-precision 1/10DIN gas measurement RTD from Omega. Anything better than this would cost closer to $1000 would probably break if you looked at it funny.

Ultra Precise RTD Sensors for Industrial Applications @ OMEGA

I already have the 4-wire precision meters.

Lloyd

This is great progress toward the goal of truly digital %ABV.

Manofconstantsorrow

Ok so back to an old Idea;

if it's so hard to beat the little glass thingy why not incorporate a bar code on the scale of the hydrometer and position the reader at parrot eye level?

Laser scanners (hell most smart phones have them) are cheap and Lloyd already has his own hydrometers being made.

CothermanDistilling

I agree, but @Lloyd has been asked this before.

Manofconstantsorrow

Ok, someone in NZ needs to contact and befriend these guys;

IKEGPS CORPORATE HEADQUARTER Level One, 42 Adelaide Rd Wellington 6021, New Zealand office: +64 4 382 8064

They make the spike which could read the height of the alcometer in a parrot. So we need someone to get one of these guy's hooked on distilling and subliminally let them think it's their idea to provide a cheap measuring device for ABV, these guy's are hackers at heart so they might just want to jump on the band wagon.

See here. I should also tell you they also used kick start to raise money.

CothermanDistilling

so we have someone like them make a little widget that clamps or screws onto the parrot, and can look at the miniscus of the liquid on the parrot as compared to the 100%abv line (needing no special alcometer), then this little widget talks to a black box (via USB, 4-20mA, etc) and that black-box displays ABV by using the formula, maybe also has temperature correction with a thermocouple/RTD (maybe a contact thermocouple built into the mounting clip).... this black box could also output i2c, 4-20ma, ethernet, wifi, bluetooth, or whatever to go to another controller or display...

chew on that for a while....

Oh, make the open alcohol area sealed in glass for my fire marshal...

Manofconstantsorrow

This fits the bill perfectly; CD22 Laser Measurement Sensor @ Optex Ramco Too bad their 480 USD even for eval.

CothermanDistilling

nice! I went googling, and holy crap, the Anton-Paar site has a handheld unit for $1286.11 did I say holy crap?, these were $3500 a couple years ago! I know what I want for christmas!

Anton Paar Snap 40

grim

Lots of money to pay for an Anton Parr that the TTB has not approved for proofing.

CothermanDistilling

I see it as $1286 is better than $3500, and you can do all your proofings with it except the final one and you break out the $300 NIST calibrated hydrometer...

I noticed the 35 also has output ability

from the rep:

Hello Michael

Thanks for your email. The difference is that with the Snap 40 you can only measure alcohol by volume. No other methods.

With the Dma 35 you can measure ABV but also Brix, Density, Specific Gravity among others.

In both cases you will need to make a distillation before you can measure the ABV.

If you have any other question please let me know.

Thank you.

Sent from my iPhone

Juan Gonzalez Anton Paar USA Caribbean and Florida office (787)590-7094

grim

Yeah, the DMA 35 isn't approved either, it doesn't have the requisite level of accuracy. You need to go all the way to the top.

Approval of Alternative Devices for Determination of Proof of Distilled Spirits For Tax Determination Purposes; Guidance and Performance Standards - July 21, 2014 (PDF) @ TTB

If you are diligent, you can actually find one of the older approved models on the used market occasionally, the re-calibration isn't cheap, but given the price of the DMA 5000, it's a bargain. I've been prowling for about 3 months now, and nothing has popped up.

Also realize that the handhelds are incredibly fragile, just like the glass ones, if you drop it or mistreat it, it'll break, and from what I've heard, they aren't repairable. Sure, it won't roll, but this thing needs to be treated like a precision instrument and probably kept cased nearly all the time.

You can get high precision hydrometers (the wide body variety) from Novatech for about $45 each. They don't carry a NIST calibration certificate but for intermediate stage proofing this is just fine. It's a little bit of a joke that that they require a calibration certificate, since they aren't capable of being calibrated at all, generally if it made it out of the factory, it'll meet calibration, because if it didn't it'd go right into the trash as a defective product.

$45 vs $1286? The $1286 doesn't buy you anything more accurate.

Also, you can not use the SNAP or DMA35 for continuous monitoring. You've got to click to take a sample, wait for the reading, and than clean it between samples (according to Anton Parr anyway). You can find non-approved U-tube densiometers that are made for continuous monitoring. They are bench top units with an input and output port.