Has anyone modded bathroom scales to gauge beer left in kegs?

Homebrew Talk - Beer, Wine, Mead, & Cider Brewing Discussion Forum

Help Support Homebrew Talk - Beer, Wine, Mead, & Cider Brewing Discussion Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

BigFloppy

Well-Known Member
Joined
Aug 28, 2013
Messages
885
Reaction score
117
Location
Marietta
I've seen a couple of hacks out there for getting an Arduino to intercept measured weights, I'm figuring on putting together a front end that basically tells me easily approximately how much beer I have left in the kegs... thoughts anyone?
 
There have been a myriad of methods - attempted and occasionally successful - on tracking keg volumes. I'm thinking I've seen most of those documented in the Kegging section (though I could be wrong). There have been schemes based on bubble meters, scales, floats of all kinds, impeller/rev counters, sonic transducers, ad infinitum.

Aside from an epic KegBot or two (truly awesome level of overkill-kill-kill dispensing system) I've yet to read of anything that I thought had the legs to work beyond the one-off level.

Anyway, using scales - and there's a crap ton of cheap digital scales that might be "interfaceable" with a controller - the relatively hostile environment at the bottom of the typical keezer may pose a life-span limiting factor. We've had some epically humid weeks here where I've had to wet-vac my keezer floor...

Cheers!
 
There have been a myriad of methods - attempted and occasionally successful - on tracking keg volumes. I'm thinking I've seen most of those documented in the Kegging section (though I could be wrong). There have been schemes based on bubble meters, scales, floats of all kinds, impeller/rev counters, sonic transducers, ad infinitum.

Aside from an epic KegBot or two (truly awesome level of overkill-kill-kill dispensing system) I've yet to read of anything that I thought had the legs to work beyond the one-off level.

Anyway, using scales - and there's a crap ton of cheap digital scales that might be "interfaceable" with a controller - the relatively hostile environment at the bottom of the typical keezer may pose a life-span limiting factor. We've had some epically humid weeks here where I've had to wet-vac my keezer floor...

Cheers!

Kegbots are def overkill for most home setups but damn are they sweet.

Also, off topic but day_trippr have u considered adding one of these?
http://www.amazon.com/dp/B00ANX1PAI/?tag=skimlinks_replacement-20

Works wonders on kegerators and ferm chambers with regards to condensation.
 
Last edited by a moderator:
Actually to alleviate any moisture issues, was thinking of coating any electronics in "NeverWet"
 
[...]Also, off topic but day_trippr have u considered adding one of these? Works wonders on kegerators and ferm chambers with regards to condensation.

I've had an Eva Dry in my keezer for the last year, with an 80mm fan driving air right into it. It'll hold five scant ounces of water before it needs a dry-cycle, but during the epic humidity we had the last few weeks it was woefully overwhelmed...


Cheers!
 
I had a few thoughts, but will probably never get around to building anything. My easy, mechanical idea was to mount the keg on a base made from plastic/MDF with springs between a top and bottom piece. Then machine a simple pair of blocks to mount on the front with a pivoted needle indicator. A bit of work, but simple in concept and requires no electricity. As the keg emptied it would raise the top part of the base.

Another thought would be to take something like a cheap junk yard Throttle Position Sensor and interface with and arduino. Same basic mechanics for the keg base, but with a digital component. It just converts the voltage variation from the TPS into a readout.

Something in between would be a battery-powered LED display that uses some type of sensor (switches, optical, light detecting, etc.) in the base to trigger a certain LED in the stack. As the keg gets lighter, a new LED will light up. Less accurate than some might like, but the LEDs could be easily mounted in the front of the kegerator and the batteries could last a long time if a simple timer circuit turned it on and then turned it off after like 10 seconds.

Another option is to use a simple weight sensing device, but I'm not sure how they look physically, so I can't really understand how to put that together. This might be something I'd work on after I get my brew rig built, which is going to be some time after I retire, at this rate!
 
Homercidal said:
I had a few thoughts, but will probably never get around to building anything. My easy, mechanical idea was to mount the keg on a base made from plastic/MDF with springs between a top and bottom piece. Then machine a simple pair of blocks to mount on the front with a pivoted needle indicator. A bit of work, but simple in concept and requires no electricity. As the keg emptied it would raise the top part of the base.

Sounds somewhat similar to what my grill does with the propane tank. Basically the tank hangs from a spring that provides visual output at to how heavy the tank is and thus how much propane is left.

Otherwise, for a digital approach, do some searching on strain gauges and wheatstone bridges; I think that's the most viable path.
 
I am working on a system like this for gas cylinders, though it will work for kegs as well. I got a load cell, making a base for the cylinder to sit on and a frame for the weighing unit. Put that through an instrumentation amplifier, then into a LM3914 with 10 LEDs. One can stack 2 LM3914s for 20 LEDs, but to me 10 is enough. "calibrate" the offset of an empty cylinder to no LED, set the full cylinder to all 10 LEDs. Not very accurate, as each cylinder will have a different empty weight, but should give good enough indication of when to start panicking whether there is a spare handy or not.
 
The biggest issue is sensor drift over time. Load cells are notorious for this, hence the need to zero the scale every time you use it. The question is does that drift cost you 1 beer resolution or 10 beer resolution.

Same can be said for old throttle position sensors. The PCM relearned zero on every key up/decel. The drift was generally only a count or two, however, if using it to measure minimal deflection, it could throw off the entire system.

Of the methods I've seen here I think the flow meters have the highest chance of success. I'd be worried about trapping crud in there, keeping them clean, etc....

Time the tap was open and an inferred flow rate would be a fairly simple assuming you are fairly consistent with you kegging pressure and line lengths.
 
Hopwza' 'bout........An oversized fish scale, (inline spring type), weigh your empties for a ball park figure, hang a full keg, and mark where "full" is..........

I know some of you folks love the tech aspect of it, ( and me too, but not a lot!), but if ya' ain't worried about "portion control" or the price per beer, can all that flow meter, load cell, see through keg crap, and look at the fish scale! :D
 
I think the best bet because of Sensor drift as MI_Troll mentioned is instead of measuring weight would be putting in a flow sensor in line on each tap.

You should know more or less exactly how much beer is going into your kegs from the fermenter, so simply keeping track of how much has been pulled should tell you fairly closely when your kegs about to go.

Something like this could be wired up to an arduino and require very minimal programming to get working. It says it supports up to 2 MPa which a quick google says is about 270 PSI...so well beyond any serving pressure.

http://www.adafruit.com/products/828?gclid=CPDXyO6btbkCFYdxQgodjloAyg

To use a force gauge to measure the weight you will constantly need to recalibrate it, and to get one with proper accuracy you would be spending a pretty penny.
 
In answer to the original question, no, but I have wondered about it. I had to fix our bathroom scale once and am waiting for it to break again so I can have the excuse to get a new one and have the old one to play with as a keg gauge.

I was thinking about using a small inner tube hooked to a pressure gauge with a plate on top to set the keg on. With about 10 square inches of contact with the inner tube and a 50 pound full keg the gauge would need to be around 5 psi full scale plus something for pressure in the tube unloaded...
 
I am yet to have the drift problem with a load cell. I know for accuracy they should be calibrated yearly. The main problem is you need to use the same weight keg every time, as different models / sizes will have different weights. From my research with commercial scales & applications I expect less than a beer / year drift, though I am yet to make the unit & see in practical DIY terms.
I want to experiment with a simple, economical FOB detector for homebrew use, can only do that when I find a way to afford kegs though. For now I can start playing with weight / gas.
 
Has anyone tried these pressure-sensitive flexible strips? This one doesn't go low enough and isn't accurate enough to perform as a scale, but something like this would be waterproof and sufficient to flip on a light when a keg went below a certain level. Not sure if all kegs have this, but mine have a sort of rounded bump in the middle of the foot of the keg, where it tapers to the base of the tube inside. Build something that causes the keg to rest against that bump and put the sensor between two pieces of stiff flat material underneath that bump. The trick would be to add weight to the top of a nearly-empty keg until you can trigger the sensor by pushing down on it lightly.

http://www.karlssonrobotics.com/cart/force-sensitive-resistor-0-5/

These kinds of pressure sensor have an exponential decrease in resistance once their minimum pressure is reached, then a more gradual (but not very accurate) decrease in resistance until their maximum pressure is reached. I think the best you could do is drive a single three-pin red/green LED so that it turns gradually yellow as the keg empties, and when the pressure sensor flies into "off" state it turns solid red. You'd need to know how to work with transistors, but wouldn't need anything more complicated than that.

There's also this piezo pressure sensor:
http://www.karlssonrobotics.com/cart/flexiforce-pressure-sensor-25lbs-1-area/

This is more expensive but looks like it has the right sensing characteristics to accurately weigh a keg. Also a laminated film sensor so should be waterproof. This should be able to tell you with reasonable accuracy how full a keg is, but since there's some curvature to the response (open the datasheet link) you'd need to do a table lookup to translate accurately into weight, which brings a rasberry pi or arduino into the equation. You'd probably mount it in the same way as the other, just underneath the rounded bump at the bottom of your keg, with something built to mount the keg so it carries the weight.

Another option:
http://www.karlssonrobotics.com/cart/load-sensor-50kg/

This is supposedly the sensor used in those bathroom scales. It's $10 on the site, but if their claim is true you could get four of them for $20 from a bathroom scale. I've taken apart bathroom scales before and I think this is enclosed inside a plastic foot assembly. If you could tear apart the plastic foot and retrieve the sensor, I'm fairly certain you could make this device waterproof by sealing it in a loose plastic bag, and use it just like the above options. Water won't destroy it (though it might rust), you mainly just have to prevent condensation from changing the resistance and throwing off your measurement. Same as the piezo film sensor, you'll probably need a lookup table to translate this thing's resistance into an actual weight measurement.

For any of the above you could also build a thin three-footed platform to hold each keg. Two feet rest on the base of the keezer, and the third has the sensor in it. The position of the keg on the platform would heavily influence the readings, so you'd have to place the keg in an exact location. Probably build the platform with glued layers of plexi. This is def. more expensive than bathroom scale hacks, but would solve condensation problems.
 
There are a few threads floating around about this. Some have some very complicated solutions, and some very simple such as lifting the keg, or heating the keg and looking for the condensation to form...and I have tried both.Then I got the idea of using one of those tally counters ( you know like bouncers use to count people entering a night club). I found this on amazon......

http://www.amazon.com/dp/B008ZVLCG2/?tag=skimlinks_replacement-20

the idea being that every time I pour a beer, I just push the button.Knowing that a corny keg holds 53 12oz pours of beer, I figure it should work out fairly close , give or take a few beers either way, but close enough to anything else this cheap.After using this system for several kegs, I should have it dialed in pretty well. Plus it will be really nice to know that I have 20 beers left, or 10 beers left just by simply looking at the counter. The nice thing about these counters is that they have a built in strap (supposedly to be used to wrap around your index finger) that can be used to strap it to the beer line right at the picnic tap, or right around the faucet if you have a real tap. Plus it will give me something to do while pouring the beer. All I have to do is get in the habit of remembering to push the button with every pour, but with it strapped right to the tap area, it shouldn't be a problem. These tally counters come in several different colors too, which I plan to use to color code my picnic taps to make it easier to see which keg I'm pouring from. I just ordered 2 from amazon for a grand total of about 8 bucks shipping included. Will follow up on here once I start using them.:mug:
 
Last edited by a moderator:
I tried the bathroom scale idea. Does not work. Number one, scales like things to be set on them and then be removed, not on continuously. Two, most scales hate the cold! So I came up with a better, more accurate method. I bought a food scale for $7 online, connected it to my Arduino Uno, and now have a beer scale that as of the second test keg was -.20% accurate! When my keg ran out, the PC app that I made to read the data sent by the Arduino via Xbee gave me that result. No more lifting of kegs, just reading the display on my PC! Check out my site,http://www.beerscale.webs.com .
 
I either pull out the keg in question or shut down my kegarator for about 20 minute to half hour and check the condensation line!
 
I tried the bathroom scale idea. Does not work. Number one, scales like things to be set on them and then be removed, not on continuously. Two, most scales hate the cold! So I came up with a better, more accurate method. I bought a food scale for $7 online, connected it to my Arduino Uno, and now have a beer scale that as of the second test keg was -.20% accurate! When my keg ran out, the PC app that I made to read the data sent by the Arduino via Xbee gave me that result. No more lifting of kegs, just reading the display on my PC! Check out my site,http://www.beerscale.webs.com .

Awesome, this is exactly what I started planning out last night. I've previously used an inline flow sensor to figure out how much was left in a keg, but this required me to reset it every time I installed a new keg, so I think a weight based approach will be more automatic.

Everywhere that carries those load sensors seems to be backordered right now, but you mention using a kitchen scale. I bought a couple bathroom scales (most kitchen scales don't seem to have the range for a full keg), and they seem to be hit or miss on the types of sensors they use. I had hoped to somehow use a single sensor per keg, though your method seems a bit more stable. It does however, use 2 of the Arduino's analog inputs right? I'm going to have to get it down to a single input, as the remainder are taken up with pressure sensors attached to my gas lines (0,1,2 for keg mass sensors and 3,4,5 for the gas pressure sensors).

What model of scale did you use to get the parts, or did you eventually end up buying them individually? Do you have more pictures covering how you build the scale parts itself to distribute the weight?
 
Awesome, this is exactly what I started planning out last night. I've previously used an inline flow sensor to figure out how much was left in a keg, but this required me to reset it every time I installed a new keg, so I think a weight based approach will be more automatic.

Everywhere that carries those load sensors seems to be backordered right now, but you mention using a kitchen scale. I bought a couple bathroom scales (most kitchen scales don't seem to have the range for a full keg), and they seem to be hit or miss on the types of sensors they use. I had hoped to somehow use a single sensor per keg, though your method seems a bit more stable. It does however, use 2 of the Arduino's analog inputs right? I'm going to have to get it down to a single input, as the remainder are taken up with pressure sensors attached to my gas lines (0,1,2 for keg mass sensors and 3,4,5 for the gas pressure sensors).

What model of scale did you use to get the parts, or did you eventually end up buying them individually? Do you have more pictures covering how you build the scale parts itself to distribute the weight?

How are your circuit design and soldering skills? I might be able to help you figure out a solution for that problem.

Also, where did you find pressure sensors for the gas lines? The places I've looked those are all industrial quality and hundreds of dollars each.

EDIT (with parts): I found an analog multiplexer for about 50 cents that should get you set up to read multiple analog channels from one arduino pin. You'll need a small protoboard and a few extra parts, but it should be able to operate from the arduino's power supply.

Texas Instruments CD4051be CMOS 8x1 analog mux
Datasheet

What you would do to use this method, if you're up to it:
1: Wire up the multiplexer with a latch on the addressing pins, as well as any of the other suggested parts in the reference circuit diagrams that seem important.
2: Design circuit to deliver voltages from your pressure sensors to the multiplexer chip.
3: Program the Arduino to, instead of just reading an analog pin, send an address and a latch pulse through its digital pins, wait for the analog circuit to settle, and then read the analog pin.
4: Repeat (3) for each addressed input, and you have read up to 8 kegs from one analog in.

Example:
DIGITAL OUT: 000 1
DIGITAL OUT: 000 0
Wait a few milliseconds
Read analog input
DIGITAL OUT: 001 1
DIGITAL OUT: 001 0
wait a few milliseconds
read analog input
DIGITAL OUT: 010 1
DIGITAL OUT: 010 0
wait a few milliseconds
read analog input

[etc]
 
How are your circuit design and soldering skills? I might be able to help you figure out a solution for that problem.

Also, where did you find pressure sensors for the gas lines? The places I've looked those are all industrial quality and hundreds of dollars each.

My digital circuit work is better than my analog, but I've done a lot with op amps before, so I'm up to the task.

I picked up http://www.mouser.com/ProductDetail/Freescale-Semiconductor/MPX4250GP/?qs=N2XN0KY4UWX3WOzXduHkkA== for the pressure sensor. As a warning though, digikey sent me a mail telling me they have end of life for Feb 2014, so there's probably a different part to replace them. I plumb them in with a T fiting on the gas line and a piece of tubing, and they've been accurate. Well, as accurate as the analog inputs on the Arduino can be. The reference voltage from the Arduino kind of bounces all over the place so you have to take a number of reads and average them.

Hopefully, I only need 6 analog inputs (3 scales, 3 pressure sensors) so I can get by without the mux.
 
It looks like these solutions still require you to take your keg out and weigh it each time you want an update. Has anyone found a solution for taking constant readings?
 
It looks like these solutions still require you to take your keg out and weigh it each time you want an update. Has anyone found a solution for taking constant readings?

I know the Kegbot project has been talked about, but they recently had a kickstarter project and are coming out with new hardware. It looks to be very user friendly, no more soldering, and a server isn't required anymore. Anybosy that was turned off to it for those reasons should take another look.

Full disclosure: I have the old style where I soldered and fiddled forever...and it's awesome.
 
Back
Top