Keg Weighing System

Australia & New Zealand Homebrewing Forum

Help Support Australia & New Zealand Homebrewing Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Nick JD said:
What about a temperature sensitive sticker that goes down the side of the keg. You open the fridge for a minute, and you can see where the fluid level is due to the empty part of the keg warming faster than where the beer is.

50c per keg.
Yep. This is what I use.

The geek in me likes the idea of a computer that could accurately measure the beer left in the keg, but the lazy in me means I couldn't be fucked unless it's a "plug and play" type deal and doesn't cost a thousand bucks.
 
I just give the kegs a bit of a tug... if it moves up, then its nearly empty ;)

funnily enough the same 'system' works on my pool chemical barrels (auto-feed system)
 
Edak said:
Yeah, it appears you are right. The website is misleading but the forums describe it perfectly. For four taps you have to either:
1. buy two arduinos, two keg boards and two coasters
2. buy arduino mega, design own keg board and two coasters
3. re-design the wheel

Personally that sounds like too much hardware.


Zizzles method is great but it requires a LOT of work:
* Build custom floor for the fridge to which you can bolt the measuing apparatus
* Build the measuring plates
* Ensure that your floor and plates aren't going to get destroyed by condensation/liquid damage
* Calibrate each measuring plate
* Build the amplification circuit
* Ensure shielding is strong
* Much more wiring


I am thinking that I will use a purpose-built data logger which can handle everything and has a webserver in-built. I already have a suitable solution with multiple high-speed counters and inputs for up to 5 load cells, but I still like a proper, customised solution.
The have a keg board designed to fit the Mega, they just dont produce it, so simple down load and send it off for fabrication.
 
Nick JD said:
What about a temperature sensitive sticker that goes down the side of the keg. You open the fridge for a minute, and you can see where the fluid level is due to the empty part of the keg warming faster than where the beer is.

50c per keg.
50c a sticker? any links to similar products on ebay? I was looking at getting some of Keg Volume Stickers but I think the price is a bit much.
 
Thinking about it further, you could have a bunch of hardware counters going into a shift register, then you could read that shift register serially with arduino, as long as the arduino can service these quickly enough then you can have many flow sensors connected via one serial interface (with one enable line per flow sensor to put counter into Z state).


SN74HC590AN --->

SN74HC590AN --->
SN74HC590AN ---> CD4021 ---> Arduino digital inputs

SN74HC590AN --->

I think that the nominal frequency output during a pour would be 500Hz (based around 6100Hz/L and about 6 seconds to pour a pint) and each counter has 256 states so you need to read each within 0.5 sec to prevent overflow.

OR!!!!

Use one of these chips for each flow meter (32bit counter > SPI Serial interface), which means you have no read time issues and only require one chip to talk to your arduino using the existing SPI interface. This would be the BEST solution!
http://www.anaheimautomation.com/products/ics/lsi-csi-item.php?sID=270&serID=17&pt=i&tID=159&cID=55
 
Edak said:
Thinking about it further, you could have a bunch of hardware counters going into a shift register, then you could read that shift register serially with arduino, as long as the arduino can service these quickly enough then you can have many flow sensors connected via one serial interface (with one enable line per flow sensor to put counter into Z state).


SN74HC590AN --->

SN74HC590AN --->
SN74HC590AN ---> CD4021 ---> Arduino digital inputs

SN74HC590AN --->

I think that the nominal frequency output during a pour would be 500Hz (based around 6100Hz/L and about 6 seconds to pour a pint) and each counter has 256 states so you need to read each within 0.5 sec to prevent overflow.

OR!!!!

Use one of these chips for each flow meter (32bit counter > SPI Serial interface), which means you have no read time issues and only require one chip to talk to your arduino using the existing SPI interface. This would be the BEST solution!
http://www.anaheimautomation.com/products/ics/lsi-csi-item.php?sID=270&serID=17&pt=i&tID=159&cID=55
hmmm, think I know where your going, but it's like the whole zizzle thing, I seriously think you two could talk this stuff for hours and of course empty the keg being counted. I can solder, I can drill, upload software to devices, I've even had a linux server. (with help) But alot of this stuff goes way over my head. Electronics and programing and me, well we have an understanding, Me: "please work" the electeonics etc .... well they do what they like. Just to ensure I haven't got you wrong; your going to shift the flow sensor inputs from the serial bus by having another chip convert it to digital. Or am I somewhere looking for the wrong star gate.
 
MastersBrewery said:
... Just to ensure I haven't got you wrong; your going to shift the flow sensor inputs from the serial bus by having another chip convert it to digital. Or am I somewhere looking for the wrong star gate.
The flow signal is already technically digital (pulses) but the first counter chip translates it to a binary number that represents the number of counts (eg. 0100 1001 = 73). The problem is that each bit is in parallel so each sensor requires at least 8 digital lines. Now we don't have 32 digital inputs on the arduino (8 digital lines x 4 flow sensors) so we use a separate chip (shift register) to convert those 8 separate lines to a single, serial line. We can get away with using one shift register to save on build cost and less data lines by allowing it to switch between which sensor it reads parallel data from.

Or, you could forget all the above and use the second chip solution that I mentioned with SPI built in, one chip per flow sensor, no timing constraints either. Four chips on a shield and plug and play.

If anyone has CPLD experience then we could also look at halving the cost of the chips by using a CPLD instead...
http://ad7gd.net/counter/
 
Edak said:
Zizzles method is great but it requires a LOT of work:
* Build custom floor for the fridge to which you can bolt the measuing apparatus
* Build the measuring plates
* Ensure that your floor and plates aren't going to get destroyed by condensation/liquid damage
* Calibrate each measuring plate
* Build the amplification circuit
* Ensure shielding is strong
* Much more wiring
Edak

Do you believe that flow sensors would be the best way to achieve keg level monitoring? The flow sensors that the kegbot use are $60 plus delivery each. Is this what you would intend on using if you were to go ahead with this project?

I was really hoping Zizzle's idea would be the way to go based only on the cost. I believe we could get it way under $100 for the whole system.

I've already started doing some extremely basic coding for a simple keg weighing system that will most likely display the % of whats left in each keg which is totally dependent on whether I can sort out the electronics side. There are a few load cell tutorials online so hopefully this won't be a problem...
 
PeteQ said:
Edak

Do you believe that flow sensors would be the best way to achieve keg level monitoring? The flow sensors that the kegbot use are $60 plus delivery each. Is this what you would intend on using if you were to go ahead with this project?

I was really hoping Zizzle's idea would be the way to go based only on the cost. I believe we could get it way under $100 for the whole system.

I've already started doing some extremely basic coding for a simple keg weighing system that will most likely display the % of whats left in each keg which is totally dependent on whether I can sort out the electronics side. There are a few load cell tutorials online so hopefully this won't be a problem...
My prob with this would be my kegs aren't all the same or even the same volume, so each keg would have to have it's own individual schema
 
I believe that the flow system should give you a very accurate result, even if you have +/-1% accuracy that's less than a pot.

It just depends on where you want to spend your time, either building mechanical things and going through a lot of trial and error, or use a sensor that's easy to install and proven to work by others. What load cells do you have and how do you plan on attaching them to weigh your keg? Do you have a diagram of how it will work?
 
I still dont have any load cells, but i do have some instrument amplifiers on the way. I'm hoping to use load cells from some cheap digital scales as per Zizzle's setup.

After further investigation it looks like Matho has done it as well - http://aussiehomebrewer.com/topic/41443-hlt-controller/page-2

Also on the bottom of that page Zizzle mentions this article and the instrument amplifiers I'm waiting on - http://www.ehow.com/how_5980113_make-load-cell-use-arduino.html

Definitely not sure if any of this will actually work, hit and hope i guess...
 
I would be very interested if someone was to build a kit in the flow sensor type that worked for a tap system. I guess i would shell out for the better sensors at $60 if the rest of it came in a plug and paly box that was a reasonabe price. I would not be interested in a weighing system for myself.

Kegbot really does look the ant's strides.
 
punkin said:
I would be very interested if someone was to build a kit in the flow sensor type that worked for a tap system. I guess i would shell out for the better sensors at $60 if the rest of it came in a plug and paly box that was a reasonabe price. I would not be interested in a weighing system for myself.

Kegbot really does look the ant's strides.
I agree
 
Edak,
The intricacies of the build will come way later on down the track. I do like playing around with things and if i can get this project working I'll be stoked but for now I'll just take it one step at a time. I was hoping if i get stuck I would bank on you guys to help me out!

Masters
I have already sorted out code for a keg zero system for each keg on each load cell so it knows the starting weight of the keg prior to racking. I'm unsure whether to do the same thing full, I'll have to see how well the load sensors work. BUT one thing I didn't think of with the kegbot is what if your keg isn't full from the start? Do you somehow calculate the amount of beer that goes into the keg when racking maybe with another flow sensor? I'm going to look into this....

Don't get me wrong, the kegbot is frieken awesome but the price is pushing me towards something a little, well, a lot cheaper. To install a kegbot on my 4 tap keezer would cost over $300, certainly money I don't have lying around for something I don't really need....
 
I looked into options to identify how much was left in the keg. Would be cool with a digital display for "Brew Porn". But I just tap the side of the keg with my finger going from the top and moving down. As soon as the sound changes to a lower note, I know where the level is. :)
 
I think as punkin said , if someone comes up with a system that works and can, do a similar thing to what matho did with the braumiser I am dead set certain you'd have a line up of takers, most of us just don't have the electronics and coding skills to pull something like this off. So next best is kegbot.... but as stated throughout this thread it'll cost a bit.

and Yes it's all about the bling, but useful bling
 
Is anyone still thinking of taking on this project using the kegbot?

I'd be very interested if someone was to build an extra. I have a 4 tap kegerator.
 
Back
Top