[Feeler] Any Interest in Digital Tap Lists?

A cursory Google search of the site didn't turn up many threads relating to digital tap lists. In fact, only one thread I can find.

I'm currently finishing up a digital taplist. Let's cut to the chase, mmkay. I received the code on which I based my version via a GNU Share-a-Like license. I'm willing to pass on all of my source code and updated graphics under the same license (free for non-commercial or with attribution).

I'm willing to create a step-by-step tutorial from unboxing hardware to up and running, written to be Marine-proof (or even drunken HBTer-proof!). That said, I'm only going to put my time towards such a massive project if it's something that the community here wants.









Would anybody here be interested in something like this for their home? Please let me know either by Liking this post, replying, or both.


UPDATE: Made plenty of tweaks since posting the screenshot above. I'm making tens or hundreds of edits a day to the code right now. So instead of updating the screenshot manually, I just made the Pi web-facing and forwarded the ports in my local network.

"What does that mean?!" you say. It means you can check out a live version: http://thadius856.servebeer.com/

Things may look weird at times, as you're seeing my behind-the-scenes edits in real time.

I like it and think it looks really cool. Just like everyone else though, I'm cheap. What's the cost and what do you actually get. I'm computer stupid so break it down for me.

Thanks!

If you have some electronic pieces laying around, you could build this for ~$50 out of pocket. What you see pictured probably cost me closer to $200, but $120 is locked up in the (24") display.

The other drywall anchors you see are for two satellite displays still in the planning phase.

Alright, cost breakdown. There's a few things you need, and a few things can be helpful but aren't required. Depending on what you have laying around, it may be cheaper or more expensive.

You need one Raspberry Pi mini computer. They're $35 + S/H from the manufacturer. Otherwise, $43 shipped on Amazon. Free 2-day Air shipping if you have a Prime membership, otherwise free ground shipping.

You need a display of some sort. I bought 24" Dell IPS panels for this project, but if you had an old monitor laying around, you could use that. I recommend 1920x1080 resolution for this. I paid $119 shipped for this as a refurb (MSRP $190), and I threw in two more. Dell regularly has refurb sales with 19" panels below $65 shipped.

You'll need a display cable. If the monitor has HDMI, you need an HDMI cable. If the monitor has only DVI, you need either 1) an HDMI cable and a DVI-to-HDMI adapter ($3) or 2) an HDMI-to-DVI cable ($6).

You need a power supply. Most people have old cell phone chargers that will work. Kindle chargers, iPhone chargers, etc. Must be minimum 700 maH (will be printed on the charger). If you can't find one, knockoff iPhone chargers are all over eBay (search "colorful iPhone charger") for $1-2 shipped.

You need a power cable. It should be USB A (standard USB plug type) to USB Micro (like a cell phone cable). I already had plenty, but there's a 5pk on Amazon for $7 shipped, under $1 ea from Monoprice + S/H, and plenty other options available.

You need either a USB or wireless keyboard and mouse (at least for setup). I got tired of swapping out my wireless keyboard/mouse dongle with my PC while programming and bought a Favi FE01-BL mini wireless keyboard, but it's completely optional. Who doesn't have a USB keyboard/mouse these days?

That's all you need.

A USB wifi adapter is helpful for web browsing with it, but not required. If you want to mount the monitor to a wall, you'll need a wall mount. This cheapie one is $10, this one with tilt is $15, and this one (I'm using) is a premium version with tilt, pan, and an easy-mount setup for $22.

I don't get it....

Does the display change as you dispense from the tap?

twst1up said:

I don't get it....

Does the display change as you dispense from the tap?

Click to expand...

Yeah I want to know too and if so, I want!

But please, don't put baby (the guitar) in the corner.

'Fraid not. It does not update as your dispense (yet).

Problem is, the cheapest reliable food-safe flow meters with decent precision are $60 ea, plus import fees and shipping. That's more than the price of a corny keg for most of the US!

If you're interested in live keg updates, check out the Kegbot project. So far, they're limited to 2 kegs on a flow meters, and they require an Android tablet for display. Talk about expensive!

If I could find a source of reliable, cheap food-grade flow meters, I'd consider developing based around an Adafruit sheild add-on. Unfortunately, the market just isn't there yet. Sorry.

twst1up said:

I don't get it....

Does the display change as you dispense from the tap?

Click to expand...

Hello said:

Yeah I want to know too and if so, I want!

But please, don't put baby (the guitar) in the corner.

Click to expand...

That old pile of crud? That's a Fender Starcaster. $80 retail at any big box store. I hang it in the corner because it's my ****tiest guitar. It's the one I hand newbs when they come over and want to play Rocksmith 2014. It exists so they don't pick stab the **** out of my other pieces.

I hang my nice guitars with pride: Gibson Les Paul '52 Tribute ('86 reproduction) in Sunburst, '92 Epiphone Les Paul in Wheat, '81 Vantage VSH-445 in Mahogany stripe, and my Warmoth custom.

Why do you need a flow meter? A calibrated load cell with ADC should do the job for around $15 per keg slot, assuming you know the exact starting quantity (same assumption as using a flow meter). If you wanted it to be super accurate you could also work in the FG of each keg.

chocotaco said:

Why do you need a flow meter? A calibrated load cell with ADC should do the job for around $15 per keg slot, assuming you know the exact starting quantity (same assumption as using a flow meter). If you wanted it to be super accurate you could also work in the FG of each keg.

Click to expand...

Every thread I've read indicates load cells and strain gauges are inaccurate at this weight. Also, they don't work if you cram your kegs in for a tight fit (which I do).

I'll gladly be proven wrong. Have plenty of Arduinos around to play with, and the GPIO is pretty simple to hook up with a few pieces in-line.

Let me know if you have spare load cells on hand.

Very nice display. I can see a commercial version of this in small tap houses and brew pubs.

thadius856 said:

Every thread I've read indicates load cells and strain gauges are inaccurate at this weight. Also, they don't work if you cram your kegs in for a tight fit (which I do).

I'll gladly be proven wrong. Have plenty of Arduinos around to play with, and the GPIO is pretty simple to hook up with a few pieces in-line.

Let me know if you have spare load cells on hand.

Click to expand...

Sorry, never actually done it... I'm strictly an armchair digital logic designer!

I've thought about a similar thing, though, and I've seen people having done it with plastic flow meters from Adafruit (10 bucks each; http://www.adafruit.com/products/828 - example keg flow majigger is in the Tutorials)) - they don't SAY they're food grade, but it seems like they're made of polypropylene which should be fine. That said, I'm not sure they would be any more accurate than a properly calibrated load cell as they rely on an internal pinwheel (like the inverse of the impeller from an impeller pump) to send pulses to the GPIO of the microcontroller. You will start stacking up error margins pretty quickly IMHO (even if you found a "food grade certified" version of the same).

If you wanted to be really fancy you could mount liquid level sensors (40 bucks a pop) inside each of your kegs - maybe the gas post could also function as a coaxial electrical connector for the level sensor? I don't have Cornys so I'm not sure if the poppet is isolated from the outside of the post (or the whole post from the main surface of the keg - could also be another way to get the signal out)

CA_Mouse said:

Very nice display. I can see a commercial version of this in small tap houses and brew pubs.

Click to expand...

I have seen a similar thing at a pub in Portland - I think it was Bailey's Taproom? It had video screens with every thing on tap and a "level bar" for each keg. Not sure what mechanism they used for it, though.

Oh, and the benefit of the load cell approach is that nothing touches your beer so you don't have to worry about it being food grade.

chocotaco said:

Sorry, never actually done it... I'm strictly an armchair digital logic designer!

I've thought about a similar thing, though, and I've seen people having done it with plastic flow meters from Adafruit (10 bucks each; http://www.adafruit.com/products/828 - example keg flow majigger is in the Tutorials)) - they don't SAY they're food grade, but it seems like they're made of polypropylene which should be fine. That said, I'm not sure they would be any more accurate than a properly calibrated load cell as they rely on an internal pinwheel (like the inverse of the impeller from an impeller pump) to send pulses to the GPIO of the microcontroller. You will start stacking up error margins pretty quickly IMHO (even if you found a "food grade certified" version of the same).

If you wanted to be really fancy you could mount liquid level sensors (40 bucks a pop) inside each of your kegs - maybe the gas post could also function as a coaxial electrical connector for the level sensor? I don't have Cornys so I'm not sure if the poppet is isolated from the outside of the post (or the whole post from the main surface of the keg - could also be another way to get the signal out)

Click to expand...

Liquid tiptubes, gas diptubes, poppets, posts, and the keg body are all connected for electrical purposes. The only exception would be the plastic versions of gas diptubes, but I've drilled out my kegs that had them to accept the metal ones instead.

Hmmm... I have a display sitting around.

chocotaco said:

Sorry, never actually done it... I'm strictly an armchair digital logic designer!

I've thought about a similar thing, though, and I've seen people having done it with plastic flow meters from Adafruit (10 bucks each; http://www.adafruit.com/products/828 - example keg flow majigger is in the Tutorials)) - they don't SAY they're food grade, but it seems like they're made of polypropylene which should be fine. That said, I'm not sure they would be any more accurate than a properly calibrated load cell as they rely on an internal pinwheel (like the inverse of the impeller from an impeller pump) to send pulses to the GPIO of the microcontroller. You will start stacking up error margins pretty quickly IMHO (even if you found a "food grade certified" version of the same).

If you wanted to be really fancy you could mount liquid level sensors (40 bucks a pop) inside each of your kegs - maybe the gas post could also function as a coaxial electrical connector for the level sensor? I don't have Cornys so I'm not sure if the poppet is isolated from the outside of the post (or the whole post from the main surface of the keg - could also be another way to get the signal out)

Click to expand...

The flow sensors I have pulse 6000 times per liter. I found that they stayed incredibly accurate through the pouring of a full keg. I poured nearly a full keg of 12oz beakers, and the reported numbers were all accurate to a few tenths of an ounce (I'll see if can dig up my dataset). Remember that all you need is accuracy down to the average glass size you pour (or maybe half of that). That's a lot of wiggle room.

Still, I've also played around with load cells as I didn't like resetting the volume of the kegs every time I added a new one. Cheap load cells are pretty inaccurate, but the bigger problem is temperature and creep.

Temperature can be solved by using a loaded and an unloaded cell to form the bridge, so that the unloaded cell cancels out some voltage from the loaded cell kept at the same temperature.

Creep is a bigger problem, even the most expensive cells require re-zeroing periodically. When left under constant load, the value will creep over time. I've got them wired up, I just haven't done any long term testing with keg weights, I'll probably get around to it in the next month or so. Though I may skip it and just go back to the flow sensors.

CA_Mouse said:

Very nice display. I can see a commercial version of this in small tap houses and brew pubs.

Click to expand...

It already exists, its called Digital Pour

http://digitalpour.com/

Many breweries and tap houses in Oregon use it if they have a lot on tap. As mentioned Baileys Taproom in Portland uses it, as well as Orenco Taphouse in Hillsboro. For example both of those places use it on a big probably 55-60" LCD TV mounted above the bar.

I would definitely build this if I could get some instructions. It's not quite like anything I've DIY'ed before.

absolutely! i love this

I've been wanting to make something like this. I have a 22" monitor haning around collecting dust.

chocotaco said:

I've thought about a similar thing, though, and I've seen people having done it with plastic flow meters from Adafruit (10 bucks each; http://www.adafruit.com/products/828 - example keg flow majigger is in the Tutorials)) - they don't SAY they're food grade, but it seems like they're made of polypropylene which should be fine.

Click to expand...

I'd seen those flow meters and dismissed them because the lack of food-grade rating (NSF stamp). If people have tested them and are not getting flavor scalping, I'll certainly reconsider.

opqdan said:

The flow sensors I have pulse 6000 times per liter. I found that they stayed incredibly accurate through the pouring of a full keg. I poured nearly a full keg of 12oz beakers, and the reported numbers were all accurate to a few tenths of an ounce (I'll see if can dig up my dataset). Remember that all you need is accuracy down to the average glass size you pour (or maybe half of that). That's a lot of wiggle room.

Click to expand...

Can you link to the model you used? This has certainly caught my attention now.

chocotaco said:

Oh, and the benefit of the load cell approach is that nothing touches your beer so you don't have to worry about it being food grade.

Click to expand...

Agreed, but again, I cram my kegs into the keezer as many here do. Load cells don't work in that situation, as a keg may be supported partially or fully by its neighbor.

CA_Mouse said:

Very nice display. I can see a commercial version of this in small tap houses and brew pubs.

Click to expand...

FuzzeWuzze said:

It already exists, its called Digital Pour

http://digitalpour.com/

Click to expand...

I have no interest in producing a commercially viable version. As you said it already exists. I don't have the time or resources to support businesses, especially when it's a product that is key to their livelihood. Business owners FREAK OUT when key technologies break on them, and I don't care to deal with all that mess.

This project would be intended for home use (free limited support). It would be designed around metrics a homebrewer would be interested in displaying (IBU, SRM, etc). A taphouse wants to display other things that maybe aren't so important in a home setting (prices, social media tickers, live reviews, etc). Of course, there's some basic elements that we probably all want (beer name, style, a description, keg levels, etc).

Should put it on github and make it open source

milldoggy said:

Should put it on github and make it open source

Click to expand...

Github will be utilized once I get a functioning build completed.

thadius856 said:

Can you link to the model you used? This has certainly caught my attention now.

Click to expand...

I have 3x SwissFlow sf800
http://www.swissflow.com/sf800.html

I got the model idea from KegBot a number of years ago. The kegbot website says that it is food safe, though the SwissFlow website doesn't say that it is. I'm not convinced that it is, but I'm also not too concerned. Only a small amount of beer has extended contact with the plastic, and the temperature is low (I wouldn't run any hot or caustic liquids through it).

I think I paid way less than what they are going for now (possibly ebay, but I can't recall).

opqdan said:

I have 3x SwissFlow sf800
http://www.swissflow.com/sf800.html

I got the model idea from KegBot a number of years ago. The kegbot website says that it is food safe, though the SwissFlow website doesn't say that it is. I'm not convinced that it is, but I'm also not too concerned. Only a small amount of beer has extended contact with the plastic, and the temperature is low (I wouldn't run any hot or caustic liquids through it).

I think I paid way less than what they are going for now (possibly ebay, but I can't recall).

Click to expand...

Indeed, I had been trying to just roll out a ready-to-use Kegbot in my home. However, they're limited to 2 flow sensors with the Arduino's limited GPIO capacity. They're also pretty firmly entrenched in the Android tablet + Arduino board + kegboard coasters realm. In the end, they seem to get a slower device, less upgradability, and a much smaller screen.

I did find a stack of SF-800s on eBay earlier, sold by a company that went out of business before they could implement them. We're looking at $26 per, plus $7 S/H instead of $60 per, plus $20 S/H and international shipping wait. It says 86 sold and "more than 10 available". I wish eBay would show the full number available by default so I'd know if he has 11 or 1100 left.

Been doing some posting over on the RPi official forums. Whereas the Arduino is limited to 2 flow meters, the RPi has 21 GPIO pins available on the board. Because there's a 3-pin harness on the SF-800's, I assume each flow sensor needs two GPIO pins, plus a ground that can easily be shared. Looks like I could build a MCP23017-based breakout board for $10-15 which would take allow me to split two GPIO pins into 16 (or I could double-up and go 4-to-32 for another $5). This also lets me nuke the MCP23017 if I mess up the voltages ($2) instead of the RPi ($42) itself.

In other words, I think the RPi could be expanded to accept 8 flow meters for ~$15, or 16 flow meters for ~$20. Plus flow meter cost, of course.

I hate you guys right now...

thadius856 said:

Indeed, I had been trying to just roll out a ready-to-use Kegbot in my home. However, they're limited to 2 flow sensors with the Arduino's limited GPIO capacity. They're also pretty firmly entrenched in the Android tablet + Arduino board + kegboard coasters realm. In the end, they seem to get a slower device, less upgradability, and a much smaller screen.

I did find a stack of SF-800s on eBay earlier, sold by a company that went out of business before they could implement them. We're looking at $26 per, plus $7 S/H instead of $60 per, plus $20 S/H and international shipping wait. It says 86 sold and "more than 10 available". I wish eBay would show the full number available by default so I'd know if he has 11 or 1100 left.

Been doing some posting over on the RPi official forums. Whereas the Arduino is limited to 2 flow meters, the RPi has 21 GPIO pins available on the board. Because there's a 3-pin harness on the SF-800's, I assume each flow sensor needs two GPIO pins, plus a ground that can easily be shared. Looks like I could build a MCP23017-based breakout board for $10-15 which would take allow me to split two GPIO pins into 16 (or I could double-up and go 4-to-32 for another $5). This also lets me nuke the MCP23017 if I mess up the voltages ($2) instead of the RPi ($42) itself.

In other words, I think the RPi could be expanded to accept 8 flow meters for ~$15, or 16 flow meters for ~$20. Plus flow meter cost, of course.

I hate you guys right now...

Click to expand...

I only used the hardware ideas, I haven't used the actual KegBot software. The swissflow only required a single pin for IO (the other two are 5V and GND), but due to the pulsing nature I've had to hook it up as an interrupt, which the Arduino has only 2 of though it looks like newer versions may support more.

While I've had a working Arduino (and later a Netduino, and then back to Arduino again) solution, I spent a long time feature creeping the PC software side (I had a half cocked idea to implement facial recognition though it was only right about 50% of the time) and it never really came together. Since I am going to be getting back into this over the next few months, I am probably going to start from the ground up, so I've also been looking at the Pi. All you need is the ability to count pulses on a GPIO in the ~1kHz range (assuming pouring one imperial pint in 5 seconds). If that can be done on any of the Pi pins, then you've got a lot of options. My only road block is that I'd have to add a separate ADC chip for my analog pressure sensors.

opqdan said:

All you need is the ability to count pulses on a GPIO in the ~1kHz range (assuming pouring one imperial pint in 5 seconds). If that can be done on any of the Pi pins, then you've got a lot of options. My only road block is that I'd have to add a separate ADC chip for my analog pressure sensors.

Click to expand...

Here's the question I asked this morning. Seems very plausible.

http://www.raspberrypi.org/phpBB3/viewtopic.php?f=91&t=64780&p=477308#p477308

I could easily see one of these for my setup if the flow meters cost was slightly lower.

Renegades_Brew said:

I could easily see one of these for my setup if the flow meters cost was slightly lower.

Click to expand...

Right. List price $60 ea + $20 S/H + international wait is pretty steep.

I noticed the guy with them listed on eBay for $26 ea + $7 S/H listed it as "Or Best Offer". I submitted an offer for $20 ea + $7 S/H and he accepted.

Now I've got 10 of these things on the way. :S

thadius856 said:

Indeed, I had been trying to just roll out a ready-to-use Kegbot in my home. However, they're limited to 2 flow sensors with the Arduino's limited GPIO capacity. They're also pretty firmly entrenched in the Android tablet + Arduino board + kegboard coasters realm. In the end, they seem to get a slower device, less upgradability, and a much smaller screen.

I did find a stack of SF-800s on eBay earlier, sold by a company that went out of business before they could implement them. We're looking at $26 per, plus $7 S/H instead of $60 per, plus $20 S/H and international shipping wait. It says 86 sold and "more than 10 available". I wish eBay would show the full number available by default so I'd know if he has 11 or 1100 left.

Been doing some posting over on the RPi official forums. Whereas the Arduino is limited to 2 flow meters, the RPi has 21 GPIO pins available on the board. Because there's a 3-pin harness on the SF-800's, I assume each flow sensor needs two GPIO pins, plus a ground that can easily be shared. Looks like I could build a MCP23017-based breakout board for $10-15 which would take allow me to split two GPIO pins into 16 (or I could double-up and go 4-to-32 for another $5). This also lets me nuke the MCP23017 if I mess up the voltages ($2) instead of the RPi ($42) itself.

In other words, I think the RPi could be expanded to accept 8 flow meters for ~$15, or 16 flow meters for ~$20. Plus flow meter cost, of course.

I hate you guys right now...

Click to expand...

Man, I've had my eye on something like KegBot for a while now but cant pull the trigger because it only supports 2 taps (of which I have 3). If it were possible to use a RPi with some flow meters and an RFID capability (or not) I'd be seriously interested!

DustyTheBrewer said:

Man, I've had my eye on something like KegBot for a while now but cant pull the trigger because it only supports 2 taps (of which I have 3). If it were possible to use a RPi with some flow meters and an RFID capability (or not) I'd be seriously interested!

Click to expand...

I hope to have flow meters, breakout boards and research done before the end of January. Hopefully I can have something functioning and a final word back. However, I'm pretty hopeful.

I think I'll need to migrate this thing away from CSV and into SQL for this to be viable long-term. I'll also need to expand my use of PHP conditionals to expand to wine/cider/kombucha/soda.

Been looking into it (and I'm not sure if this gets a bit off topic from your list project) but it looks like the arduino "mega" should be able to handle up to six taps. This is pretty far over my head, however and I wouldn't know really how to implement it. Just to add some fodder to your brainstorming battle

https://kegbot.org/kegbb/index.php?p=/discussion/616/three-taps

DustyTheBrewer said:

Been looking into it (and I'm not sure if this gets a bit off topic from your list project) but it looks like the arduino "mega" should be able to handle up to six taps. This is pretty far over my head, however and I wouldn't know really how to implement it. Just to add some fodder to your brainstorming battle

https://kegbot.org/kegbb/index.php?p=/discussion/616/three-taps

Click to expand...

Yup. Looks like it. However, you'd still need additional GPIO pins if you wanted solenoids. So, you're looking at 3 taps + 3 solenoids.

Still, the Mega certainly costs more than the Pi, and is a lot less powerful.

question? love the set up - looks like you have 10 kegs on tap, what size are they, that seams like a lot to fit in that size chest freezer, only curious as I would love to increase that amount I can fit if I am doing it wrong.

cjones said:

question? love the set up - looks like you have 10 kegs on tap, what size are they, that seams like a lot to fit in that size chest freezer, only curious as I would love to increase that amount I can fit if I am doing it wrong.

Click to expand...

This GE 15 cu ft freezer holds 11 if you use the hump as well. I went with 10 + 20# CO2 on the hump, hence the tall collar.

Check out the spec sheet using the link in my sig.

def interested in this - thanks

Great project. Absolutely interested.

Interested as well as I already have a 42" screen mounted next to my kegerator as my brewing computer.

It's not mounted in portrait orientation though. How hard do you think it would be to adapt?