OpenBar

Some time ago my flatmate built a lazydrinker automatic bar (lazydrinker.com). Basically it is an array of microcontroller activated solenoids which allow compressed air into bottles, forcing liquid out to a nozzle. While this is a pretty cool system it has some issues:

  • It's expensive ($90 board only)
  • It's closed source
  • The software is unfinished and poor
  • It's not extensible

So I decided to make my own. The concept is for an open-source harware and software combination with community involvement. Hardware is an arduino-based basic 16 solenoid driver. Components are an ATMega 168 with arduino bootloader, ftdi style serial header, two 74*4051 multiplexers and two ULN2803 darlington arrays to sink the solenoid current. The DB25 is pin compatible with the lazydrinker system. 595 shift registers would probably be better, but I didn't have any and I'm not sure that the darlingtons could sink enough current to allow more than one solenoid to be open. A proto-board test article has been made, and I've designed a prototype PCB for fabing, but it probably wont get made until next year (moving agian). Software is currently a processing sketch for test purposes, may move to Java for a better UI but maintaining cross platform capabilities.

A website will be set up in due course. This will have code, circuit diags etc.

proto-board version

Alec

Nice. I love PRACTICAL projects, and this would be awesome.

I've been following Clint Rutkas' "Drunktender" project. He's a microsoft employee and has written some awesome software for looking through lists of drinks based on what knows how to mix and what fluids it has available. His hardware is pretty cool too though, you should check it out. i know I always like seeing how others did similar projects.

http://www.betterthaneveryone.com/category/10.aspx

Video - http://www.betterthaneveryone.com/archive/2009/07/31/877.aspx

Thanks for the link, it's pretty cool, but C#, SQL (MSSQL I assume) and I think .NET framework. Win32 only is a bit of a killer. I've started playing around with SWT in elipse, and looking at that HTML/java might be the way to go...

Alec

Website up: http://openbar.2kmegs.com/

You're my hero! ;D

A few ideas:

  • Have the pour head on an X/Y CNC-like axis so it can pour in multiple glasses, you could have them sit in fixed recessions on a platform beneath which have sensors to detect their presence, so as to avoid spills. It should be pretty easy since it'll be aiming for about a 1/4" target to be centered on the glass, which to a stepper motor is like trying to throw a baseball in the grand canyon, so being ultra-precise in the mechanics isn't a priority. This would also allow for having the ability to purge the pour head with clear water between drinks, perhaps having quick flushes between drinks and more thorough cleansings and perhaps blowing out the nozzle with air during shutdown.

  • Giving it the ability to add ice would be a cool feature too. 8-)

  • Maybe using recycled bottles of all the same shapes and sizes to standardize everything would make maintenance easier? You could fill each bottle out of the originals, using clear ones so making sensors to keep track of the levels of content would be easier, and to occasionally run them through a dishwasher to keep them clean.

I stumbled upon these valves in another thread, thought anyone thinking of taking on this project might find them usefull:

http://www.robotshop.us/dynalloy-flexinol-air-valve-2.html

Tornadoboy, those valves are interesting, although not exactly suitable for this kind of project. The CNC positioning head would be cool, but I think it's a little bit down the priority list. :D Possibly a more practical way would be to have a rotating carousel below the head to allow a new glass to be presented.

I considered level (weight) sensing as a stock capability. It seems weight sensing can be either good or cheap. Strain gauge based systems are expensive, even diy it would be about $15 per bottle (strain gauge, beam and instrumentation amp), but would give accurate and repeatable weights (with a little calibration). Capacitance based weight sensing would be much cheaper, but there is not COTS solution so a lot of development work would be required. Finally optical sensing is much, much cheaper, and can be done with widely available hardware, but requires standard bottles and a lot of calibration work. Believe me, it's on the list, but someway down the track.

Priority one is getting the new java based software started. Anyone any good with lib rxtx?

Standard bottles would make making the caps simpler. Alcohol bottles all seem to have different caps (size, threads etc). The lazydrinker set provides a number of plastic discs with barbed fittings for the air and fluid line, and a plumbers hole cutter can be used to make holes in the cap (from a hardware store). However, PET bottles don't like alcohol (they taint) so something like polycarbonate or glass would be better.

Also website has been updated with a test video and more information.

Alec

I like your carousel idea! Simple and cheap, just the way I like it! [smiley=thumbsup.gif] Wish I could be of assistance on writing the software, unfortunately I'm more of a hardware guy and programming has never been one of my strengths.

I've been pondering different ideas on how to minimize the need to buy valves by maybe using just one, a stepper motor and an improvised mechanism to move the pathway between bottles, if I think I've got anything viable I'll post it here. Mechanically I'm aiming for things that are cheap, easy to make and "good enough" rather than perfect to get things moving along, kind of like what the REPSTRAP is to the REPRAP project, perfection can come later with tinkering and experimentation.

Getting standardized bottles shouldn't be too tough or expensive since they can be ordered from a home brewing supply source, in fact they're everywhere so you may be able to check some out in person to get ideas before buying.

Do you plan on adding forums to your website? [smiley=beer.gif] Cheers!

Actually after looking at how cheaply solenoids can be gotten on the surplus market I think I’m going to drop the whole stepper motor idea for the valves and try to work something else out. :o

Speaking of which, anyone know what/where the cheapest but still usuable solenoid valves can be gotten? Maybe I’m just over-complicating things.

Also I think I might have caught myself in a misunderstanding of what exactly the valves are supposed do in this type of system. Do they directly control the flow out of the bottles, in other words do they have fluid run through them and the bottles are under constant pressure? Or are they strictly pneumatic and control the flow of gas to the bottles, which are only pressurized when being tapped?

If they have to handle fluids directly then that shoots down a lot of my ideas. Grrrr… damn reality…

The system will work much better if the valves control the flow of fluid. Fluids are largely incompressible, and by controlling the fluid flow you'll get an output that starts and stops nearly instantaneously, and gives near constant flow rates (if you have a big enough pneumatic reservoir). Liquid valves have the advantage that they can be placed very close to the nozzle, reducing wait times. By controlling the air flow you'll get a stream that starts and stops slowly and takes time to reach peak flow and to stop. However you avoid problems with the valves getting dirty or sticking. I was thinking about a system using both, with a single liquid valve to get the on-off response and a full set of air valves, but in the end I could see no great advantage to this and a number of disadvantages.

Valves seem to be the one area that costs can't be reduced too much. If you can afford it I would recommend simply shelling out for a good set (8 or so to start with). If you find a good supply of valves or solenoids give me a shout and I'll put it on the website.

You could build your own valves out of blocks of plastic (I'm going to put up a reference for food safe plastics soonish) using a mill or you could get away with a drill press.

I'm interested to see your ideas.

P.S. There is now an RSS feed for the website news. openbar.2kmegs.com/fnews/fastnews-rss.php

Yeah absolutely, a fluid valve system directly on the bottle outlets with constant pressure would be the best setup if one has the money to do it, but I'm guessing decent valves are at least $30 a piece and for just a meager 4 valve setup we're talking $120 for those parts alone.

What I had in mind would be based on cheapo solenoids either salvaged or bought off the surplus market, springs and flat aluminum stock from the hardware store. To put it in overly simplistic terms you would have the inlet mounted in a hole in the top bar, another movable bar with a hole sandwiched underneath between that and another fixed bar on the bottom with the outlet coming up through another a hole. The pathway would then be opened or blocked by sliding the center bar back and forth using the solenoid to pull and the spring as tension to bring it back to a normally closed position.

This of course would need lots of tweaking but the materials would be very cheap, maybe $5 for the solenoid, spring and various hardware and perhaps $10 for the metal stock which would probably be enough to make several.

The problems with all this of course is the fine balance between keeping an air-tight fit and still being able to move the parts, and lubrication would probably not be possible since at least some of it would somehow wind up getting into the fluids. This would also no doubt leak at least a little unless one was really precise building it, but if the bottles are only pressurized during pours the amount leaked would probably matter little either for the pouring bottles or those which are not. Also if one were to have one main gas shutoff valve (a bought one) feeding the impromptu valve setup that would keep the gas from bleeding away needlessly between pours. In theory you could make a whole rack of these between two longer top and bottom bars.

But again the fluid valves would work a lot better, they would be much faster and allow more precise measurements, so if there's a way to adapt this idea to work with fluids effectively while not needing an engineer's degree then I think we're really on to something! 8-)

Just make a plunger valve, most of the commercial solenoid valves are this type, see http://en.wikipedia.org/wiki/Solenoid_valve

Every other automatic bar project seems to be using solenoid valves

Sealing isn’t too hard, just put some recesses for o-rings, either buy some or make them! Teflon does a good job if made with plenty of interference. Lubrication shouldn’t be a problem if you make the valve body it with a little clearance. the other option is to use a membrane seal.

I would strongly recommend food grade plastic or stainless steel, commercial grade aluminium doesn’t like food acids too much. Price difference shouldn’t be too much either.

You should be able to keep the leaks down pretty easily.

Alec

I've thought of using o-rings but I figured without lubrication (or even with) the sliding friction would wear them out pretty quick, especially when they travel over the edges of the holes. Aluminum produces a toxic salt when corroded by carbonic acid (soda) so that's another reason it isn't good for fluid valves, and as for plastic on sliding valves I'd have to wonder if excessive wear would be an issue to. Stainless steel of course would be ideal but I'm not sure what's available off hardware store shelves, I'll have to poke around and see what's available. Like aluminum Stainless would work better from a sealing standpoint because you could use lapping compound between the surfaces to wear everything really smooth and tight, I've thought of using the reciprocating action of an old handheld jigsaw to help wear them in. Here's a crude mockup of what I was thinking:

http://www.flickr.com/photos/44008864@N07/4049985934/

This of course doesn't mean working out some kind of other style of valves using plastic like a membrane is a bad idea, I'll have to chew on that concept for a little while and see if I can't figure something out.

The OpenBar Wiki is now online at openbar.2kmegs.com/pmwiki/pmwiki.php this is the place for you to get involved in OpenBar and publish your experiences, designs, tips, mods and ideas.

TornadoBoy, I'll start to put some information about making plunger valves on in the next few days. In brief, you're not going to put any significant wear on a plastic part, especially if it's ABS or something similar. Also, rings should be fine, just think teflon rings that are more like the sealing rings on your car rather than rubber o's (I think even rubber o's should work), and remember you're going to have plenty of lubrication, like a hydraulic system, it will be lubricated by it's working fluid.

Awesome! I signed up for the wiki but apparently my membership is pending or something because it's not letting me log in.

I think I'm going to build one of the el-cheapo valves I was talking about just to see how well it works, they might be useful for non-food grade applications where lubrication with grease isn't an issue. I was thinking it might be fun to try to make an Arduino controlled hamonica and those valves might be perfect for it. It would look pretty damn cool too, lots of shiney metal moving parts (aluminum is EASY to polish) that look complicated as hell but really aren't tough to make.

Weird, did you get the confirmation email? Give those valves a go, an arduino controlled harmonica sounds like a cool project, I'd love to see one built!

Alec

Give those valves a go, an arduino controlled harmonica sounds like a cool project, I'd love to see one built!

Yeah, too bad I have no idea how to play one, but it'll make all kinds of cool noises! ;D Yes, I was the kid at family gatherings whom always liked to bang the keys on the piano.

I'm still having trouble signing into the wiki, I tried signing up with an alternate email address under the user name tornadoboy2 and didn't get any confirmation message either, not sure if I'm doing something wrong or what. :-?

oops, maybe it doesn't make it clear on the site, but I disabled the confirmation email bit. Your account is be ready to go!

Alec

Really interesting project, looking forward to see how this goes. I found a source of valves for another project but I guess I can't post a link until I post once ( :-? ) so see below for details!

Here's the site I was taking about. I bought a few valves through ebay and got em in less than a week. Good variety (including low flow/gravity fed valves which most solenoid valves cannot control) and most are $13-$16 with flat rate shipping on multiple valves.

http://stores.ebay.com/valves4projects