BRBU Builds a bar bot

[HeK]

I was thinking that with the collective skill set within BRBU, we can build a bar tending bot! Perfect to unveil at the next ACEN.

Something that we can load with a number of core liqueurs and mixers but can still prompt the user for additionals.


I'm thinking along the lines of a raspberry pi, using a webserver and python to control a PWM I2C board that runs a series of peristaltic pumps.
Database full of recipes and the user interface is done on a tablet.


Ideas? How many products should we try to dispense?

[Didzo]

Ideas? How many products should we try to dispense?

When I played around with HPLC machines a while back, I imagined that they could probably be re-purposed as extremely expensive and inefficient cocktail makers. I think your idea of a series of pumps, each attached to a different type of booze or mixer would make most sense though. Also, call it the Pootis and have the TF2 sound clip play every time someone orders a drink.

[Eightball]

There was a kickstarter recently for something along these lines: https://www.kickstarter.com/projects/som...escription

Seems like a neat side-project. Let me know if there's any way I can be of support.

EDIT: Pre-orders are open for shipping this summer. If this project gains traction, consider if the cost can be kept below 429USD, plus the value of experience.

[FlyingMongoose]

I think we can do this actually. Plenty of us have the programming knowledge.
And plenty have general engineering knowledge.
Could make it interface with a mobile app. Want to throw something up on Trello for some project plans? I use trello for work so if you do add lbabuder to it.

A cocktail I'd like to put in it is an espresso martini but adding chilled coffee might be difficult.

Name for the device:
brb, uninstalling bartender

Sent from my SCH-I545 using Tapatalk

[HeK]

First, we should do some basic system analysis.

Should it mix? I think having n-number of spouts lined into a glass should be good.
Should it have low level detection on the reservoirs? We can meter when dispensed. Give the operator a 'fill to line and click ok' option?
How many things should it hold? No idea. Need help here.
How much of things should it hold? 1L? 2L? Less for some, more for mixers?
Ice?
Cup size selection? Yes! and a drain grate.
Drink menu limited to whats in the system, or include the wetbar?
Random button? Duh.
Sound effects? It's been voted in!
220v? Fuck off Europe.
Glass placement detection? Possibly....


What colour should we paint it?

[Gasman]

I can help with numbers for the fluid mechanics of the piping and pump systems when it comes to that. "Borrowing" some ideas from that kickstarter project isn't a bad idea, either, since the concepts behind it sound solid.

A cocktail I'd like to put in it is an espresso martini but adding chilled coffee might be difficult.
Sent from my SCH-I545 using Tapatalk

It might be possible to install a refrigeration unit that chills the fluids as they are mixed/flow through the piping, eliminating the need for ice or chilling the fluids beforehand. I don't know how small and efficient refrigeration units can get, which may impact this idea. What about having the entire robot be constructed out of a mini fridge?

Should it mix? I think having n-number of spouts lined into a glass should be good.

Honestly, it all depends on how much agitation the machine will cause when combining the ingredients. My gut personally tells me it should mix them at least partially, since just dropping them into a glass may not be enough.

Should it have low level detection on the reservoirs? We can meter when dispensed. Give the operator a 'fill to line and click ok' option?

Yeah, but I think that should be an afterthought, at least until the base design is created.

How many things should it hold? No idea. Need help here.

What are the average number of ingredients in a typical drink? IMO, it should be able to hold that many, plus a few more for other drinks. Maybe make the holders modular, like in the kickstarter? It would solve the issue of refilling for different drinks by us preparing more containers beforehand, and swapping them out when necessary.

How much of things should it hold? 1L? 2L? Less for some, more for mixers?

I dunno, smaller volumes for more specific ingredients, and larger volumes for more commonly used ingredients?

Ice?

See above.

Cup size selection? Yes! and a drain grate.

Yes.

Drink menu limited to whats in the system, or include the wetbar?

How about having buttons measure out a chosen amount from each container, and have the containers labelled with what is in it? That way, we are only limited by what we have on hand, and not what is programmed into the system.

Random button? Duh.

Sound effects? It's been voted in!

220v? Fuck off Europe.

Glass placement detection? Possibly....

No idea.

What colour should we paint it?

Iron man. Make it gaudy and give it the personality of Jarvis.


Also, the pump graphic in the kickstarter is confusing me for two reasons. First, it appears that the pump is working backwards, but I'm not familiar enough with that type of pump to make that judgement. Secondly, shouldn't that design on the blades of the pump lock itself up as it spins?

[HeK]

That's the same type of pump that makes up a automotive supercharger. The fluid is transported on the outside as the middle is occupied by the paired blade.

[FlyingMongoose]

This is meant to be a low cost small design endeavor right? The majority (in my experience of a family who drinks a lot of mixed drinks with a father who at at least one point in his life was a bartender) of Most mixed drinks have 2-4 ingredients.

The espresso martini my dad makes involves espresso, bailey's, chocolate something (can't remember what it's called), and vodka. A basic martini is gin and vermouth. Flavored martinis are usually actually vodka and a mixer.

I think 6-8 reservoirs can provide enough variety But i think we can do 10 or 12. Configured however someone wants to configure it. But also think about the number of drink combinations as the reservoir number goes up.

Also remember 1 shot is one ounce. (Weight sensors? Or measurement triggers?) Weight sensors can confirm the presence of a cup.

Or in the case of a lightweight cup what about some sort of infrared or visual trigger that the system can detect?

Also, cup size selection? We aren't always going to have the right rock glass or martini glass.

Sent from my SCH-I545 using Tapatalk

[Gasman]

That's the same type of pump that makes up a automotive supercharger. The fluid is transported on the outside as the middle is occupied by the paired blade.

Fascinating, and makes sense. It never occurred to me that the space in the middle was always occupied by the blades.

[Eightball]

It might be possible to install a refrigeration unit that chills the fluids as they are mixed/flow through the piping, eliminating the need for ice or chilling the fluids beforehand. I don't know how small and efficient refrigeration units can get, which may impact this idea. What about having the entire robot be constructed out of a mini fridge?

I think refridgeration would have to happen before the beverages are mixed. Even small amounts of water/alcohol need a pretty substantial amount of time to chill unless you can maintain a very high temperature gradient (IE not gonna happen with ice or anything we can reliably get our hands on) or do something to increase conduction area (wat)

[quote author=HeK link=topic=7513.msg284979#msg284979 date=1432264440]
Should it have low level detection on the reservoirs? We can meter when dispensed. Give the operator a 'fill to line and click ok' option?

Yeah, but I think that should be an afterthought, at least until the base design is created.[/quote]
Agreed.

[quote author=HeK link=topic=7513.msg284979#msg284979 date=1432264440]
How many things should it hold? No idea. Need help here.

What are the average number of ingredients in a typical drink? IMO, it should be able to hold that many, plus a few more for other drinks. Maybe make the holders modular, like in the kickstarter? It would solve the issue of refilling for different drinks by us preparing more containers beforehand, and swapping them out when necessary.[/quote]
If our experience last weekend is any indication, 6 canisters should be adequate. 2 liquor bases and 4 mixers could make 8-10 cocktails off the top of my head. More canisters though would allow an exponential increase in drink variation.

[quote author=HeK link=topic=7513.msg284979#msg284979 date=1432264440]
How much of things should it hold? 1L? 2L? Less for some, more for mixers?

I dunno, smaller volumes for more specific ingredients, and larger volumes for more commonly used ingredients?[/quote]
Considering that this will probably get transported from Winnipeg, i'd advise 1L. Don't bother with more (bulk) or less (bitters? Put that shit in yourself)

For the other bells and whistles: Cup size? IE take drink formula and multiply by X? Sure. Paint or style it as janky as possible. Like sticks and woodglue janky.

[Eightball]

RE: Cooling; how robust are these pumps? Can they overcome 5ft of hydrostatics? Then let's keep the canisters in an iced cooler and ground level and just feed up. How much volume would this leave in the tubing? Is that a problem?

[FlyingMongoose]

I'm asking if we should detect cup sizes and if so mix accordingly if not enforce a standard size (the latter being easier the former making it more universal but harder to implement)
Of course using the infrared beam method with 3 heights allowing a cup to break a certain beam would indicate cup size.
Place these across from each other in a recess then we don't have to worry about cup position either.
Sent from my SCH-I545 using Tapatalk

[Gasman]

RE: Cooling; how robust are these pumps? Can they overcome 5ft of hydrostatics? Then let's keep the canisters in an iced cooler and ground level and just feed up. How much volume would this leave in the tubing? Is that a problem?

From what I remember, larger pumps can get rather expensive for more power, but I don't have the specifics on prices of pumps vs. power. It is very possible to tweak the pipe diameter and change the flow rate to accommodate for a weaker pump, of which I can calculate, but I need more specifics before I can work on that, such as the power of the pumps and overall height of the system plus height from ground level.

Edit: I just had a thought: what if we don't use canisters, but just have the bottles of alcohol/mixers and have tubes that run from the system into the bottles? that would save space on canisters and just require the system itself and a bucket of ice water along with the alcohol.

I'm asking if we should detect cup sizes and if so mix accordingly if not enforce a standard size (the latter being easier the former making it more universal but harder to implement)
Of course using the infrared beam method with 3 heights allowing a cup to break a certain beam would indicate cup size.
Place these across from each other in a recess then we don't have to worry about cup position either.
Sent from my SCH-I545 using Tapatalk

I like it!

[StolenToast]

I'm asking if we should detect cup sizes and if so mix accordingly if not enforce a standard size (the latter being easier the former making it more universal but harder to implement)
Of course using the infrared beam method with 3 heights allowing a cup to break a certain beam would indicate cup size.
Place these across from each other in a recess then we don't have to worry about cup position either.
Sent from my SCH-I545 using Tapatalk

I don't think it would be that hard, but just out of practicality I think it could just as well have three size buttons.

[FlyingMongoose]

[quote author=FlyingMongoose link=topic=7513.msg284988#msg284988 date=1432267641]
I'm asking if we should detect cup sizes and if so mix accordingly if not enforce a standard size (the latter being easier the former making it more universal but harder to implement)
Of course using the infrared beam method with 3 heights allowing a cup to break a certain beam would indicate cup size.
Place these across from each other in a recess then we don't have to worry about cup position either.
Sent from my SCH-I545 using Tapatalk

I don't think it would be that hard, but just out of practicality I think it could just as well have three size buttons.
[/quote]

I was thinking more along the lines of being more than a few drinks in with intent to consume more. An automated system might be better in that case. Also infrared sensors are not that expensive.  Or invisible lasers etc. Buttons work too though

Sent from my SCH-I545 using Tapatalk

[Didzo]

First, we should do some basic system analysis.

Should it mix? I think having n-number of spouts lined into a glass should be good.

No. The simpler, the better. A mixing function would require a wash system, which adds unnecessary complexity. A single line linked to a single fluid with individual spouts is the way to go. If a drink requires vigorous mixing, then a tumbler can be used instead of a glass.

Should it have low level detection on the reservoirs? We can meter when dispensed. Give the operator a 'fill to line and click ok' option?
Yes on the low level detection. I'm not sure how that would be accomplished in the simplest manner. Maybe a buoy of sorts with a line attached to it. When the level is too low, the tension in the string activates a low level signal. It automatically disables the selection of any recipe requiring that fluid until the tension is released.
 
How many things should it hold? No idea. Need help here.

???

How much of things should it hold? 1L? 2L? Less for some, more for mixers?

If you use a low level warning system that has a minimum level warning but no maximum, then it really doesn't matter. Large volumes for stuff like mixers and low volumes for things like bitters. I don't think the reservoirs need to be fancy either.

Ice?

Nah. Ice can be added manually to user preference. I think having ice be automated would add a lot of weight and cost. It needs to be refrigerated, refilled, and you'd need to deal with melting.

Cup size selection? Yes! and a drain grate.

Agreed.

Drink menu limited to whats in the system, or include the wetbar?

Limit it to what's in the system.

Random button? Duh.

Just set a minimum and a maximum volume for each ingredient and have the total volume of the generated mixture not exceed the total volume of the selected cup size. Also, a log of all randomly generated recipes should be kept.

Sound effects? It's been voted in!


:-*

Glass placement detection? Possibly....

A simple IR sensory would do the trick. It would help avoid spills.

What colour should we paint it?

Moist pink.

[Didzo]

Also, there should be flashing lights involved. This is very important.

[FlyingMongoose]

Also, there should be flashing lights involved. This is very important.

And something should spin or move. It's not high tech if something doesn't spin or move.

Also, for low reservoir could use a simple float and lever mechanism. Once the side opposite the fulcrum from the float gets to a certain point it could hit a contact and as long as this contact is "ON" then this reservoir is not available. This could just be a push switch (low resistance to pressure as long as the float is heavy enough). Anything near the reservoirs should be liquid proof and sealed.

ASCII Rendition (underscore represents contact and line represents fulcrum):
___
-----|-----o

Sent from my SCH-I545 using Tapatalk

[HeK]

Going with peristaltic pumps may be our best option as they are accurately metered (based on run time and PWM) and keep the fluids separate from the drive mech. This aids in cleaning.
http://en.wikipedia.org/wiki/Peristaltic_pump

Adafruit has one for $25 that is rated for 100 ml/minute. I also see these on eBay/AliExpress for ~$10 each.
Add the cost of replacing the hose with one that is food-save silicone.

They are easy to control, such as run for 18 seconds per ounce. However this may make drinks take 40-60 seconds to be prepared. I'll see if I can source a higher flow and low cost unit.

Also remember 1 shot is one ounce. (Weight sensors? Or measurement triggers?) Weight sensors can confirm the presence of a cup.

That is one fluid ounce, or 30 ml. Since the mass of each beverage will be different you can't use weight. With an accurate pump, you know how much as been used.

RE: Cooling; how robust are these pumps? Can they overcome 5ft of hydrostatics? Then let's keep the canisters in an iced cooler and ground level and just feed up. How much volume would this leave in the tubing? Is that a problem?

I like this idea. We can keep mixers in a cooler of ice. However, even with common 2mm hose, that leaves quite a lot of volume in the hoses. Go go silicone?

Of course using the infrared beam method with 3 heights allowing a cup to break a certain beam would indicate cup size.

We would need an array of sensors to measure the hight, width and taper of the glass. What if the glass is clear? Or conducts IR?
What about a solo-cup button? Shot glass? Other common sizes?

Edit: I just had a thought: what if we don't use canisters, but just have the bottles of alcohol/mixers and have tubes that run from the system into the bottles? that would save space on canisters and just require the system itself and a bucket of ice water along with the alcohol.

While this will reduce the cost and the mobile size of the machine, it does mean open bottles unless we devise a universal fit stopper (or a series of stoppers. Available at beer making stores) with both an outlet for the supply hose and a vent/airlock. Remember that alcohol likes to vaporize at room temperature.
It does allow easy storage of non-empty bottles after use. If we build the machine with a manifold bank on the back and just connect hoses as needed, that may be the most simple solution.

I was also thinking of mason jars. They are available in 1L and 2L sizes. They are clear to see the level. They have a two part lid so you don't spin the hoses when you unscrew. They are heavy glass though.

And something should spin or move.  It's not high tech if something doesn't spin or move.

If you want to be the one cleaning up afterwards...

[FlyingMongoose]

The other downside to the pump from bottle is you have two options when it comes to refill. Basically you need to be able to either enforce the same size bottle and possibly liquor type per hose or allow the user to change that information as it becomes necessary. Long term maintenance makes the reservoir the better option. Plus you could even make one reservoir a pre-mixed drink for more variety.

Along with the fact that we can standardize the reservoir size and as hek said a good pump can tell you (or you can decide) how much a hose uses. Which removes the necessity of a float mechanism because you can have it calculate based on usage and turn off that reservoir after a certain amount was used . I'd probably still want the float as a back up though.

Also the moving light up thing doesn't have to do anything or affect the drink in anyway. It's just a trope that I feel is necessary.

Sent from my SCH-I545 using Tapatalk