BRBU Builds a bar bot

[Gasman]

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

I was thinking that we could use this type of pump. It is ideal for what we want, but there are two possible issues with using it. First, the tubing used in it will break eventually from use, and need replacing after a number of uses. It should still be cheaper to replace the tubing than other pump prices are ($25? Wow). The other issue is will the pump be able to handle the head necessary for five+ feet of tubing?

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.

The system will need to be cleaned regardless, since the internal components will be exposed to the fluids. Easy solution is to run a cleaning agent through the tubes before replacing the reservoirs with other liquids.

[Didzo]

[quote author=Didzo link=topic=7513.msg284992#msg284992 date=1432274231]
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.

The system will need to be cleaned regardless, since the internal components will be exposed to the fluids. Easy solution is to run a cleaning agent through the tubes before replacing the reservoirs with other liquids.
[/quote]

Just to clarify, I was thinking of a wash step between each drink as something to avoid. If the drink is mixed in the system, then you don't want residue of whatever you just made to carry over into the next drink you're making.

[Gasman]

[quote author=Gasman link=topic=7513.msg285000#msg285000 date=1432321196]
[quote author=Didzo link=topic=7513.msg284992#msg284992 date=1432274231]
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.

The system will need to be cleaned regardless, since the internal components will be exposed to the fluids. Easy solution is to run a cleaning agent through the tubes before replacing the reservoirs with other liquids.
[/quote]

Just to clarify, I was thinking of a wash step between each drink as something to avoid. If the drink is mixed in the system, then you don't want residue of whatever you just made to carry over into the next drink you're making.
[/quote]

Agreed

[HeK]

[quote author=HeK link=topic=7513.msg284997#msg284997 date=1432294592]
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

I was thinking that we could use this type of pump. It is ideal for what we want, but there are two possible issues with using it. First, the tubing used in it will break eventually from use, and need replacing after a number of uses. It should still be cheaper to replace the tubing than other pump prices are ($25? Wow). The other issue is will the pump be able to handle the head necessary for five+ feet of tubing?

[/quote]

It's a positive displacement pump, so there is no priming required and no backflow when the pump is stopped. It should be able to overcome several hundred feet of hose.

I've found a good source for 100ml/minute pumps for $10 USD each and these fantastic looking 500ml/minute food-grade pumps for $30 USD each.

The PCA9685 control IC that I am looking at supports up to 16 channels per, so 4x 500ml for mixers, 10x 100ml for spirits and 2x channels for lights?

I guess pump selection will depend on mixers and spirits in use.

[FlyingMongoose]

From what I'm seeing we have a really solid concept here so far.  Can we get an early draft drawing from someone?

Sent from my SCH-I545 using Tapatalk

[WaR1761]

Maybe some other ideas.

http://gizmodo.com/meet-the-robots-who-w...1706316473

[FlyingMongoose]

No drawings?

Sent from my SCH-I545 using Tapatalk

[HeK]

Weekend = housework
Monday = actual work.

I'll spend some time today in Visio.

[CaffeinePowered]

I leave for a weekend and we lay the foundation for building an automated bartender, nice

[FlyingMongoose]

We've laid the foundation, yes, but we need a quick mock up of it. I mean we have the component parts more or less decided and an idea of what we're building and how to built it.

But we need some plans people! Or...

HeK you don't have to do the drawing, someone else can do it. I'm a TERRIBLE artist, like my drawing would make you vomit.

I just think a general design visual would be useful to all parties involved.

Can we get the first post organized into the key parts and components though? Like what the final decision for each item for it is?

Oh... and to be quite frank... it seems like we're building a Booze Keurig with multiple reservoirs and no heating element based on what is being designed so far, anyone else getting that same picture in their head?

[WaR1761]

From a brief chat I had with Hek I realized a challenge, water dilution, texture, and chilling of the beverage. Chilling can be overcome by keeping the ingredients cold to begin with, dilution and texture I am lost on how a vacuum pump system would achieve this. That is unless you mix/stir post dispersion.

https://www.youtube.com/watch?v=kxPtyFqvTgA

As for the menu focus on drinks that don't have a shit ton of ingredients(More is not always better). I would start with this list. If you wanted to maximize the number of drinks to make can also play with http://www.cocktailbuilder.com/

[Dtrain323i]

quick mockup

[FlyingMongoose]

quick mockup

Close, but I think there's a few aesthetic features that might need modification.

On the shake vs. stir item, ultimately our device goal is intended to dispense the booze more than anything. Unless we have a secondary "mixing" reservoir added for when the drink is dispersed. For example; from source reservoirs into a mixing reservoir then from there to disperse into the cup. This, however, creates the cleaning between drink problem again. So I'd say we'd emphasize people stir things after dispersal.

[HeK]

I sent my motor controller circuit draft off to a team of electrical engineers to laugh at, I should get some feed back soon.

Re: Stir vs shake vs mix. How about a few cup size options when selecting drinks: Shot glass, high ball, solo cup, mixer, shaker.
The bot can be a glorified measuring system, let the end user worry about the rest. Same concept when it comes to bitters, garnish, etc.

[FlyingMongoose]

I sent my motor controller circuit draft off to a team of electrical engineers to laugh at, I should get some feed back soon.

Re: Stir vs shake vs mix. How about a few cup size options when selecting drinks: Shot glass, high ball, solo cup, mixer, shaker.
The bot can be a glorified measuring system, let the end user worry about the rest. Same concept when it comes to bitters, garnish, etc.

That's pretty much what I was thinking too. If we make the machine do too much that creates too much additional steps and potentially moving parts, and as is the general rule with everything; the more steps the more points of failure possible. I think the only auto-mixing possibility is the secondary reservoir option.

Of course... if we're controlling with a Pi... we could do alternating outputs from each original reservoir, that can give a partial mix at least.

I'm trying to think of the code that would involve that. So far I have this (which I know is probably way way wrong, but am getting kind of tired and am not mathsing hard enough)

Code:

if(reservoir-request-count > 1){
    var drinkratio = 1 / reservoir-request-count;
    var currentfill = 0;
    for(new i = 0; i < reservoir-request-count; i++)
    {
        OpenDispensing(drinkratio,reservoirID);
        currentfill += drinkratio;
    }
}

This code is way off, because ultimately the idea is that each time the loop iterates it selects a different reservoir to come from (I actually think it needs a nested loop)

Someone else can fix it or I will later.

[StolenToast]

Yeah just think about this, this is the proof of concept.  It doesn't have to make GOOD drinks, it just has to make drinks.  Smaller increments are easier to do than big leaps.

[at0m]

Pardon my late arrival and idea vomit.

Size/number of drinks? It should all fit inside a 4.3ft^2 minifridge or smaller, or on top of/next to said fridge with a harness. Next to would allow for the dispensing portion to overhang a counter into a sink, or have the drain grate have a hose which can go into the sink. Perhaps some nozzles for mounting room-temp bottles on top that then feed inside where all the guts are. This also allows for 'freezables' to sit in the freezer section of it (like jaeger) for proper storage temperature but still work (well, depends on draw straw design, but still).

Ice? nah, do that shit yourself. If it was in a full-size fridge, then maybe, but otherwise nah.

Mixing? maybe a powered shake mixer that sits in the cup and can optionally stir while dispensing, depending on the drink selected. Alternately, mix it yoself lazyass.

Number of beverages inside really depends on volume of fridge selected, bottle storage orientation (are we filling bags or tanks inside of it? Or just mounting the bottles upside down? If upside-down, gotta put airflow ball valve into the stopper, and something to let air into the fridge as liquid leaves it). I'd suggest hoppers with bags (similar to the inside of a milk or juice dispenser), but the alcohol may attack the plastic. Would have to research whether this will happen with tanks as well. Sticking with bottles may be the best way to do it.

Multiple nozzles and pumps is definitely the way to go, unless gravity feed and volumetric measurement is acceptable. (see: 'optics' used in bars in Europe, or things like this: http://www.amazon.com/Rotary-Bottle-Stan...B003TSK9G0). May simplify design if optics will work, then all you'd need is an actuator (RC Servo?) and hose on each leading down to the mixing point. This would decrease the resolution of mixes down to the size of the optic, which is usually 1oz (so mixed shots would not be a thing). This could allow for expandability, but would be more difficult to refrigerate. Maybe skip on refrigeration altogether?

[HeK]

Maybe skip on refrigeration altogether?

Lets keep the scope creep to a minimum. Though I do like the idea of making able to mount above a mini fridge.

At this point, I am thinking it to be a fancy PC-controlled series of DC motor pumps, backed with a database, accounting and UI.

[Greatbacon]

Got bored, so I made a super simple system architecture diagram based on HeK's previous comments.
https://docs.google.com/drawings/d/1Liia...pH6Y0/edit

I think a couple of simple, preliminary use cases would help give us a high level vision for framing the initial work.

For example:

• I put in a cup, select 2 ingredients to put together, and the machine pumps equal parts of them into the cup.
  
• If an ingredient is running low, display a warning.
  
• If an ingredient runs out, it should no longer be available for selection.
  

[HeK]

Recipe database select based upon available ingredients.