Can Inkjet Print Head Technology Be Used for Precise Liquid Dispensing?

[ChesireCat5]

Hi, I'm new here! I'm designing a dispensing system with several liquids. The liquids need to be dispensed into a circular area about 3/8" (opening of a bottle). I'm looking for advice, and would like to share my two ideas:

1. 
   A single peristaltic pump where the tube is manually changed. Each liquid would have it's own tube.

1. 
   A pressurized system with a centrifuge and tube for each liquid. All but one centrifuge would be programatically opened to allow the liquid to be dispensed.

Details:
Not every liquid is used each time. There will eventually be about 500 liquids, currently ~100, and about 5 are used with each dispense.
The liquids cannot be cross-contaminated.
The precision must be accurate to roughly 0.025mL (about one drop of water) if not possibly better.
Some liquids are more viscous than others.
The system will be controlled by an Arduino or Gadgeteer board for the dispensing volume.

With plan A, I would calibrate the DC motor for the peristaltic pump by evaluating how many revolutions relates to volume.
With Plan B, I would find an acceptable pressure, keep the system at that pressure, and would manually evaluate how much time the centrifuge being open relates to volume. (Since the viscosity varies, I more than likely will have to calculate this time each liquid).

The hole in both of these plans are that I can foresee many problems related to accurate measurement.

I hope I have adequately explained my puzzle! There may perhaps already be a solution to such a task that I am not aware of, but Google hasn't told me otherwise.

 

[Simon Bridge]

You need to get liquid, maybe sequentially, definitely in exactly measured amounts, from any combination of separate containers into a single receptacle.

I'd do this by keeping the containers + their tubes still and move the receptacle.
Mount the containers above the receptacle in a grid. The receptacle can be moved to the (x,y) location of a particular liquid using belts and stepper motors. It is even possible to raise the receptacle once it is in position so the tube definitely sits inside the opening.

Tubes can be kept short then (helps accuracy) - containers can be piston, gas or just gravity driven - with a spigot at the bottom (to avoid drips) - or maybe just a small pump in the tube if the seals are good.
This is actually the main part of your problem.

Varying viscosity is a problem - you'll need some way that the tech filling the containers can input the required adjustment as they go.

You will need some small-scale experimentation.
Have you had a look at how industry solves the same problems?
(i.e. hospitals use a machine to automedicate patients - this is what I'm thinking of.)

I don't know the scale you want - manually changing 500 tubes seems rough - but I'm thinking either a lab-like setup where the "containers" are 5ml syringes and the "receptacle" is a small conical flask... or an industrial setup where the containers are kegs of liquor and the receptacle is a small cocktail shaker.

Another option is to feed all the tubes into the same funnel, now nothing needs to move - but you don't want to leave stuff behind on the funnel huh?

If you must keep the receptacle still, you should investigate moving the rack of containers.

 

[ChesireCat5]

Thank you Simon! I can see for a fully automated system, I must make use of a conveyor of sorts. I thought this over and decided to keep this partially automated: I can keep the bottle still and manually move the dispensing tubes.

I'm working with 4 to 16 fl oz containers to make perfume/cologne with aroma ingredients dispensed into small bottles. Most industries have assembly lines for each product, I'm looking for a more cost-efficient way of accomplishing multiple tasks in one system.

I woke up and jotted this down, I am not positive the mechanics behind it are right but let me know what you think:

1. Processor tells piston to enter syringe at an exact distance
2. Solenoid # opens, letting pressure into ingredient resevoir
3. The correct volume of liquid enters recepticle
4. A window of time is allowed for more viscous fluids to flow
5. The # solenoid closes
6. The A solenoid opens, allowing air into the system
7. The syringe returns to start position
8. The A solenoid closes
9. Repeat

 

[Simon Bridge]

Or you mix a large batch and dispense that into many small bottles ... that gets you the accuracy too.

 

[ChesireCat5]

Right, I could do that as well, , but I'm usually low on stock of at least one ingredient. Will the exact volume of the syringe dispense the exact liquid using my example? If so, then this method could be the most accurate and useful application :)

 

[Simon Bridge]

I think you can rig a syringe to dispense exact-enough quantities for what you want.
You'll have to experiment though. There are lots of options.

 

[Baluncore]

Peristaltic pumps are hard to calibrate and are temperature dependent.

Maybe consider using inkjet print head technology.