How to control liquid flow ?

[roborg]

Hi,

I am new to electronics and robotics and I have been going through various tutorials to understand basics. I am trying to build something that drains few drops of water (in general, any liquid) from a (small) container at specific pace and time. The pace of draining (ml/sec) and time are programmable parameters. I am having difficulty in figuring out how to do the mechanical and electronics part of this. I am thinking of this like a syringe which is "somehow" interfaced with a motor which in turn is controlled by micro controller.

micro controller---- stepper motor ---- container

Essentially every now and then (whenever timer fires) micro-controller turns the stepper motor a bit at required pace (based on ml/sec leakage required) which in turn applies pressure on the syringe/container that leaks the liquid.

My questions are:

1) As the movement required is small, I am wondering what kind of motor is desirable here ? Is stepper motor the right one for this ? If not, what else can I use. Also, there seems to be few types of stepper motors out there.

2) what type of micro-controller can be used here ?

I am looking for something with a very small form-factor here as I am dealing with milli-litres.

3) Regarding the container, I need this to close every time once required amount of liquid is pushed out (so that there is no seepage). Any suggestions on how to build or where to get such a thing ?

4) How exactly can I interface this container with the motor ? For pro-typing, I can use a syringe but it is not clear how to connect the motor to it.

5) What other electronic components would I need to get this started ?

Thanks in advance,
RO

[Soeren]

Hi,

1) As the movement required is small, I am wondering what kind of motor is desirable here ? Is stepper motor the right one for this ? If not, what else can I use. Also, there seems to be few types of stepper motors out there.

What is it for?
What is the viscosity of the liquid?
What is the approximate pH value of the liquid?
What is the min and max amounts each squirt?
What is the min and max intervals?

(Now please don't respond with "I'm a paranoid inventor and don't wanna reveal the needed parameters", because that sort of make it impossible to help. 

2) what type of micro-controller can be used here ?

I am looking for something with a very small form-factor here as I am dealing with milli-litres.

Any!
If you need really really small, then shoot for a PIC10F2xx, it's in a SOT23-6 housing (same size as an SMD transistor).

3) Regarding the container, I need this to close every time once required amount of liquid is pushed out (so that there is no seepage). Any suggestions on how to build or where to get such a thing ?

Depends a lot on the parameters I asked you above.

4) How exactly can I interface this container with the motor ? For pro-typing, I can use a syringe but it is not clear how to connect the motor to it.

If there is a reasonable budget, then you can buy such stuff ready made.
If this is a hobby project (which I kinda doubt), take a look at eg. how blood dialysis machines work and scale down.
Are you familiar with infusion counters (the machines that're used for counting how fast and how much of a bag of medicine is infused into a persons blood stream)?

5) What other electronic components would I need to get this started ?

If you have to ask that, you're probably better off throwing some money at a solution, because you would have to learn building the hardware part, learn programming the controller, buy or build programming equipment for it etc.
Pretty soon, the expensive consultant/designer for hire seems cheap in comparison. And then you add the value of the time you have to spend on it.

[nottoooily]

I don't think you need to close it to prevent leaking. As long as the outlet hole is small enough it should just stay there. Might evaporate tho.

Threaded rod to drive the syringe?

You might also do something like a fuel injector, with a constant, regulated pressure behind a valve, which you open for a known amount of time.

[dunk]

to add to Soeren's list,
what accuracy do you need in the volume? ie. do you need an exact sample size?

the correct pump for high accuracy sealed applications is a Peristaltic pump.
a small diameter flexible tube is passed between a "cog" and a flat surface.
as the "cog" turns it's teeth pinch the tube closed. a small gap is left between the "teeth" which moves along the tube as the "cog" turns.
http://images.google.ie/images?q=Peristaltic%20pump
http://www.answers.com/topic/peristaltic-pump


dunk.

[roborg]

Hi,

Please look for [RO] for my responses inline ...

Hi,

1) As the movement required is small, I am wondering what kind of motor is desirable here ? Is stepper motor the right one for this ? If not, what else can I use. Also, there seems to be few types of stepper motors out there.

What is it for?
What is the viscosity of the liquid?
What is the approximate pH value of the liquid?
What is the min and max amounts each squirt?
What is the min and max intervals?

(Now please don't respond with "I'm a paranoid inventor and don't wanna reveal the needed parameters", because that sort of make it impossible to help. 

[RO] This is for use in a chemical lab. I am interested building only a prototype to demonstrate the feasibility of such a working system and the liquid used for the prototype will be water (at room temperature). Min. amount is 1ml and max. amount is 15ml. Regarding the interval, this doesn't really have to be periodic but pre-configured (like 1ml at 7am one day one and 5ml at noon on day two or at multiple times on each day at specific timings etc.). Again, since this is a prototype once a day at a specific time is good enough. All of this logic can be controlled by the software, so I am guessing that it should not be a tough task.

2) what type of micro-controller can be used here ?

I am looking for something with a very small form-factor here as I am dealing with milli-litres.
Any!
If you need really really small, then shoot for a PIC10F2xx, it's in a SOT23-6 housing (same size as an SMD transistor).

[RO]  Ok thanks, I will take a look at this. Please point me to some online resources on this.

3) Regarding the container, I need this to close every time once required amount of liquid is pushed out (so that there is no seepage). Any suggestions on how to build or where to get such a thing ?

Depends a lot on the parameters I asked you above.

4) How exactly can I interface this container with the motor ? For pro-typing, I can use a syringe but it is not clear how to connect the motor to it.

If there is a reasonable budget, then you can buy such stuff ready made.
If this is a hobby project (which I kinda doubt), take a look at eg. how blood dialysis machines work and scale down.
Are you familiar with infusion counters (the machines that're used for counting how fast and how much of a bag of medicine is infused into a persons blood stream)?

[RO] I can spend couple of hundred bucks for this prototype. So, if I can buy something ready-made and reusable, that would be great so that I can focus on rest of the system.

I haven't heard of infusion counters, but looked at some patents on infusion pumps/systems like the one here:

http://www.faqs.org/patents/app/20080255517

Other than that there is not a whole lot of information on how to build one.

5) What other electronic components would I need to get this started ?
If you have to ask that, you're probably better off throwing some money at a solution, because you would have to learn building the hardware part, learn programming the controller, buy or build programming equipment for it etc.
Pretty soon, the expensive consultant/designer for hire seems cheap in comparison. And then you add the value of the time you have to spend on it.

[RO] I can take care of the programming (c, c++, assembly language, java, scripting languages etc.) as I have been doing that for years. I just need proper tools and development environment to load the software on the ucontroller. For the embedded software that I develop, development environment is provided by the board manufacturer itself and I am hoping the same is true with these micro-controllers.

Thanks,
RO

[Soeren]

Hi,

Just dropping a link for now:
http://www.harvardapparatus.com/wcsstore/ConsumerDirect/images/site/hai/techdocs/Pump%20Selection%20Guide.pdf
This file might be a good read for you.

[roborg]

Thanks for the link. Looks like they have programmable syringes, except for the form-factor (they look big) this is what I am looking to build, more or less. These might be expensive, I will give them a call and see.

Thanks,
RO

Hi,

Just dropping a link for now:
http://www.harvardapparatus.com/wcsstore/ConsumerDirect/images/site/hai/techdocs/Pump%20Selection%20Guide.pdf
This file might be a good read for you.

[roborg]

Hi,

Thanks for the response. Can threaded rod mechanism provide accuracy ? or does it depend on how fine grained threads are ? Are there any tutorial robots online that use this mechanism ?

I was also looking at ways to convert rotary motion to linear motion and came across "rack and pinion" - http://www.flying-pig.co.uk/mechanisms/pages/rackandpinion.html. Has anyone tried this ?

Thanks,
RO

I don't think you need to close it to prevent leaking. As long as the outlet hole is small enough it should just stay there. Might evaporate tho.

Threaded rod to drive the syringe?

You might also do something like a fuel injector, with a constant, regulated pressure behind a valve, which you open for a known amount of time.

[Soeren]

Hi,

Can threaded rod mechanism provide accuracy ? or does it depend on how fine grained threads are ?

Maybe and yes.
The rise (not sure it's the right word in English, the center distance between 2 consecutive "valleys") of the tread could be eg. 1/10 mm or 2mm. The former will give 20 times as little linear move each revolution, so this affects the precision of course.
Whether it can provide accuracy, does depend on what you define as a large enough accuracy, but imagine a single revolution of a 1mm rise.
With a 200 step/revolution stepper motor (or a geared down DC motor with a resolver of the same resolution), a single step is around 5µm (you cannot get this repeatedly from single steps, but 10 steps will push the syringe close to 50µm with a reasonable repeatability. How much that equates to in liquid measure depends on the diameter of the syringe, so it should be selected for the purpose.

I was also looking at ways to convert rotary motion to linear motion and came across "rack and pinion" - http://www.flying-pig.co.uk/mechanisms/pages/rackandpinion.html. Has anyone tried this ?

The accuracy/resolve of a rack and pinion system will depend on the size of the "teeth". with a distance of 1mm and eg. a 10 tooth wheel, one full revolution will move the pinion 10mm, so accuracy will be much worse.

If you go with a screw to push the syringe, a "bed" for the syringe with a treaded nut mounted in the right place (a sort of shim in between the screw and the syringe will make it more stable) will go a long way. The motor should attach to the bed and a spring or similar to keep the syringe open' without the screws influence, makes the entire mechanical assembly fairly simple to construct.
Material should probably be surgical steel, as this is very inactive and can stand acids, alkalines, alkohol etc. making it easy to clean.
 

[waltr]

The rise (not sure it's the right word in English, the center distance between 2 consecutive "valleys")

Just a FYI Søren, the English word is 'pitch'.

The stepper motor and threaded rod is what I envisioned and wouldn't be that hard to build.

[roborg]

Hi,

May be this question belongs to "Mechanics and Construction" forum,  but since this is for the same topic, I am posting it here.  Please let me know if you want me to post this elsewhere.

Please look at the attached picture, which is a top level view of what I am designing. Essentially I would like to use one motor to control 4 syringes (each of which can be injecting liquid at different rates and times in any given day). Here I have four chambers (syringes for now) S1, S2, S3 and S4 with some liquid (just water for now) in it, with outlet (not shown in pic) pointing outwards. The disc shaped thing is just to hold them together (or you can think of the 4 syringes as cylindrical chambers inside the three dimensional disc). In the center is the shaft of the stepper motor (so seen from top, motor is beneath the disc structure). Now, I need to drive the 4 syringe plungers using the stepper motor.

Initially I was thinking about 4 threaded rods with different pitches so that the liquid can be injected at desired rate but I am not sure mechanically how to connect these 4 threaded rods (connected to plungers of the syringe) to one motor shaft. Any idea how this can be done ? The shaft itself very small for 4 rods to be attached to it directly. I was thinking of connecting a bigger gear and then drive the threaded rods somehow.  I can also see that this method is less flexible with respect to the following - whenever the motor turns liquid is injected from all syringes. We cannot, for example, choose only liquids from S1 and S3 should be injected but not from S2 and S4 at some point in time. Ignoring this drawback, can you think of how I can get this to work ?

To fix the drawback, I think I need to connect the motor shaft and threaded rods (or gears attached to them) dynamically instead of connecting them all the time. So, when liquid needs to be injected from S1 and S3, only S1 and S3 will be connected to motor shaft. I couldn't figure out a way to do this. Any suggestions ?

Another design I thought about was, each chamber will be connected to valve with 4 incoming pipes (coming from the 4 syringes/chambers) and one outgoing pipe (similar to what is shown here - http://www.ouqiuvalve.com/Q635-china/Ball-valve-Three-way-ball-valve.jpg). Each incoming pipe will have a gate and at any given point of time, only required incoming pipes will be connected to outgoing pipe (by opening a gate). I am not sure which design is better as the aforementioned drawback still exists.

I was looking to see if there are any programmable valves or pumps esp. suitable for this kind of application. I see some solenoid valves, but not sure if they are suitable or not. If anyone has any experiences with them please share.

Thanks,
RO.

[Soeren]

Hi,

Just a FYI Søren, the English word is 'pitch'.

Thanks! 
This is just one example of a word that has several meanings in Danish. I knew "rise" was not the right word, but in Danish, "pitch" as well as "rise" is both called "stigning" (with a silent "g", pronounced something like "stee-ning") - allways nice to learn

[coldwaterak]

I think the best approach would be to use four different solenoid valves all on the same input (pressure) rail.

There are tons of various dc solenoid valves out there that could be fitted/mounted to a home-made manifold for exactly what you are trying to accomplish.

For instance with a block of aluminum and a drill press you could port your manifold.  With a little bit of guess work you could drill the ports for the valves according to viscosity and with the right amount of duration your valve could shuttle and allow enough liquid for a drop to enter into the pressure rail and it would flow out.  If you wanted several of the liquids to come out at once or in sequence just program ladder logic.

I know this is vague and over-simplified...  but just hope it helps you along.

Here is an example of what I'm talking about...   keep in mind that there are much cheaper variations out there.
http://www.wandfluh.com/fileadmin/user_upload/files/A_Dok/reg_1_2/1_2_26_e.pdf