Questions: CNC for use in life sciences lab automation

[xavier_romerez]

I have spent some time lurking in this forum and would now appreciate some opinions on my proposed use for a CNC machine.

Basically, the idea is to build a 3-axis machine capable of being used to automate a number of procedures that are routinely performed in biology research labs. A number of commercial systems are available for this (for example see https://www1.fishersci.com/Coupon?ci...5194&details=Y) but the price tag is depressing - besides, I want to build it myself.

Some proposed details:

1. Small footprint (1m x 0.5m x 0.5m)
2. Sterilisable / Easy clean construction material
3. End effectors - syringe pump (liquid dispensing) and seperate gripper (moving items around on the table)
4. CHEAP!!!!

Opinions appreciated on the following:
Drive mechanism - Screw or belt?
Structure - Full gantry or cantilevered?
Motor requirements (dare I ask...?)
Feasibility

And, of course, anything else you would like to add....


All opinions will be valued.

Regards,

Xavier

[WhiteTiger]

You should be able to cantilever with no problem unless you expect to pick up something heavy such as isotope containers

The 8020 build method might work for you. Their plastic bearings and aluminum rails would seem to be perfectly suited to your need to be sterilizable and seem to be plenty accurate for pick and place operations. You may have to swap out for stainless fastener hardware such as screws and nuts; I'm not sure if they offer ss hardware.

Myself, I'd box out the X axis (four equidistant rails at the corners of a square profile) and run a stainless ballscrew in the center of the profile.

You will probably need to provide for hermetic motor enclosures of some sterilazable material, and avenues fo recirculating cooling air to the enclosures.

Motor size is going to depend on accelleration rates and traverse speeds factored around the masses and moment arms. Some of the people here can probably give you ballpark values off the top of their heads.


Tiger

[kaye7877]

Saw one of these go on ebay about a couple of months ago for virtually nothing.

[vacpress]

i hate to say it, but i think this is an 'easy' project in many ways...

1. you can get lots of prefab usefull bits off ebay... light duty actuators, new surplus pneumatic equipment and scientific surplus..

2. the load to move is very, very light in comparison to even a dremel\plywood cnc cutter..

3. the thing can probably be 'homed' alot easier durring cycles than a cnc router, for instance, therefor tolerances dont get more and more out-of-whack the longer the machine runs.

4. while accuracy IS required, it isnt the .005~ that the machine-tool types are looking for.. i think that if you want to pick-n-place samples, and do precision pipette-type measuring\metering, you can get by with something much looser... maybe even .05, for instance.. just home it between cycles to keep within that margin and not have it add up..

5. the drivers can be something much simpler than you would want on a gantry router(for instance) xylotex would be very nice for this... but i bet a half stepping allegro 5804(or whatever the PN is) chips would work well also... and that way the 3 or 4 axis driver board would only cost $50~

6. To do the precision measurement, one option is (i forgot the name) using a stepper with an offset cam and a piece of silicone tube... these are often available new on ebay, and for $10-50/each they are a great deal consiering the same thing might be around $500 from the manufacturer.
Possible Electronics for Such a thing (3 axis assumed)

1. 3 half-stepping driver boards, good upto maybe 1amp
2. 3 stepper motors.. probably 50oz/in~ minimum.. anything over 100 would probably be overkill
3. 6 limit\home switches.. i would go optical, more precise and repeatable
4. power supply - another benefit of the lower power motors and driver boards.. a PC supply may work.. Or, better, look for a smallish 24v, 3amp +/- linear amp (no switching)
5. Breakout board, wire, etc.

Cost Estimate if you build driver boards: $150.00
Estimate if you buy xylotex(or simialr): $250.00

Mechanical

You could go cheaper, but I think 8020 is made for these applications.
1. 8020 frame material from the ebay '8020garage' seller
2. linear bearings, either 8020 stuff (im sure it would work great for this) or IGUS would also be a good option. The simple IGUS linar slides would be affordable and easy to sanitize
3. measuring hardware - ebay! find new surplus.. it may take a few months, but you can probably make an amazing scientific device with ebay surplus..

Ok. I am sick of typing. I have more to say on this, but I am supposed to be 3D modeling in my cubicle, not typing rapidly!

one last thing, i gather you are a student? i dont recomend 'stealing' software, but i have seen labview on bittorrent.. labview would be 'the way' to control something like this.. i think.... if i were to do it, i would either write everything 'script style' in qbasic, or i would use labview...

Cheers.

[unterhaus]

Student version of labview is cheap, and many universities have a full license available for use. On the other hand, I don't think it would work with step and direction drives, you need a VI that will make the motion work.

[lakeside]

The Most Expensive Part Of This Project Is Other Than A Studant Project Would Be Fda,osha And Liablity Insurance

[vacpress]

if it isnt for bonifide biomedical research, i am sure all manner of goofy contraptions have been used to automate experiments.. I mean, not all scientists have a bunch of engineers on hand. I dont think they do, anyway...

besides, this person probably wont build such a thing anyways..

[unterhaus]

most universities don't have too many osha visits. We recently had a huge effort to come into compliance with the EPA guidelines which are a pita.

The stupid crap that engineers can come up with is pretty amazing. Personally, if it doesn't look good, I will not let the students build it. I've seen too many problems start up and get a life of their own just because something didn't look right.

[sdantonio]

Just a suggestion, move the arm to the tubes as in the picture, don't move the tubes to the arm as many fraction collectors do.

If you are working with sticky solutions or solutions that can dry and leave crud behind (I love using technical terminology like "crud") you will have a problem with the tractor mechanish sticking. This was a major problem for some of the peptide synthesizers and fraction collectors when they missed or overfilled a tube. I once managed a biophysics lab with a bunch of peptide synthesisers. One moved the amino acids to the syringe (and was constantly needing to be cleaned), the others had a moving needle on a gantry and had stationary vials, they worked great every time.

[vacpress]

i think 1 reason to leave the tubes\samples stationary is the sensitive nature of many experiments..

however, i bet we will engineer this whole idea out, and this person will never show back up.

[xavier_romerez]

On the contrary, vacpress, I'm still here and greatly appreciate all of the opinions provided so far. It's just that I felt I had little to add to the discussion and was trying to absorb all the comments!

Vacpress & WhiteTiger, you seem to have confirmed pretty much what I was thinking on both the mechanical and electronic side of things. One question that remains is the material / diameter to use for the rails on each axis in order to provide sufficient support for the mass to be moved. All parts will be COTS.

One interesting approach may be to colocate two of the steppers (Y? and Z), with the Z axis being driven by means of a square-profile shaft/rod passing through a pinion (pinion is free to slide along legth of square rod). This could also serve as one of the guides for the Y?-axis. This should minimise the mass of the entire assembly. Any thoughts?

Regarding control software, this will be vendor neutral (self coded).

Some facts:
I am a biomedical graduate student.
This is a hobby project / personal challenge.
Intend to move the arm, not the tubes.
This WILL be built once the plans have gone through several rounds of revision, although there is no time restraint on just when that will be.

[vacpress]

well. to get closer, figure out a machine spec... figure out:

1. what is the application(s), specifically
2. what is your budget in time and money
3. what bits can you get for free?
4. what is the maximum mass you need to move,
5. how fast?
6. accuracy? (.05, or .002)
7. technical skills, tools available, resources, etc..

this is how i do everything. not ideal in some ways, but it allows a single person to define a complex problem and begin to tackle it...