I have been toying with the idea to make a cnc machine that will inject or pour melted plastic into molds (small molds, cavities less than 1" wide by 10" long by 1" deep). I have several rough plans on the cnc machine, and really like the one posted on this site, however I do not know how the injection piece will work into the system.

I will typically be pouring 3 colors in each cavity in 3 passes (in controlled amounts along the path). I have looked at several stepper driven metering pumps that look as if they'll do the job but I am not sure how this will work as most controller boards seem to take up to 4 axis drivers. In this scenario the x and y axis will be guiding the pour nozzles, while I need 3 more controllable motors turning 3 seperate pumps.

I am fairly new to robotic automation, and am doing this as a hobby/learning project. Please advise as to any solutions you may have or guidance you may give. Is there a way to control 5-6 axis without spending 40k? or 1k for that matter! Also, if someone has done something similar I would love to hear about it.

Thanks,
-Abram

You could add a second parallel port to your computer and hookup another controller for the additional motors. Most software will let you choose which parallel port each axis is connected to.

Thanks! Now on the subject of controllers, motors and drivers, what brand/model would you suggest for this application? The machine will be approx 2'x3' Gantry style (would like to if possible be able to convert to a mill to mill aluminum molds, but that isn't top priority). I am planing on using stepper motors for the pumps and probably for the xy as well. My goal is to spend as little as possible without sacrificing too much quality...I guess that is everyones goal...

And yet another question regarding the plastic dispensing piece. I plan on using the stepper driven pumps, but would it be more accurate to build a screw injection pump (keeping in mind the small cavity size)? Does anyone have suggestions on a heater for the lines that will carry the material (length approx 5')?

Thanks again!

Hi. Interesting project. What kind of temperature do you need to maintain on the line? There are different electrical heat tapes available, but what range of temperatures you can expect I don't know.
Another thought would be a steam trace-steam has a lot of heat-5 times more than water at 212 degrees! They use it a lot in industry to keep pipes from freezing in the cold. Just copper tubing tied along the length of the plastic tube.
The cool thing about using a screw pump is that it can be a simple on/off unit. Several of the CNC packages allow control of "extra" digital or bit outputs.
What about no pump at all? It occurred to me you would want to purge the lines between colors. How? Blow them out with air. Then it struck me that you could used compressed air to push your molten plastic through the lines.
Control would be via a two-way solenoid valve that would allow pressure to be applied when activated, then vent all pressure off when deactivated. Just some things to think about. :p

Thanks for the input! The material is to be maintained at 350F. I like the compressed air idea, but would this be easy to controll small predictable amounts? Most of the molds I use are open molds that the plastic is hand poured into in up to three passes, each pass consisting of a different color and amount (cavity size is typically the size of a ink pen). The method of pouring/injecting must pour slowly as to not splatter in the open faced molds.

Also, the plastic I use sets up really fast, within 30 seconds in a ink pen sized cavity it is already solid enough to remove, though not good as it is still hot enough to damage quality. I am looking at a heating device to keep the molds warm long enough to make 3 passes on 30 or so cavities, I may still have to keep it down to just a handfull at a time.

My thoughts are to have 3 stepper contolled pump, screw, air (or whatever will work on a budget) driven dispense nozzles, each pouring a different color. These could all be mounted on the would be "z axis" of the Gantry style CNC machine, and controlled via G-code (offsetting each by the distance of the nozzle(s) in front of it). Each pass would dispense different quantities of plastic, in possibly different locations of the mold.

I would have to purge the lines after the entire run is completed, which I am still scratching my head about, but I like your idea of compressed air to do that.


There are different electrical heat tapes available, but what range of temperatures you can expect I don't know.
Another thought would be a steam trace-steam has a lot of heat-5 times more than water at 212 degrees!

The "hoppers" in this system will be similar to crockpots with temp controll and a line to the pump from the bottom and from the pump it would travel up to 5' (probably closer to 3') to it's final destination in the mold. Would I need something like the steam trace for keeping the plastic at 350F or would the heat tapes work (where to find electrical heat tapes?).

Thanks again for all your wisdom (and please keep it comin!)

Have you considered having your 3 pots of molten plastic stationary and moving your molds to each individual pot and triggering a valve.
Then move the next mold (on a X Y axis) into position and repeat, should be possible I would think. That way 3 axis would be plenty, X and Y to move into position and Z to trigger molten plastic, from a stationary position.
With this setup you would not need a moving gantry machine, but rather a fixed Z axis, and a moveable X and Y.
Just a thought......

Hi Jimini, thanks for the input!

This is an idea I have actually pondered, and am still considering. In fact my initial sketches were of a machine where the molds were on a rotary table (x axis) and the dispensing nozzle was on a y axis arm. How would I be able to control which of the three valves to open? The plastic I use at 350F is similar in consistency as corn syrup, so the idea of an open/close valve may work, but would it be possible to predictably pour measured amounts at different points in the toolpath?

I'm not absolutely convinced, but I think I will still need 5 axis controllers, one for each plastic source and two for the movement. The pots will remain stationary with a hose line from the pots to the dispensing nozzle.

Another hope of mine is to be able to "retool" this with a z axis mill spindle to mill its own molds. This is the main reason I've been leaning towards the Gantry style, not to mention the many resources, plans, and expertise this site and others provide on such a machine.

I'm with Jimini
Keep it simple, the fewer pumps, long hoses,stepper motors the better.
I say use a fixed gantry with Y axis moving your three hot pots right above your moving X table with solenoid valves on the bottom of your pots and gravity feed, easy to heat up easy to clean few things to go wrong. Now you just need 2 axis and three on off signals.

onthemarc

If the process is to be exactly the same all the time, then CNC so to speak may not be the answer. You could use a PLC and normal motors that are just on and off devices.
You could fit either limit switches or beam break switches to the different parts of the machine. You can then program the machine to do whatever function you like once it hits each limit/beam switch.
This method would eliminate the need for a computer to drive it, and also expensive software,steppers,driver cards,etc.
While I have never used a PLC b4 I know what they can do and it sounds like this is more what you need.

EDIT: A Plc is basically a computer with lots of inputs to fit sensors to. Like switches etc. And then it has lots of output relays or electronic switches. You program it to turn on and off swithches when its sensor switches are activated. Does that make sence?

Have a read of this, and tell us if this is more what you need http://www.plcs.net/chapters/whatis1.htm

You mentioned less than a 1000 bucks. Well here is a PLC for 50 US bucks that has 6 inputs and 6 outputs. Possibly you will need more than 6 inputs, but this will give you an idea http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&category=4660&item=3874231929&rd=1
You could use windscreen wiper motors from a car for your drives. You could build a machine for next to nothing. Just so long as it is the same routine all the time.

Your getting some good input-this is why I like this site so much. If we go off of the "less is more" principle, then I like the idea of moving the molds under a stationary spout. That's how they fill the torpedo cars with molten steel in the mills: a train moves the next car under a runner. This whole assembly is under the furnace, one level down. They basically poke a hole in the side of the furnace and the stuff runs down a runner, through a hole in the floor, and into the car below. Simple. Nothing to clog.
I'm pretty sure you would be hating life if your plastic hardened inside a pump or the lines-so do away with them. You could put solenoid valves at the bottom of your heated pots, that would give you the precise metering you are looking for. Get the right kind and the body will handle the heat just fine.
You could either use gravity feed or use air pressure (adjustable to adjust flow rate).
Finally, don't get confused by axis control: you only need that for speed/position control. You don't need it to turn a 12v or 24 v motor on or off-a single bit can handle that. Same for solenoid control: a single bit driving a transistor driving a relay and you can turn hundreds of amps on and off.
But, to keep things simple, an adjustable timer outside the PC would work just as well. And limit switches that detect a mold in place could trigger that.
It could be hardwired, but the other member that recommended a small PLC was right on track: think of them as small programmable logic blocks. They run almost everything in industry. There is a small learning curve to learn how to program them, but nothing major. Ladder logic (the most prevalant way to program PLCs) is fairly simple.

Great ideas! I will surely spend hours sifting through PLC related material!

Before I do though (actually while I do!), my plan is to be able to line the table with many molds for making 30 or so small objects in a single run. Same shape/size of molds for a single run but will need to change out with different types each run (even changing from open faced molds to two part molds). Also each run will require different amounts of each color plastic. Would this be doable with PLC without major reprogramming? I was also hoping to be able to store on disk all the code for the runs I am doing to easily setup for a new run, perhaps this is all up the PLC alley?

I do agree simpler is better. I will re-approach this project with the z axis stationary holding the three pots and solenoids to dispense as well as the possibility of using PLC.

You don't need it to turn a 12v or 24 v motor on or off-a single bit can handle that. Same for solenoid control: a single bit driving a transistor driving a relay and you can turn hundreds of amps on and off.

Please excuse my ignorance, but how do you send the on/off bit to the selenoids? Would this be controllable by means of G-Code?

Thank you everyone for being so helpful (and patient!) I am sure with your help I'll be up and running in no time!

Hi! O.K, as far as the bit thingy: it was somewhat of a generic statement. Depending on what software you would be using you might be able to control a few of the lines on the parallel port as on/off outputs. Specifically I was thinking of the enable outputs for drive control or perhaps a spindle on/off line. You could "steal" the spindle on/off line and use the G-code that controls it to run something else. I can't comment on what software, exactly, gives you the most options: I just don't know.
You stated you wanted to do a mill to make molds. I don't know if you have the room for two machines, but think about a dedicated mill/router set up just for moldmaking. Use the CNC stuff to run it and be happy.
A dedicated "casting" machine doesn't really need fancy axis control. Let me explain-at lunch time. I'm at work and this "goofing off" is going to me in trouble! More in a while...

Yeah, I know... I'm trying to combine two projects into one. The mill is secondary and as you mentioned will probably be best to leave it a separate machine. I come from a object oriented software development background so I'm always looking at ways to re-use objects that perform similar tasks to minimize development time/money.

Looking forward to hearing more....

If only I had money for food. Oh well, I need to lose some weight anyway. Back to business. Like you, I do programming. I'm hoping to sell artificial intelligence components I've written online in the near future. Pretty much all I use is Borland's Delphi which is object oriented Pascal.
In a prior life I did tons of PLC programming for a system integrator. Every different kind of project you can imagine: steel mills, pharmaceutical, chemical, food, and so on.
Your typical PLC (or Programmable Logic Controller) is nothing more than a dedicated processor that scans input I/O, determines logic conditions based on the inputs, and sets output I/O accordingly. One iteration or "scan" might only be a couple of milliseconds.
Over time these units have evolved and they generally have racks with plug-in I/O so you can use what is needed for the job at hand. Thus you might have 110 VAC I/O, low voltage I/O, relay cards, analog I/O, encoders, etc.
The little cheapies have fixed I/O-say twelve inputs and eight outputs, usually all 110 VAC. Each vendor has their own software for development.
There are many players: Allen Bradley, Siemens, G.E., and many, many more.
In this arena low cost is a relative term: a "cheap" PLC might be several hundred dollars, software not included-you buy that separately.
So what do the working poor of the world do?
Given that you already have a programming background why not try something fun and cheap? Look into using any of the several small single chip controllers. Lets look at the Basic Stamp as an example. Go to http://www.parallax.com/ and do some browsing. Here is a small processor, programmed in BASIC, with lots of I/O, interface boards to allow control of motors (DC and Stepper) and other things, and onboard serial ports that can communicate with your PC. And it's cheap!!!
You could either program it to do some control locally or read it's inputs and write it's outputs via RS-232 and do control from your PC. There are many serial port components available for Delphi, VB, C, C++, or whatever that would allow you, Mr. Programmer, to write your own front end to run a machine. Sound like fun?
Most industrial applications of PLCs rely on the state of inputs (limit switches, etc.) to tell them what the outside world is doing, then they react. The programs are nothing more than state machines. You would want your Basic Stamp contolled thingy to do the same: IF the machine is enabled THEN start the plastic heaters. WAIT 12 minutes, then ACTIVATE filling cycle. IF in filling cycle AND a mold is detected MOVE the mold positioner UNTIL the mold is seen by the proximity switch under the nozzle, then STOP. IF a mold is in place, OPEN the nozzle for 1.3 seconds. Blah, blah, blah.
You get the picture.
No numerical control involved, cheap, fun, and you already know how to program.
It's something to think about. :o

Thanks for the info. Could you please help me to understand the various switches I would need, and how they may be used? I can se installing limit switches to indicate when the mold is in place, but the proximity switch...will that allow me to locate the cavity?

///I'm off to parallex....

Perhaps you can tell all of us a little more about what you are doing: what you are making, mold size, mold material, etc. Is it necessary to have multiple molds or could the thingy do one, eject the part, do another with the same mold, and so on. Any pics would help if you have something to show. The real point is that there is no one way to go, and possibly several good ways. More info helps everyone provide input.

The primary purpose of this machine is to pour soft plastisol into aluminum molds to make soft fishing lures (worms, grubs, frogs, etc...). The mold size is typically 4"x8" , which would accomidate 4-6 worms, a frog or two, etc... but can potentially be 10"x15". Most molds I use are open face molds, but some are 2 part molds. Neither of which require high pressured injection. The style of lures I pour consist of multiple layers of colors which give a worm a bloodline or holographic scale effect to mimic a rainbow trout, etc...

When the plastic sets it is very soft, so I think automated ejection would be difficult, though I'm always open to ideas. I could have it pour one mold at a time and manually extract the product, that's not ideal but it would work.

As I mentioned before I initially sketched a rotary table design that would basically pour one mold at a time, then turn the table to the next mold. This seems like it would work well with breaker switches and PLC or a Basic Stamp. I diverted from this approach when I dreamed of having it capable of converting to a mill to mill its own molds but am re-considering as I absorb more info from you experts here.

See attached for example of a couple of my open face molds.

Ah, now I see! So even with a single mold you need to move some to pour worm #1, then worm #2, and finally worm #3. Let me ask you this:

When pouring manually do you have to move the stream around to fill one item in the mold or can you pour in one place (i.e. the worm's butt) and have it fill the whole worm?

Can the pour rate (volume per second) vary much if you are filling a thick frog versus, say, a skinny worm? Would you want this to be adjustable or just pick one safe rate for all the mold types?

Does the consistancy of the plastic vary much when at the working temperature? In other words is it corn syrup at 350 degrees and tar at 325, or is it more or less the same "goopiness" over a wide range? What is the minimum working temperature? If you are going to use a timed injection, then a constant flow rate (and viscosity) will be important. With a measured volume injection this isn't as much as a factor.

Can you remelt this stuff once it has set? If the machine cooled and something clogged, would simply reheating the assembly "fix" it?.

What about non-aluminum molds? I've got some weird teflon muffin molds that will survive in the oven, yet flex and allow you to pop out the muffins easily. Could something like this be done to allow repetative filling of a single mold?

Sorry I have so many questions but these things intrigue me. Hope you don't mind. :eek:

When pouring manually do you have to move the stream around to fill one item in the mold or can you pour in one place (i.e. the worm's butt) and have it fill the whole worm? This is a variable. Some molds you just pour in one spot and it fills the mold, especially in the two piece molds. More often than not you must pour different amounts along the cavity. For instance the worm in the picture would require that you pour heavy at the "head" and very light towards the tail. This is especially true for multiple color passes.Can the pour rate (volume per second) vary much if you are filling a thick frog versus, say, a skinny worm? Would you want this to be adjustable or just pick one safe rate for all the mold types?
This I'm not sure of. I do know I need to adjust the amount of plastic poured at specific points in the cavity, but I don't know if that should be movement controlled or volume per second controlled. Pouring fast may cause plastic to splash or spill. Again, I'm not sold on either technique.
Does the consistancy of the plastic vary much when at the working temperature? In other words is it corn syrup at 350 degrees and tar at 325, or is it more or less the same "goopiness" over a wide range?I don't have the specifics on temp ranges, but as long as it is heated to 350 it will maintain the corn syrup consistancy to at least 325. I'll have to verify that for sure though.Can you remelt this stuff once it has set? If the machine cooled and something clogged, would simply reheating the assembly "fix" it?.Yes.What about non-aluminum molds? I've got some weird teflon muffin molds that will survive in the oven, yet flex and allow you to pop out the muffins easily. Could something like this be done to allow repetative filling of a single mold?I have used plaster, RTV and aluminum molds and have found the quality of the product is much better with aluminum. The lures have a glass like finish from the aluminum molds which is very important with translucent type colors.Sorry I have so many questions but these things intrigue me. Hope you don't mind. Not at all!

Hi! A dead connection has kept me offline. Wasn't blowing you off, promise!

Still thinking about how best to approach your task, but I don't have time to write more now.

I'll try to get back to you later. Good luck!

We have a hot glue machine at work that would be just the thing for you. It has adjustable temp pot, a pneumatic over hydraulic pump, heated hoses and a heated, air solenoid valve controlled discharge gun. I cannot for the life of me remember the name of the manufacturer, but I can get it Monday if it will help. I'm sure they sell seperate parts as well as complete systems. The name is something like Nordstrom. You migh try Google for hot glue equipment. It's a very well built, and trouble free system, and it runs your temps easily. When you're done with it, you just let the stuff harden up in it, then when needed again, just reheat.

Thanks for the suggestion! I had actually run across some parts on eBay for Nordson hot melt gun a while back but didn't realize what it was. I just went to their website and couldn't believe my eyes! Their machines would be perfect...however I'm assuming they are a bit out of my price range, but they have great looking machines that I can get many good ideas from!

It looks as if their XYZ system is very similar to my original plan (if equipped with a hot melt gun) of a gantry style machine. I would need to create a heater wrap for my hoses if I were to use pump/valve dispensing, but I think that could happen. I'm looking into some stepper driven pumps that are relatively inexpensive as well as solenoids valves.

If I went with stepper drivers for xyz, how difficult would it be/what would it consist of to control 3 on/off fluid pumps and 3 solenoid valves? Check out this video of the Nordson XYZ system and you'll see exactly how I want this machine to end up, only with three hot melt dispensers: http://www.efd-inc.com/media/index.html.

As far as how to control your stuff, I'm not the one to ask. I'm more of a commercial cnc guy. I would question the need for stepper driven pumps when pumping hot stuff, can you isolate the heat from the stepper? Or, if you do use stepped pumps, I don't think you'll need solenoid valves. Calculate the time the pump needs to run for your discharge amount, then back the motor up a few steps to eliminate the dripping. We use stepper pumps where I work also, and that's how we do it, in packaging of heavy, hard to deal with liquids or semi solids. Or if you're gonna use solenoid cutoff valves, just use a continous pressure system. You don't need both. I know our hot glue machines have the temp controls on the pump (constant pressure) for the tank, hose and the gun. I also remember that the company does sell replacement heated hoses of different lengths. They hook up with standard AN fittings, and an electrical plug to the machine and gun. Hope this helps.

Thanks for the info. So you recommend cutoff valves in my case? I thought perhaps I would need the steppers to increase/decrease pumping speed slightly to vary the amount of plastic being poured at different points in the cavity. I suppose I could increase/decrease speed of the xy axises to achieve this.

How would you recommend I make the constant pressure system? Thanks for the tip on the heated hoses. I'll be searching for a bargin to get my machine fitted with these (or something similar).

I would'nt say I reccomend cutoff valves, if you're going to use stepper driven pumps, I would start without cutoff valves, try reversing the stepper's rotation at the end of the cycle to eliminate dripping (if that's a problem) and then use cutoff valves as a last resort. The less things you have to heat up to use, the simpler it will be, and simple is almost allways better. If you choose to use a constant pressure system, then you'll need cutoff valves. The constand pressure can be obtained thru an air over hydraulic, or electrical type of pump, even pressurized containers will work. The hot glue machine uses constant pressure, I believe the pump is air over hydraulic, but it only pumps when the cutoff valve is opened.An on demand sort of thing. That still seems the easiest to me, having your heated resovoirs, with the pump built into the bottom of them, so you only have to heat one thing. Good luck

Those videos from that link are really damn cool. :D

When I was a lowly restaurant worker, I used to help the repair-main when our dish-machine would break down. They used Peristaltic pumps (like this guy (http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&category=46547&item=3878187801&rd=1&ssPageName=WDVW)) to give precise measured amounts of cleaning fluid, etc. Would that type of pump, with heated hose, work in your application? I would imagine it would make precise metering much easier (they are Stepper Driven), and since you don't need high-pressure it makes sense. Plus, since they are stepper driven, I don't see why you couldn't write G-Code that says "Move X-Axis blah blah, Move Z-axis blah blah, Move Y-axis (Pump) blah blah" you get the idea. Just a thought...

Thanks again to all who are helping with this project! I really appreciate all the input I am getting. I would be lost without this place!

At this point I am back to my original idea of using pumps. With the xyz moving the mold and nozzle into place and then somehow controlling the three pumps (maybe on an "a" axis or whatever the 4th is called) one at a time, perhaps some kind of re-tooling sequence (switch to pump1-1st pass, switch to pump2-second pass, switch to pump3-third pass). This theoretically would be controlled by an on/off bit as discussed in an earlier post.

The pumps will be installed directly under the heating pots, which will be stationary above the z-axis with hoses (somehow heated) to the nozzles mounted on the z-axis. I don't have it figured out yet how to shield the pump motor from the heat...any suggestions?

I have a question now about stepper drivers/controllers. Searching ebay I found this interesting product that appears to be a 4 axis stepper driver/controller combo:
http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&category=71394&item=3877761462&rd=1
Does anybody have any experience with this one? I'm curious as to if it would work for my machine. The price is right, but all too often you get what you pay for.

Also, when looking at stepper motors, what should I be looking for as far as specs? As you know I am new to this, so I'm not sure what would be too small, or just a waste of money.

I don't have it figured out yet how to shield the pump motor from the heat...any suggestions?

Did you look at my post? :p

I looked at the link. The main thin I saw is that it is full or half step-no microstepping. And you are limited to unipolar use. On the flip side the current limits are ample. I would wonder how they are handleing their current limiting if it isn't a chopper type drive.

I use the Geckodrives and love them. Go to www.geckodrive.com. Other use Rutex and like them too. There are other options-I'll try to look a few up this evening.

One thing you have going in your favor is that you are not trying to push a milling bit through material-your strictly positioning. This being the case, you can get by with less than the guys building routers or mills. You migh be fine with some stout little NEMA 23 motors (NEMA 23 refers to the motors frame size). Next size up is NEMA 34, then NEMA 42, and so on. For you more power will simply mean better acceleration and deceleration. If you don't need the speed, your OK with smaller motors.

At work and gotta go-I'm gonna get myself in a bind otherwise!

Bye for now.

JavaDog,
You're right. That pump would work great to shield from the heat. Would it be able to dispense small enough volume of the plastic? Probably would. Now if I can find 3...

Thanks Evodyne for the info.

JavaDog,
You're right. That pump would work great to shield from the heat. Would it be able to dispense small enough volume of the plastic? Probably would. Now if I can find 3...

Thanks Evodyne for the info.

Yeah, since they are driven by a stepper motor, I would imagine that they are pretty flexible when it comes to dispensing. They are used in the medical industry quite a bit as well.

You may even be able to make your own, as their design is pretty straight forward.

Okay, got my motors on the way. I went with powermax II Nema 23 motors, primarily because I found them cheap on ebay. Hope they will work!

I'm looking into a controller board and found a couple, but I'm not sure about them. This one:
http://www.hobbycnc.com/driverboards/4aupc/4aupc.htm
and either the 3 or 4 axis one from here:
http://www.xylotex.com/

Any suggestions? The price is right for this project, but will it work?

Take a look at STAMP microprocessors (lots of Google links). Pretty cheap easy way to set up logical motor control based on switch/timing inputs. I really like them. They're very configurable and there's a lot of good user projects posted for you to steal code snippets from.

What plastic are you planning to use? A lot of plastics suffer from quite a lot of shrinkage. Especially the crystalline ones (PE, PP, nylon etc) Expect quite a lot of sink in the middle of your part. 1"x1" is quite a lot of cross section for thermoplastics.

MadMax,
Thanks for the info.

The plastics I use are a liquid plasitisol that sets to a rubbery plastic. The plastic does shrink as it sets, but I should be able to compensate for that with timing.