I'm unclear after reading all that what the goal now is. Just a copy of a makerbot? Or something that does everything?
Thanks for the feedback. This is the purpose of review of the initial high level requirements, I'm glad the process is functioning correctly..
The overall, highest goal is to make a base design for a motion control machine.
The base design can be optional reinforced to perform as a CNC machine, or made less expensively as a 3D printer.
The base design can accommodate beefier motors for CNC, or cheaper motors for 3D print.
The design of the head is to include a standard mount so that an appropriate head may be attached at the users option. This is not limited to:
Small CNC head (made a Dremel type tool if the material is small and soft)
"Standard" CNC head (what ever that may be at this time)
Plastic extruder
Ceramic mix extruder
Hot Wire cutter
Laser light (for exposing photo sensitive etching material)
Pen or other marking utensil
[anything the user wishes to mount]
My thought is that the common denominator of CNC, 3D, plotter, etc is the bed, arms, motors, and motion control software. This is the base design.
Any particular machine can be built by extending the base design to that specific purpose.
Also the physical size is to be scalable, so (the first one I build) can fit on my desktop, and the one my friend builds can take up the whole garage. But the software interfaces, etc etc are recognizably similar.
The intent is not to re-make the maker-bot; unless that's what the maker;bot already does, in that case the work is done and I just go buy a maker bot and save space on the forums.
I don't think the makerbot is a universal base design intended to be expanded, but maybe I didn't read the article carefully (happens all the time to me).
It is only my opinion, but the first thing that should be established is a reliable printing solution. The success of this project is ultimately dependant upon the printhead, the media, and resolution. Perhaps several solutions would be best for different types of endeavors.
Secondly, I would suggest that you set a size parameter, that way everyone is on the same page. What is the largest item that everyone wants to be able to print? This should help establish guidelines for X, Y, and Z axises, software parameters, etc. Be realistic though, because the larger you go, the higher the cost. Of course everyone may have different views and/or play money to invest in the project. Once again it is only my opinion, but the height of the printing will ultimately determine the complexity and cost of this project, from a mechanical standpoint. Of course this pertains mostly to a CNC mill type of setup.
IMHO, the size parameter is already determined for some printing processes, like extrusion, by the material properties. Models printed open air really can't exceed 4x6x6 or so. Larger models will suffer significant shape distortion, or must be built extremely slowly. The commercial extruders demonstrate this.
For some other processes, overall manageability of the material determine model size. The raster based ones require a considerable amount of material be on hand for the process to work. A cubic foot is about where those play well.
Some use layers of material, spools, and they are limited by spool size, and the practicality of cleanup and handling waste. Current commercial solutions have a 6-8" spool width, and a similar model size.
The older "dot matrix" wax type printers offered half a cubic foot or so build space.
A base machine that can address roughly a cubic foot would be sufficient for moving forward. The commercial solutions and processes involve printing partial models for these reasons.
A base machine that can address roughly a cubic foot would be sufficient for moving forward.
Although I am not that familiar with existing commercial processes, I would have to agree that a machine that addresses a cubic foot would provide a good modeling space.
I'm really not sure what the point of making your claim was. I honestly and truly could not believe what I was reading. I could not imagine under any circumstances making a claim such as that whether true or not, especially when I would not be able to back it up.
Please stay and contribute, or if you don't want to then don't, but I think you have to stop throwing your toys out of the pram every time you are in some way challenged. I've read many of your posts and the only thing telling me we would be losing a great deal is you. Let your advice and your wisdom and your published progress speak for you, saying you have a patent does not prove a thing and it is strange to think it does. You know who you are, you know you are great, let us find this out for ourselves from measurable output not here-say.
This is why I respect PhiPi, Mike Green, Heater, Jazzed etc etc etc. I have absolutely know idea about their credentials (I think Mike is a doctor!) but my god these guys know their stuff and I know that simply because they provide good information which is well backed up, they make reasonable arguments and try to be helpful.
It is only my opinion, but the first thing that should be established is a reliable printing solution. The success of this project is ultimately dependant upon the printhead, the media, and resolution. Perhaps several solutions would be best for different types of endeavors.
Absolutely right I think, as in CNC machining you should work back from the "cutter". Actually back from the part. Decide what parts you REALLY want/need to make, what material they might be made out of and then to what resolution. That will determine the material and the process you might use and might even inspire new process creation. Then you are also in a position to work on the machine as you know what it needs to move and how fast and with how much force.
I also think the prof is right, motion control is the first propeller task, it is needed for all machines.
I think an interesting start to this would be to implement the protocol of an existing open system like a reprap/makerbot. It is fairly well documented.
My only reqeuirement for size (so far) is that it fit on my desk, which is 24" deep (and 48 inches wide, but I would like to leave some space for crayons, etc).
Lets say the requirement is a 609mm (square?) maximum foot print for the "small" option.
Further, lets say that a scaled up or large version is out of scope of the current effort. My friend won't commit his garage to a giant DYI CNC until I at least show the little 3D printer works.
So let me ask about the differnece between a 3D printer and a CNC at this "desktop" size. Would there be a substantial cost to construct a 24" (609mm) square foot print device that is capable of 3D printing and CNC? For example, could we limit it to software materials and use something along the lines of a Dremel tool as a cutting head? Or is "changeable heads" a poor idea?
I'm very intrigued by this concept now. I really want to make one, but lack the time and mechanical expertise to make my own.
Perhaps the easiest way to do it is to buy the print head and basic hardware from the Thing-O-Matic people and then copy their design for the rest of it.
I'd almost buy the whole thing, but I have a problem with paying up front and then having to wait 7 weeks. Also, I'd rather make it Propeller powered...
They sell most of the parts to make it on their website...
I think using some existing print head is a great idea. There are lots of fickle things about printing that have taken a long time to reach viability. The RepRap has fairly coarse resolution, and that's the big limiting factor in it's overall utility.
A machine that could do both could be very worth building. Lots can be done with even just a dremel...
So let me ask about the differnece between a 3D printer and a CNC at this "desktop" size. Would there be a substantial cost to construct a 24" (609mm) square foot print device that is capable of 3D printing and CNC? For example, could we limit it to software materials and use something along the lines of a Dremel tool as a cutting head? Or is changeable heads a poor idea?
You must decide what it is you actually want to be able to make.
Let me assume you want to print plastic with FDM and machine plastic, wood, G-10 and other soft materials. You could use a dremel but they are not great, there are some small routers that are reasonable (Bosch, Kress etc) and you can get a better version of the Dremel from Proxxon. To make the machine I'd recommend a fixed gantry router with moving platen. These are simple to construct, stiff and given the small size of the required work envelop the loss of working area compared to the moving gantry type is fine. Drive with belt drives using stepper motors, enough torque for the light cutting and plenty of speed.
Or you could use my 2-axis parallel kinematic concept and add a platen beneath it with an adjustable height. Fast (very low inertia and axes do not carry each other), stiff, novel but a little more tricky when it comes to motion. Again use belts and perhaps 4 leadscrews for the z-axis platen but with some decent bearings. This design is also good for raster work as both motors work together in one axis giving extremely high speed.
But I say again, decide what you want to be able to make. For me I actually want to make precise patterns for casting as I already have access to FDM and I have a CNC, for you I'm not sure, your call!
I like the idea of the basic concept to have a basic design that can be varyied.
CNC-milling of aluminium or even harder (in a double sense) CNC-milling of steel combined with productivity and a precision of 0,05 mm requires different technologies than a 3D-printer.
So if the basic design should be as broad as from a few 100 dollars 3D-printer up to a steel milling CNC with a resolution of 0,05 mm or even 0,02 mm the hardware-design has to to be developed from the upper end
which means CNC-milling. To achieve this resolution a lot of parts have to be very precise (ball-screw and ballscrew-axles that are polished and not rolled) and a very VERY STIFF construction.
Motion-control should include PID-control of position, speed and acceleration/deceleration. Which would be best achieved by servo-motors instead of stepper-motors.
So I guess such a broad scope would be hard to achieve.
If you make compromises about
- precision only 0,1 mm,
- productivity (milling only 0,1 to 1 mm thin) using just a dremel (instead of a 1000 W spindle) press start button at 7:00 AM a single steel-part is finished at 7 PM
- start-stop driving of stepper-motors
the options on how the design could be get much broader.
Another aproach could be to design central functions modular:
developing standards for interfaces like
position-feed-back (endswitches, reference switches, encoders)
GUI or GUI-elements
command-set for communicating with a motion-controller
To keep it standardised all the special things of the included machine-types 3D-printer, laser-enlighter, laser-cutter, CNC-mill etc. must be considered.
I haven't thought very much about all these machine-types. But to satisfy all users it has to be very universal.And maybe with this broadness the universality
changes into a brake because there are too many parameters or too many commands. I want to emphasise this MAYBE could happen. If somebody has good ideas
how to achive a "compact" universality please post your ideas as this will forward the project a lot.
I don't see why the motion control should include PID, we can leave that to the driver. I'd say it is more sensible to have step and direction output which can then go to a standard stepper motor drive or a servo drive such as those from Gecko. Actually even high end AC servo drives will often accept step and direction, my Yaskawa drives do. You can certainly machine steel without that level of complexity.
The whole encoder feedback thing, well that can be an object, use it if you want it. I've never bothered and never missed it. If I had it I am sure I would love it, if in production I might well need it.
Do you suppose that the object based nature of the propeller might take care of the modules naturally?
I imagine a decent motion engine, a g-code parser, encoder reader, limit monitor as objects.
... lack the time and mechanical expertise to make my own...I'm thinking about a 24" cube for the box...
Same here. I hope to see more about idbruce's 3-axis CNC assembly.
I'm liking the changeable head idea, and allow printing and CNC (maybe even CNC detailing on the object that was printed?)
I need to find out more about "fixed gantry router with moving platen" and "better version of the Dremel from Proxxon".
Drive with belt drives using stepper motors also sounds right. (so precion of .01mm is too strict?)
Is there a picture of "2-axis parallel kinematic concept and add a platen beneath it with an adjustable height"?
Since this is a first attempt, I would propbably settle for a printer that makes a candle out of wax, or a CNC that makes a small square candle out of a big round candle.
A machine that can function as a 3D printer that can make a fork and spoon, and as a CNC that can work jewelers wax and drill a PCB would be sufficient. Anything more is gravy, but gravy will be accepted.
So far, we have narrowed the size, now we are looking at functions. Is this getting more reasonable or further away from "possible"?
Cool ! Complete different design to what I have seen so far. So thank you for sharing this idea.
I know: asking questions is so much easier than answering them or compete by showing another concept developed on my own.
So if I ask it is not to be skeptic it is just out of curiosity. Can you already say something about the precision that is achievable with this concept?
Of course. Well in terms of resolution you can imagine that if both actuators move together you get one axis of motion (the long axis). In this direction the resolution is as per your actuators, so if you use belts which is most likely that comes down to your pulley size. On this test machine which is made from epson printer parts with arms machined on my CNC the pulleys are small, about 6mm diameter I guess, so you can say for half stepping a 200 step motor about 6*pi/400 = 0.047 mm resolution. The other axis can be considered as being related to the differential position of the two actuators. Here is an image that shows the variation in resolution across the working area:
You can see that the grid you get is not square but that should not worry you as this effect would be occurring with 0.04mm between the lines. Resolution is about the same as the actuator's in the middle but decreases towards the edge. You can also see the working area is not square, I imagine it best to pick a square or rectangle in that area and consider it the working area. It is one disadvantage of parallel machines. By changing the spacing between the rails you can also change the working area, you can even toggle the mechanism in some cases on purpose.
Here are some sketches I made. I was considering a laser cutter:
And here is why it is a reinvention only
Notice that you can use a parallelogram to create a plate for the attachment of the tool that does not rotate.
For more info including the fairly simple mathematics see my page:
As both Graham and Bruce pointed out, it would be best to work from the "tool" backwards through everything. It seems as though the primary goal is a 3D printer, so start with the "print head" technology. It isn't easy or the the guys that made up that rep-rap thing would have come up with something better by now. I have not looked into it for a while, but the resolution of their output was far less than desireable. Someone suggested a resolution of .01mm (.0004"), I suspect that would be very difficult to achieve. Even a lower resolution of .1mm (.004" - slightly thicker than paper) will be tough but at least it will result in a decent product.
The mechanical framework and kinematics to move the print head is the least of the problems at this point. There are dozens (hundreds, thousands?) of designs for the mechanicals. Be it a: bridge mill (often incorrectly called a fixed gantry), bed mill, gantry mill, knee mill, articulated arm, or whatever, there are many choices to work with. If you can't achieve a decent print method, the framework and kinematics means nothing.
The drive system for the axes is another, "worry about it later". You can use: belt drive, rack and pinion, lead screw, or ball screw - many proven designs to work from. The resolution you want to achieve here is dictated by the resolution of the print head. Having greater resolution on the axes than on the print head serves no additional function. Using: PID, open loop, closed loop, servos, or steppers - who cares at this point! An open loop stepper drive system would be very reliable for the 3D printer and even for many light duty machining applications. Stick with something easy and economical - open loop steppers fits the requirements nicely.
Motion control software - why spend time worrying about this at this point in time. I mentioned that the MACH software and a PC can handle what you need for 3D printing (and much more). The demo version is free and it would be very suitable for use during testing. If you want to incorporate a PROP into the system, worry about that after you have solved the tough engineering aspect of the project which is the print head.
My whole point is this, you have to solve the problem of the printing method first, that is the unknown in the system. All the other required elements of the system have already been done and there are multiple, pre-existing solutions available.
If this is going to be some sort of kit or something to be sold, an even better starting point is to determine what the build budget should be - what is it worth to the end user. If this thing ends up costing $20,000.00, who wants it at that price?
I agree a lot of this is just worry about it later stuff, but only if you are experienced. I've just been trying to put some flesh on the bones to provide some ideas of how things might be for those who have admitted having less experience, particularly as the question of whether a combined machine could be created and how big and what it might look like was on the table.
Assuming the printing method is just FDM then Mach or similar could be used, for other printing methods this might be less practical (syncing a jetting head with motion).
So if the real issue is the head and the mechanics and the prop might be another way to implement the control at a later date it does rather sound like we are on the wrong forum at this stage???
I can tell you exactly how the FDM head in our uprint works, very simple. It pushes the plastic rod using pinch rollers through a heated tube which has a small nozzle. Overall dimensional accuracy of the part in x,y comes down to the machine pretty much, you can get decent tolerances. Minimum feature comes down to the width of the extruded filament, this also affects print speed. Resolution in z also comes down to the filament thickness, the uprint does 0.254mm thick layers. The print heads are carefully temp controlled, the nozzle is purged and wiped (with brush and squeegy) on a regular basis and the printing environment is also heated. On the Uprint which has two nozzles the same drive motor is used for both nozzles, it is changed over by driving the head against stops on the machine (a clever reuse of the axis motors).
Graham
p.s. Sorry for using the wrong term for a bridge mill (or portal mill as they seem to be sometimes called.) I'd be interested to find a glossary of the accepted terms so heinous crimes like this can be avoided in the future.
Looking at the FDM head you have at your disposal, does it look like something that can be made by the average person here on the forum? If not, do you know of a source for them? The .254mm (.010" about 3 thicknesses of paper) thickness is okay but like yourself and others have stated, that resolution depends on what people want to make with these things. Do you also happen to know what the diameter of the nozzle is? Do you know why the enviroment is heated? Possibly to help the plastics adhere better?
I would think that most people would want this 3D printer to print with plastic as the resulting product would be useful. I have seen some other 3 printers that use some sort of white powder that is sprayed (ink jet printed) with a color binder. This created a 3D color model of the object but I don't recall the durability of the object. Perhaps the others in the thread might want to voice their ideas of what they plan to make with this thing.
As for the wrong forum, I don't know. Eventually it could lead to a propeller related project. Heck, even the controller for the extruder and other heating elements could be propeller controlled - we would need some electronic gurus to help with that stuff.
No worries on the bridge mill miss-use, the hobbiest groups everywhere use that term, drives me nuts, why not call it what it is. As for a glossary, I can only reflect back to the engineering text books from years gone by. Perhaps there is something on-line somewhere. Another source would be machine tool company websites, one would hope that a company selling a machine tool for a quarter million dollars or more would know what to call it.
For me I actually want to make precise patterns for casting
*If you want to create precision parts from metal, build a milling machine.
*If you want to create prototypes or casting patterns, build a 3D printer.
Of course, there are many grey areas in between.
Graham was talking about printhead technology. I once read an article where these students created a 3D printer and used sugar as the printing medium. The article said the results were very satisfactory, but then again, I did not hold the finished pattern in my hand. There are options, and I think we should research to find economical, but reliable print technology. I am sure plastic printing would be the best solution for sand casting, however I believe it would be costly and time consuming to implement. I think I would lean more toward wax printing and lost wax casting.
EDITED: But there is a downside to lost wax casting, you always lose your pattern, whereas in sand casting, you can reuse the patterns.
I want a design that is stiff enough for CNC on my 24 inch footprint motion bed.
I want to pop in a head for 3D printing, or one for PCB board drilling, or laser exposing.
I guess what I what is the capability to mill the item I just 3D printed.
I want start by specifying a target for precision that is 10x better than the best currently available, although I realize I will eventually have to compromise at some point, I want to know what would be needed to to surpass the current commercial machines. Equipped with this knowledge we may find that the precision ends up worse than existing machines at a given cost, at that point the cost would be one of the scaling factors.
Graham Stabler - fixed gantry router (sorry bridge portal mill) looks like what I had in mind. 2-zxis parallel kinematic concept looks insanely cool. What about 2 of these set perpendicular to each other? In post #48, the group of three pictures show examples with two separate rails. Have you tried moving one end of the rails? I guess that wouldn't be parallel anymore, but that's how I imagined it wanting to move when I first looked at it.
Chris_D working backwards from the tool - OK, since we're still at high level design, I would like to entertain a marginal alternative. Can we examine defining the "tool" as an interchangeable plastic extrusion head -AND- a Dremel type CNC cutting head? This is what I really (imagine I) want, and might be different from existing designs. Since CNC requirements are stricter than 3D printer requirements, I guess that If we design for CNC most of 3D printing might be at least partially addressed.
Also I agree that the existing software for PC should be able to do a large part of the job (at least at first), until we find that we find a need for something custom. I suggest that the prop only needs to provide a standard interface between the motors and geometries of the hardware, and the existing software that is already proven to work well on a PC. (at least for the time being.)
Price - we know the cost of the prop, what is the cost of the ball-screw and ballscrew-axles that stephaL38 has recommended?
$20K might be a little high for me. I was thinking about $200 (at least as the amount I tell the wife when I start). Can we fashion something out of a single common resource? Around here, kids use ABEC 7 bearings in their skateboard wheels. I could easily scare up a pile of disused skateboards, is there a possibility of harvesting parts?
FDM head- I did not catch what FDM stands for, can you expand this to words? Can we get these separately? .254mm filament sounds fine, especially if design the head to be interchanged with a CNC-dremel for further finishing.
Heated environment - This might not be feasible, but could one use heat-gun, thermostat, and fans to control the environment? The prop should be able to control any of these if direct control of the heatgun by the thermostat is insufficient?
"students created a 3D printer and used sugar as the printing medium" - SWEET! Just in time for Valentines Day too! I see a biz op right off the bat!
But I would like the option of changing between plastic and wax printing if the extrusion head can accommodate it.
Improving over the precision of commercial CNC machines means spending more money on the linear bearings, building stiffer machine frame and buying a spindle with exceptionally low run out. Just spend many more $
But for a printer to beat the current state of the art at least in FDM(see below) you will need to decrease the nozzle size, this may make it very difficult to make it work at all due to the viscosity of the plastic. You can beat the repraps and makerbots by just making a better machine mechanically speaking which is not hard, a typical CNC router is.
I think specifying that you want 10X greater resolution when you don't know what you want to make or what resolution you actually need is bad engineering. If you want to discuss how to improve things then fine, but perhaps start by understanding what defines the resolution. Do you really need better resolution than the commercial machines? Remember that if you specify a 10X thinner filament you can expect prints that take 10 times longer, couple that with the use of screws not belts and the machine may be rather a lot slower performing rapids (and there are lots of those). You could end up quite a sad bunny.
You talk about ballscrews and then about skateboard bearings. Both are perfectly good hardware at opposite ends of the spectrum, mixing them looses the benefits of each. Ballscrews are not cheap, you also need bearings for each end (that must restrict axial motion) and shaft coupling. All ballscrews are not equal either, some are ground and some are rolled and some have more backlash than others. Anyway you can't have significantly better and cheaper than the current hobby level machines, only one or the other.
FDM is the process that reprap and makerbot and Dimension's uprint use, fused deposition modelling (don't have google? :P ):
I'd recommend reading up on current 3D printing and rapid prototyping types, might come in handy. It might also be useful having a look around at other peoples CNC efforts, trying to learn it all here means we repeat a heck of a lot that is already out there.
Graham
Sorry I couldn't understand the comments on the parallel kinematic machine.
I want a design that is stiff enough for CNC on my 24 inch footprint motion bed.
I want to pop in a head for 3D printing, or one for PCB board drilling, or laser exposing.
I guess what I what is the capability to mill the item I just 3D printed.
I want start by specifying a target for precision that is 10x better than the best currently available, although I realize I will eventually have to compromise at some point, I want to know what would be needed to to surpass the current commercial machines. Equipped with this knowledge we may find that the precision ends up worse than existing machines at a given cost, at that point the cost would be one of the scaling factors.
Chris_D working backwards from the tool - OK, since we're still at high level design, I would like to entertain a marginal alternative. Can we examine defining the "tool" as an interchangeable plastic extrusion head -AND- a Dremel type CNC cutting head? This is what I really (imagine I) want, and might be different from existing designs. Since CNC requirements are stricter than 3D printer requirements, I guess that If we design for CNC most of 3D printing might be at least partially addressed.
Also I agree that the existing software for PC should be able to do a large part of the job (at least at first), until we find that we find a need for something custom. I suggest that the prop only needs to provide a standard interface between the motors and geometries of the hardware, and the existing software that is already proven to work well on a PC. (at least for the time being.)
Price - we know the cost of the prop, what is the cost of the ball-screw and ballscrew-axles that stephaL38 has recommended?
$20K might be a little high for me. I was thinking about $200 (at least as the amount I tell the wife when I start). Can we fashion something out of a single common resource? Around here, kids use ABEC 7 bearings in their skateboard wheels. I could easily scare up a pile of disused skateboards, is there a possibility of harvesting parts?
I think I have a can of fairy dust laying around somewhere, I can just sprinkle it on a workbench and we will have anything we want for free ;-)
You have got to do a lot more research on this before going any further. I spent a few minutes this morning to gather a few details about 3D printing with plastic, the extruder alone will cost you $150.00.
10X the precision? What are you kidding? You have to get realistic somewhere. A commercial CNC machine tyically has .0001" resolution, to drop to .00001" resolution would require glass scales, fully closed loop control, and a temperature controlled enviroment to work in.
$200.00, again, do some research - see what is out there. I would say that the typical bench top mill with steppers and made out of something more than sawdust glued together is about $1500.00 to $2000.00 - and that isn't with premium components. I have not priced linear bearings or ball screws in a while but don't expect to find these cheap!
Rethink everything, do a little research on prices, capacities, precsion required, etc.
There is a very good effort, that uses the FDM method, open air, very lean material feed system, lean print head, and incorporating a PC for the control work being produced out of China in fairly good numbers. It was sold originally at $1500, and now runs about $3K or so.
Resolution and such are comparable to the uPrint devices, which are reasonable resolution. Not great, but just reasonable.
Beating that $1500 number is going to force a trade off of speed, resolution, model size, etc... and probably all of the above. Rep Rap machines are coarse by comparison, but they do run a few hundred bucks!
If anyone is interested, I have a proposed contribution to offer.
I visit the patent office website on a fairly regular basis. Within their website, there is all kinds of technology available. However it is all patented.
If anyone would be interested in reading patent documents that pertain to 3D printing, please let me know, and I will provide PDF patent documents pertaining to this subject.
I am not saying that I will proof read all these documents, however they will apply to 3D printing. I have provided a sample just to give you an idea of whats available. And there are many more.
Comments
Thanks for the feedback. This is the purpose of review of the initial high level requirements, I'm glad the process is functioning correctly..
The overall, highest goal is to make a base design for a motion control machine.
The base design can be optional reinforced to perform as a CNC machine, or made less expensively as a 3D printer.
The base design can accommodate beefier motors for CNC, or cheaper motors for 3D print.
The design of the head is to include a standard mount so that an appropriate head may be attached at the users option. This is not limited to:
Small CNC head (made a Dremel type tool if the material is small and soft)
"Standard" CNC head (what ever that may be at this time)
Plastic extruder
Ceramic mix extruder
Hot Wire cutter
Laser light (for exposing photo sensitive etching material)
Pen or other marking utensil
[anything the user wishes to mount]
My thought is that the common denominator of CNC, 3D, plotter, etc is the bed, arms, motors, and motion control software. This is the base design.
Any particular machine can be built by extending the base design to that specific purpose.
Also the physical size is to be scalable, so (the first one I build) can fit on my desktop, and the one my friend builds can take up the whole garage. But the software interfaces, etc etc are recognizably similar.
The intent is not to re-make the maker-bot; unless that's what the maker;bot already does, in that case the work is done and I just go buy a maker bot and save space on the forums.
I don't think the makerbot is a universal base design intended to be expanded, but maybe I didn't read the article carefully (happens all the time to me).
Is this too broad scope?
It is only my opinion, but the first thing that should be established is a reliable printing solution. The success of this project is ultimately dependant upon the printhead, the media, and resolution. Perhaps several solutions would be best for different types of endeavors.
Secondly, I would suggest that you set a size parameter, that way everyone is on the same page. What is the largest item that everyone wants to be able to print? This should help establish guidelines for X, Y, and Z axises, software parameters, etc. Be realistic though, because the larger you go, the higher the cost. Of course everyone may have different views and/or play money to invest in the project. Once again it is only my opinion, but the height of the printing will ultimately determine the complexity and cost of this project, from a mechanical standpoint. Of course this pertains mostly to a CNC mill type of setup.
Bruce
For some other processes, overall manageability of the material determine model size. The raster based ones require a considerable amount of material be on hand for the process to work. A cubic foot is about where those play well.
Some use layers of material, spools, and they are limited by spool size, and the practicality of cleanup and handling waste. Current commercial solutions have a 6-8" spool width, and a similar model size.
The older "dot matrix" wax type printers offered half a cubic foot or so build space.
A base machine that can address roughly a cubic foot would be sufficient for moving forward. The commercial solutions and processes involve printing partial models for these reasons.
Although I am not that familiar with existing commercial processes, I would have to agree that a machine that addresses a cubic foot would provide a good modeling space.
Bruce
I'm really not sure what the point of making your claim was. I honestly and truly could not believe what I was reading. I could not imagine under any circumstances making a claim such as that whether true or not, especially when I would not be able to back it up.
Please stay and contribute, or if you don't want to then don't, but I think you have to stop throwing your toys out of the pram every time you are in some way challenged. I've read many of your posts and the only thing telling me we would be losing a great deal is you. Let your advice and your wisdom and your published progress speak for you, saying you have a patent does not prove a thing and it is strange to think it does. You know who you are, you know you are great, let us find this out for ourselves from measurable output not here-say.
This is why I respect PhiPi, Mike Green, Heater, Jazzed etc etc etc. I have absolutely know idea about their credentials (I think Mike is a doctor!) but my god these guys know their stuff and I know that simply because they provide good information which is well backed up, they make reasonable arguments and try to be helpful.
Graham
Absolutely right I think, as in CNC machining you should work back from the "cutter". Actually back from the part. Decide what parts you REALLY want/need to make, what material they might be made out of and then to what resolution. That will determine the material and the process you might use and might even inspire new process creation. Then you are also in a position to work on the machine as you know what it needs to move and how fast and with how much force.
I also think the prof is right, motion control is the first propeller task, it is needed for all machines.
Graham
Hey! Excellent! Looks like the big part of the work is already in place. Has anybody played with this yet?
My only reqeuirement for size (so far) is that it fit on my desk, which is 24" deep (and 48 inches wide, but I would like to leave some space for crayons, etc).
Lets say the requirement is a 609mm (square?) maximum foot print for the "small" option.
Further, lets say that a scaled up or large version is out of scope of the current effort. My friend won't commit his garage to a giant DYI CNC until I at least show the little 3D printer works.
So let me ask about the differnece between a 3D printer and a CNC at this "desktop" size. Would there be a substantial cost to construct a 24" (609mm) square foot print device that is capable of 3D printing and CNC? For example, could we limit it to software materials and use something along the lines of a Dremel tool as a cutting head? Or is "changeable heads" a poor idea?
Perhaps the easiest way to do it is to buy the print head and basic hardware from the Thing-O-Matic people and then copy their design for the rest of it.
I'd almost buy the whole thing, but I have a problem with paying up front and then having to wait 7 weeks. Also, I'd rather make it Propeller powered...
They sell most of the parts to make it on their website...
I'm thinking about a 24" cube for the box...
A machine that could do both could be very worth building. Lots can be done with even just a dremel...
You must decide what it is you actually want to be able to make.
Let me assume you want to print plastic with FDM and machine plastic, wood, G-10 and other soft materials. You could use a dremel but they are not great, there are some small routers that are reasonable (Bosch, Kress etc) and you can get a better version of the Dremel from Proxxon. To make the machine I'd recommend a fixed gantry router with moving platen. These are simple to construct, stiff and given the small size of the required work envelop the loss of working area compared to the moving gantry type is fine. Drive with belt drives using stepper motors, enough torque for the light cutting and plenty of speed.
Or you could use my 2-axis parallel kinematic concept and add a platen beneath it with an adjustable height. Fast (very low inertia and axes do not carry each other), stiff, novel but a little more tricky when it comes to motion. Again use belts and perhaps 4 leadscrews for the z-axis platen but with some decent bearings. This design is also good for raster work as both motors work together in one axis giving extremely high speed.
But I say again, decide what you want to be able to make. For me I actually want to make precise patterns for casting as I already have access to FDM and I have a CNC, for you I'm not sure, your call!
Graham
I like the idea of the basic concept to have a basic design that can be varyied.
CNC-milling of aluminium or even harder (in a double sense) CNC-milling of steel combined with productivity and a precision of 0,05 mm requires different technologies than a 3D-printer.
So if the basic design should be as broad as from a few 100 dollars 3D-printer up to a steel milling CNC with a resolution of 0,05 mm or even 0,02 mm the hardware-design has to to be developed from the upper end
which means CNC-milling. To achieve this resolution a lot of parts have to be very precise (ball-screw and ballscrew-axles that are polished and not rolled) and a very VERY STIFF construction.
Motion-control should include PID-control of position, speed and acceleration/deceleration. Which would be best achieved by servo-motors instead of stepper-motors.
So I guess such a broad scope would be hard to achieve.
If you make compromises about
- precision only 0,1 mm,
- productivity (milling only 0,1 to 1 mm thin) using just a dremel (instead of a 1000 W spindle) press start button at 7:00 AM a single steel-part is finished at 7 PM
- start-stop driving of stepper-motors
the options on how the design could be get much broader.
Another aproach could be to design central functions modular:
developing standards for interfaces like
position-feed-back (endswitches, reference switches, encoders)
GUI or GUI-elements
command-set for communicating with a motion-controller
To keep it standardised all the special things of the included machine-types 3D-printer, laser-enlighter, laser-cutter, CNC-mill etc. must be considered.
I haven't thought very much about all these machine-types. But to satisfy all users it has to be very universal.And maybe with this broadness the universality
changes into a brake because there are too many parameters or too many commands. I want to emphasise this MAYBE could happen. If somebody has good ideas
how to achive a "compact" universality please post your ideas as this will forward the project a lot.
best regards
Stefan
I don't see why the motion control should include PID, we can leave that to the driver. I'd say it is more sensible to have step and direction output which can then go to a standard stepper motor drive or a servo drive such as those from Gecko. Actually even high end AC servo drives will often accept step and direction, my Yaskawa drives do. You can certainly machine steel without that level of complexity.
The whole encoder feedback thing, well that can be an object, use it if you want it. I've never bothered and never missed it. If I had it I am sure I would love it, if in production I might well need it.
Do you suppose that the object based nature of the propeller might take care of the modules naturally?
I imagine a decent motion engine, a g-code parser, encoder reader, limit monitor as objects.
Graham
Same here. I hope to see more about idbruce's 3-axis CNC assembly.
I'm liking the changeable head idea, and allow printing and CNC (maybe even CNC detailing on the object that was printed?)
I need to find out more about "fixed gantry router with moving platen" and "better version of the Dremel from Proxxon".
Drive with belt drives using stepper motors also sounds right. (so precion of .01mm is too strict?)
Is there a picture of "2-axis parallel kinematic concept and add a platen beneath it with an adjustable height"?
Since this is a first attempt, I would propbably settle for a printer that makes a candle out of wax, or a CNC that makes a small square candle out of a big round candle.
A machine that can function as a 3D printer that can make a fork and spoon, and as a CNC that can work jewelers wax and drill a PCB would be sufficient. Anything more is gravy, but gravy will be accepted.
So far, we have narrowed the size, now we are looking at functions. Is this getting more reasonable or further away from "possible"?
A fixed gantry router:
And my 2-zxis parallel kinematic concept
http://www.youtube.com/watch?v=QMJE7o_3aFo
It does not have an adjustable height platen but it could be added.
0.01mm is far more than required given the likely run out of the cutter. And our commercial 3D printer uses steppers and belts.
Graham
Cool ! Complete different design to what I have seen so far. So thank you for sharing this idea.
I know: asking questions is so much easier than answering them or compete by showing another concept developed on my own.
So if I ask it is not to be skeptic it is just out of curiosity. Can you already say something about the precision that is achievable with this concept?
best regards
Stefan
Of course. Well in terms of resolution you can imagine that if both actuators move together you get one axis of motion (the long axis). In this direction the resolution is as per your actuators, so if you use belts which is most likely that comes down to your pulley size. On this test machine which is made from epson printer parts with arms machined on my CNC the pulleys are small, about 6mm diameter I guess, so you can say for half stepping a 200 step motor about 6*pi/400 = 0.047 mm resolution. The other axis can be considered as being related to the differential position of the two actuators. Here is an image that shows the variation in resolution across the working area:
You can see that the grid you get is not square but that should not worry you as this effect would be occurring with 0.04mm between the lines. Resolution is about the same as the actuator's in the middle but decreases towards the edge. You can also see the working area is not square, I imagine it best to pick a square or rectangle in that area and consider it the working area. It is one disadvantage of parallel machines. By changing the spacing between the rails you can also change the working area, you can even toggle the mechanism in some cases on purpose.
Here are some sketches I made. I was considering a laser cutter:
And here is why it is a reinvention only
Notice that you can use a parallelogram to create a plate for the attachment of the tool that does not rotate.
For more info including the fairly simple mathematics see my page:
http://www.indoor.flyer.co.uk/kinematics.htm
Graham
The mechanical framework and kinematics to move the print head is the least of the problems at this point. There are dozens (hundreds, thousands?) of designs for the mechanicals. Be it a: bridge mill (often incorrectly called a fixed gantry), bed mill, gantry mill, knee mill, articulated arm, or whatever, there are many choices to work with. If you can't achieve a decent print method, the framework and kinematics means nothing.
The drive system for the axes is another, "worry about it later". You can use: belt drive, rack and pinion, lead screw, or ball screw - many proven designs to work from. The resolution you want to achieve here is dictated by the resolution of the print head. Having greater resolution on the axes than on the print head serves no additional function. Using: PID, open loop, closed loop, servos, or steppers - who cares at this point! An open loop stepper drive system would be very reliable for the 3D printer and even for many light duty machining applications. Stick with something easy and economical - open loop steppers fits the requirements nicely.
Motion control software - why spend time worrying about this at this point in time. I mentioned that the MACH software and a PC can handle what you need for 3D printing (and much more). The demo version is free and it would be very suitable for use during testing. If you want to incorporate a PROP into the system, worry about that after you have solved the tough engineering aspect of the project which is the print head.
My whole point is this, you have to solve the problem of the printing method first, that is the unknown in the system. All the other required elements of the system have already been done and there are multiple, pre-existing solutions available.
If this is going to be some sort of kit or something to be sold, an even better starting point is to determine what the build budget should be - what is it worth to the end user. If this thing ends up costing $20,000.00, who wants it at that price?
Chris
Assuming the printing method is just FDM then Mach or similar could be used, for other printing methods this might be less practical (syncing a jetting head with motion).
So if the real issue is the head and the mechanics and the prop might be another way to implement the control at a later date it does rather sound like we are on the wrong forum at this stage???
I can tell you exactly how the FDM head in our uprint works, very simple. It pushes the plastic rod using pinch rollers through a heated tube which has a small nozzle. Overall dimensional accuracy of the part in x,y comes down to the machine pretty much, you can get decent tolerances. Minimum feature comes down to the width of the extruded filament, this also affects print speed. Resolution in z also comes down to the filament thickness, the uprint does 0.254mm thick layers. The print heads are carefully temp controlled, the nozzle is purged and wiped (with brush and squeegy) on a regular basis and the printing environment is also heated. On the Uprint which has two nozzles the same drive motor is used for both nozzles, it is changed over by driving the head against stops on the machine (a clever reuse of the axis motors).
Graham
p.s. Sorry for using the wrong term for a bridge mill (or portal mill as they seem to be sometimes called.) I'd be interested to find a glossary of the accepted terms so heinous crimes like this can be avoided in the future.
Looking at the FDM head you have at your disposal, does it look like something that can be made by the average person here on the forum? If not, do you know of a source for them? The .254mm (.010" about 3 thicknesses of paper) thickness is okay but like yourself and others have stated, that resolution depends on what people want to make with these things. Do you also happen to know what the diameter of the nozzle is? Do you know why the enviroment is heated? Possibly to help the plastics adhere better?
I would think that most people would want this 3D printer to print with plastic as the resulting product would be useful. I have seen some other 3 printers that use some sort of white powder that is sprayed (ink jet printed) with a color binder. This created a 3D color model of the object but I don't recall the durability of the object. Perhaps the others in the thread might want to voice their ideas of what they plan to make with this thing.
As for the wrong forum, I don't know. Eventually it could lead to a propeller related project. Heck, even the controller for the extruder and other heating elements could be propeller controlled - we would need some electronic gurus to help with that stuff.
No worries on the bridge mill miss-use, the hobbiest groups everywhere use that term, drives me nuts, why not call it what it is. As for a glossary, I can only reflect back to the engineering text books from years gone by. Perhaps there is something on-line somewhere. Another source would be machine tool company websites, one would hope that a company selling a machine tool for a quarter million dollars or more would know what to call it.
Chris
*If you want to create precision parts from metal, build a milling machine.
*If you want to create prototypes or casting patterns, build a 3D printer.
Of course, there are many grey areas in between.
Graham was talking about printhead technology. I once read an article where these students created a 3D printer and used sugar as the printing medium. The article said the results were very satisfactory, but then again, I did not hold the finished pattern in my hand. There are options, and I think we should research to find economical, but reliable print technology. I am sure plastic printing would be the best solution for sand casting, however I believe it would be costly and time consuming to implement. I think I would lean more toward wax printing and lost wax casting.
EDITED: But there is a downside to lost wax casting, you always lose your pattern, whereas in sand casting, you can reuse the patterns.
Bruce
I just lost my reply, I typed in the quick reply but then pressed the reply to thread button!
It was long so I will regain my strength before retyping, in fact I might just takes some pics or a video.
Graham
Chris
http://forums.reprap.org/list.php?35
babbelfish or google translate
EDIT: I just notice this is the German sub-forum of the English language site
I want a design that is stiff enough for CNC on my 24 inch footprint motion bed.
I want to pop in a head for 3D printing, or one for PCB board drilling, or laser exposing.
I guess what I what is the capability to mill the item I just 3D printed.
I want start by specifying a target for precision that is 10x better than the best currently available, although I realize I will eventually have to compromise at some point, I want to know what would be needed to to surpass the current commercial machines. Equipped with this knowledge we may find that the precision ends up worse than existing machines at a given cost, at that point the cost would be one of the scaling factors.
Graham Stabler - fixed gantry router (sorry bridge portal mill) looks like what I had in mind. 2-zxis parallel kinematic concept looks insanely cool. What about 2 of these set perpendicular to each other? In post #48, the group of three pictures show examples with two separate rails. Have you tried moving one end of the rails? I guess that wouldn't be parallel anymore, but that's how I imagined it wanting to move when I first looked at it.
Chris_D working backwards from the tool - OK, since we're still at high level design, I would like to entertain a marginal alternative. Can we examine defining the "tool" as an interchangeable plastic extrusion head -AND- a Dremel type CNC cutting head? This is what I really (imagine I) want, and might be different from existing designs. Since CNC requirements are stricter than 3D printer requirements, I guess that If we design for CNC most of 3D printing might be at least partially addressed.
Also I agree that the existing software for PC should be able to do a large part of the job (at least at first), until we find that we find a need for something custom. I suggest that the prop only needs to provide a standard interface between the motors and geometries of the hardware, and the existing software that is already proven to work well on a PC. (at least for the time being.)
Price - we know the cost of the prop, what is the cost of the ball-screw and ballscrew-axles that stephaL38 has recommended?
$20K might be a little high for me. I was thinking about $200 (at least as the amount I tell the wife when I start). Can we fashion something out of a single common resource? Around here, kids use ABEC 7 bearings in their skateboard wheels. I could easily scare up a pile of disused skateboards, is there a possibility of harvesting parts?
FDM head- I did not catch what FDM stands for, can you expand this to words? Can we get these separately? .254mm filament sounds fine, especially if design the head to be interchanged with a CNC-dremel for further finishing.
Heated environment - This might not be feasible, but could one use heat-gun, thermostat, and fans to control the environment? The prop should be able to control any of these if direct control of the heatgun by the thermostat is insufficient?
"students created a 3D printer and used sugar as the printing medium" - SWEET! Just in time for Valentines Day too! I see a biz op right off the bat!
But I would like the option of changing between plastic and wax printing if the extrusion head can accommodate it.
Improving over the precision of commercial CNC machines means spending more money on the linear bearings, building stiffer machine frame and buying a spindle with exceptionally low run out. Just spend many more $
But for a printer to beat the current state of the art at least in FDM(see below) you will need to decrease the nozzle size, this may make it very difficult to make it work at all due to the viscosity of the plastic. You can beat the repraps and makerbots by just making a better machine mechanically speaking which is not hard, a typical CNC router is.
I think specifying that you want 10X greater resolution when you don't know what you want to make or what resolution you actually need is bad engineering. If you want to discuss how to improve things then fine, but perhaps start by understanding what defines the resolution. Do you really need better resolution than the commercial machines? Remember that if you specify a 10X thinner filament you can expect prints that take 10 times longer, couple that with the use of screws not belts and the machine may be rather a lot slower performing rapids (and there are lots of those). You could end up quite a sad bunny.
You talk about ballscrews and then about skateboard bearings. Both are perfectly good hardware at opposite ends of the spectrum, mixing them looses the benefits of each. Ballscrews are not cheap, you also need bearings for each end (that must restrict axial motion) and shaft coupling. All ballscrews are not equal either, some are ground and some are rolled and some have more backlash than others. Anyway you can't have significantly better and cheaper than the current hobby level machines, only one or the other.
FDM is the process that reprap and makerbot and Dimension's uprint use, fused deposition modelling (don't have google? :P ):
http://en.wikipedia.org/wiki/Fused_deposition_modeling
I'd recommend reading up on current 3D printing and rapid prototyping types, might come in handy. It might also be useful having a look around at other peoples CNC efforts, trying to learn it all here means we repeat a heck of a lot that is already out there.
Graham
Sorry I couldn't understand the comments on the parallel kinematic machine.
I think I have a can of fairy dust laying around somewhere, I can just sprinkle it on a workbench and we will have anything we want for free ;-)
You have got to do a lot more research on this before going any further. I spent a few minutes this morning to gather a few details about 3D printing with plastic, the extruder alone will cost you $150.00.
10X the precision? What are you kidding? You have to get realistic somewhere. A commercial CNC machine tyically has .0001" resolution, to drop to .00001" resolution would require glass scales, fully closed loop control, and a temperature controlled enviroment to work in.
$200.00, again, do some research - see what is out there. I would say that the typical bench top mill with steppers and made out of something more than sawdust glued together is about $1500.00 to $2000.00 - and that isn't with premium components. I have not priced linear bearings or ball screws in a while but don't expect to find these cheap!
Rethink everything, do a little research on prices, capacities, precsion required, etc.
Chris
Resolution and such are comparable to the uPrint devices, which are reasonable resolution. Not great, but just reasonable.
Beating that $1500 number is going to force a trade off of speed, resolution, model size, etc... and probably all of the above. Rep Rap machines are coarse by comparison, but they do run a few hundred bucks!
If anyone is interested, I have a proposed contribution to offer.
I visit the patent office website on a fairly regular basis. Within their website, there is all kinds of technology available. However it is all patented.
If anyone would be interested in reading patent documents that pertain to 3D printing, please let me know, and I will provide PDF patent documents pertaining to this subject.
I am not saying that I will proof read all these documents, however they will apply to 3D printing. I have provided a sample just to give you an idea of whats available. And there are many more.
Bruce
http://www.google.co.uk/patents (I love the hilarious example patents on the front page)
http://ep.espacenet.com/ (searches all patents, free pdf downloads and no messing around)
The company related to the original FDM stuff is stratasys if anyone wants to have a look.
Graham