This is why I don't visit the patent office. Just saying
I don't really care to know how they did it, because it's infringing to use with out a license, and... we need serious patent reform anyway. Software and business process patents are killing innovation big. Not cool.
I think it is fair for a creator to own exclusive rights for 20 years. What's wrong with that? The only reform that I would like to see is in their processing time
Please, let's stop talking about patents unless we are talking about their contents or we are going to have several pages of Smile in the middle of the thread.
I was thinking a central location to post the technology that is available. I believe that most likely these will be homebrew machines, instead of a commercial venture, but you never know. I was just saying that a lot of good information can be retrieved from a patent. I will provide them if anyone thinks it is worthwhile, otherwise I won't. Simple as that.
I was thinking a central location to post the technology that is available.
I was thinking of way to collect and organize our discussions and any concensus or divergence.
I use google.code for the software project I am involved in (propforth) and I started one for a robotics club that did not pan out and I have since lost the url, but I believe this was on google groups.
Do you guys feel this thread is sufficient for capturing our plans as a "running discussion" or would something else be more suitable?
I tend to like using the thread as a running discussion, but have a separate repository for requirements, instructions, references, etc.
Perhaps the method of updating the first post in the thread with files to the "current state" would be sufficient? Your opinions please.
Thanks for the link on Fused Deposition, I think we should use the term Fuse Filament Fabrication (FFF) so we avoid using terms with legal problems.
>> ...bad engineering...
To me, it isn't yet the engineering phase, this is still the research phase. This is the decision process. "Just spend more money" is not clear, we need to establish what to buy and how much it would cost to determine what the trade-offs are. I have found that "bad engineering" is the result of skimping on preparatory research. Please bear with me, I'm slow. I knew nothing about CNC and 3D printing four days ago, and now I have a good start on requirements AND there's even a couple local folks that are interested in helping with the build. This phase is still examining the trade offs and narrowing the requirements.
I see we have some numbers: if current best is .0001mm, and improving it 10x to .00001mm would be too difficult, then fine. Somehow I thought .01mm was mentioned as the current precision for something in the discussion; I wanted to examine the degree of difficulty to improve this. Sorry that this was unclear. 0.01mm is fine, if that precission is the same as other units, we can figure out a way to make the cost lower.
I see an important factor in printing precision is the nozzle size. I didn't realize that, I thought the stepper motor would be the main factor in determining precision.
The printing mechanism need not be commited to any particular medium, the design should accomodate motion suitable for our choice of head. Similarly, the CNC option can be limited to material softer than steel.
There have been some interest in the following:
- Interchangeable heads, both printing and CNC.
- CNC capability, meaning very rigid for a 3D printer, but maybe not so good for something as hard as steel. Possibly slower milling time as a result; Proxon was mentioned.
- Initially a 24" x 24" footprint (desktop size)
If we further say that the current portion of the project is focused on the hardware for the bed and motion, and the software is limited to (using a prop for) controlling the motors, does that give us a narro enough focus? [Still assuming that the PC software can handle covering CAD into the list of motion instructions, and the prop runs the motors, raises/lowers the extrude/cutter/pen head, and activates the extruder/cutter/pen]
With this set of trade offs, does this sound like a fixed gantry/moving platen arrangement would be feasible?
What kind of cost might we be looking at? Any early guesses?
I don't see what legal issues we could face saying FDM, it is what is in the text books.
You stated that "I want start by specifying a target for precision that is 10x better than the best currently available". This sounded like you had already made decisions instead of a thought experiment.
0.01mm is probably better tolerance than most DIY efforts manage. Spend more money means buy better and more accurate components, you can't do better than your worst components. Positional resolution comes down to the lead of your screws, if we assume a stepper motor it is best to assume that the smallest step will be a half step (even if micro stepping is employed) because this is the smallest step size with high torque. But a fine pitch screw may still have backlash or lead error (variation in the lead along its length) so you spend more money to get a better one. Of course if the machine flexes during machining or there is run-out in the spindle then you lose precision again.
If we are talking about using a PC in the first instance for motion control then the propeller is not even needed to control the motors, just buy cheap off the shelf stepper drivers run from the parallel port like most home CNCs.
I don't see any trade offs listed but a bridge router can of course have a 24"X24" size and interchangeable heads. If you feel that is what you might like then you have two goals, build a small router and build a print head of some sort. The former has been done ad infinitum so there is a lot of info online, the latter has been done but could be improved or perhaps new methods developed.
The propeller could do it easily. A DLP projector shines UV onto the bottom of a glass dish of resin, this hardens a layer selectively, a build base is then lifted up and the process repeated, layer by layer in an inverted process.
Extremely fast, very easy (no real motion control) and high resolution. You could just output bmps to the projector from the propeller via an SD card.
But no milling functions!
I have some papers on this process deep in my vault, I always thought it looked like a cool process but could not afford a projector at the time, some years on they are fairly cheap. I'll try and dig the papers out.
Note that Zcorp now have a printer based on this technology which came from another company I forget the name of. However I have a feeling they don't do it inverted and that this inverted method is not patented.
I don't see what legal issues we could face saying FDM, it is what is in the text books.
LOL - I mis-read wiki, I thought it was a trademarked term. Shows what I know. I don't care what it is called as long as its consistent
0.01mm is probably better tolerance than most DIY efforts manage.
This is fine as a target then. Of course this can be changed as specific motors etc are spec'd
Spend more money means buy better and more accurate components, you can't do better than your worst components.
Understood. Do you have any specifics so we can look at trade-offs? IE What is you favorite stepper motor for this type of application?
propeller is not even needed to control the motors, just buy cheap off the shelf stepper drivers run from the parallel port like most home CNCs.
Understood, but the point of the project is to try controlling motion for CNC/3D printing with a propeller
bridge router can of course have a 24"X24" size and interchangeable heads.
Agreed.
small router and build a print head of some sort.
I was hoping for someone with experience with a particular small router and/or particular head(s) to tell us the trade-offs of each so we could evaluate if it is desirable to buy one off the shelf or build one with X-improvements.
I was also hoping to track the consensus of the experience contributors, or at least record where opinions differ. Being a software guy, I have no basis for opinion on mechanical or electronics, so I have to ask lots of n00b questions.
DLP projector shines UV onto the bottom of a glass dish of resin
Now THIS is totally awesome. In addition to CNC and 3D printing via extrusion etc, I would also like this option.
How about in addition to interchangeable heads, also design in an interchangeable bed so a glass sheet can be placed as the bed? Also, the "head" accepts an inverted base as in the post, the Z motion is provided by the head?
Even if the machine does not turn out to have the highest precision available, it certainly will be versatile, and definitely the coolest one on my block.
Oh, forgot to ask, anybody know how dangerous are the the chemicals in the UV exposure method? I don't want Minamata disease in my area.
I think the fact is it trademarked does not effect us using the term in a forum, if you make a product you might have to change the term.
Stepper motors are not very exciting beasts, they are all pretty much a muchness. But I would always go for 200 step, bipolar motors. If rated appropriately they step reliably (they don't loose steps randomly) and do their jobs well. You run them over the rated voltage with current limiting if you want performance, if you can stand low speed you can run them at the rated voltage with simple electronics. But given the cost of bipolar drives with current limiting these days I think it is best to buy the drives and get the propeller to produce step and direction signals.
If the "whole point" is to use the propeller then this is a complicated way to find a use for a propeller In the first instance use PC software, debug the process then introduce the propeller. A stand alone machine would be great and we can divide the coding effort into useful objects.
No specific experience with small routers but I can spot a good one. Chris' looks rather nice, I won't try and state what kind it is in case I get the name wrong but it does have a gantry and it does move. The pros of the bridge type I suggest are that it is easier to make stiff, it is easier to move a table back and forth with a single screw or belt than a moving gantry (which will often have two screws linked by belts or some other less nice arrangements.) But it will have a smaller working area than the type Chris has built. For a 24"X24" machine you can almost get that in travel, for the bridge machine you will get more like 24"X12" (or visa versa I forget). But that sounds OK to me anyway.
I think if you look around on the web you can get an idea of the pros and cons yourself assuming you have some grasp of the physical world. You might like to collect a few images as points of discussion.
Graham
p.s. The resin is nothing all that special, you can get resin kits for making rubber stamps that must be something similar. But not cheap. Don't drink it mind.
Yes, the one I showed in the video posts earlier would be a gantry machine - it's mechanics are like that of a gantry crane. I agree with your assesment that a bridge machine is naturally more rigid than a gantry machine. A bridge mill, has the upper portion fixed, just like a bridge that we drive our cars over. I have built 3 styles of machines at home, Bed mill, Bridge, and Grantry. The Bridge mill and bed mills are the most rigid and probably the easiest to make.
As for the propeller requirement, I would think that working with MACH first makes the most sense. Even the machines I built at home for the specific purpose of designing a CNC control started out being controlled by MACH. It serves as a good baseline and benchmark for the control design.
I still think the OPs and other's interested in doing this project need to start by defining their budgets. There is no point at all doing anything on this project if the budget is unrealistic - and at this point in time it is VERY unrealistic. Just guessing, I would say we are looking at between $2000.00 and $3000.00 to make something useful.
Quite true although I wonder how many of my projects I would have started if I had been as sensible as that
A few more pros and cons on the configurations. The bedmill has one axis stacked on the other as well as the work table. You can imagine therefore a price in terms of accelerations and speeds for a given motor or a price in motor size to get the same performance. The column that supports the spindle must overhand a large amount for significant travel, this may make it less stiff, hence it is common to see wide but shallow beds on bedmills.
If we look at reprap, makerbot and uprint then you have a gantry system that moves and a table that raises and lowers. The uprint uses a single ballscrew and decent bearings on a cantilevered bed, the others tend to use multiple lead screws. The advantage is that you can get large x-axis travel easily but it would be harder to make it stiff enough for machining I suspect.
I don't know about the costs but that does not seem unreasonable. If I had a DLP projector however I could build that system from parts I have lying around. I may look into making one for my work as it would be useful.
There is no point at all doing anything on this project if the budget is unrealistic - and at this point in time it is VERY unrealistic.
Good point. Like all great pioneers have discovered, Now Is Always the Best Time to Give Up.
However, I only have dozen or so projects going at this time, so I will continue my research.
Maybe its time to clarify. Please indulge me for another post.
You guys with all the experience appear to have very clear ideas of what one can and can't do. This is very good, and finding you and this knowldege is the point of me asking. However, you also appear to expect a non experience experimenter to have the same understanding without your basis of experience, which is the driver of me asking. I appreciate your frustration with a n00b, please stay with me we are making good progress.
So we are talking budget, great. How much do the options you describe cost? What are the trade offs for each? These are the things I am trying to determine.
There needs to be some basis for a budget, as in what will my $$$ get me; I don't think its sensible start with a dollar amount and work backward when a project is in the research phase. Although I have seen businesses do it this way, their consistent failure is the basis for my opinion that the research phase is critical and must be explored and understood before setting the project in stone. (This is the term failure prone individuals haved used when they wish to skip following through on the research phase).
Incidently, I have a source of discarded equipement, from industrial microscopes to engines, and my plan is to harvest parts whenever possible.
The things that cannot be easily harvested, like the extruder head, I plan to purchase or othersie obtain later, so this is not included in the budget.
The mounting for the head need only to be designed to accept a unit of similar size and mass.
SCOPE: I am trying to limit the scope of the project to table that holds the work, the arm or bridge or gantry that holds the tool, and the motors and associated mechanicals that make it all move. I don't know if this is the right term, but I think of this project as the "motion" part of a CNC/3D printer.
We've already established that a prop will be used to control the motor, I would like to further assume that the prop is already purchased and is not included in the budget.
Also the software for CAD and any translation or conversion needed will be some popular existing software, so this is not included in the budget.
USE-CASE: Perhaps a use-case is usefull here: With the design I have in mind for a machine, the requestor makes a drawing and converts it into a stand formatted list of motion instructions.
The Requestor give this CAD development output to this machine on perhaps an SD card as a preosrted list of motion instructions. The machine (prop and mechanicals) only reads the instructions and performs them in the order given. It is the responsibility of the requestor to make the correct motions in the correct order.
Summary of SCOPE of research phase: Table and mounting for the tools. What is needed to make is move, and how strong does it have to be? What are the costs of the vaious options, or at minimum which are more expensive and which are less expensive and what are the (quantifiable) trade offs for each?
The anwers these questions will give an indication of what parts are needed, and the parts list will give us an indication of necessary cost for benefit achived, and THEN the trade offs can be determined. At that poit is where I would start scaling back requirements from millionths to thousandths to tenths and maybe back up to hundredths.
This is what I have in mind, however, I plan to accomplish the project for substantially less than $3000 (at least as far as my wife knows) since I don't plan to purchase all parts new retail.
There are numerous factors that ultimately affect the cost of a machine. Take a screw for example, slot head zinc coated screws are reasonably cheap, but now compare that to a socket head black oxide cap screw and there is a huge difference. You can wrap up hundreds of dollars just in screws.
As I told someone once before, you are the painter (the creator), it is your canvas, now paint the picture. Your needs and desires for this machine will affect any price that we might suggest. And there are any number of other factors that will also affect the price, such as:
How many parts can you make yourself?
How many parts can you salvage?
How many parts must be custom made?
Can you use stock components for many of your parts?
How accurate does it have to be?
ETC...
To accurately answer these questions, you should first create drawings of what you want to build, and then create a parts list.
However, I would guess that a decent "homebrew" 3D printing machine could be made for as little as $1500, but could easily reach $3000 - $5000 depending upon your needs and desires.
EDITED: Just to establish this point a little further, everyone has their own view. Take for example drive setup and the platform bed. While belt driven axises are much faster and cheaper to implement then screw driven axises, in my opinion they are much more difficult to design, but thats just me, I like belts, but I use screws. I also prefer a stationary platform bed. So you can see, we all have different opinions, and I am certain we all would attempt it in a different fashion. Every decision you make about the construction of the machine will ultimately affect the overall cost of the machine.
I have no problems with "noobs" but as far as I can tell you have not as yet done much research on this subject and that is not because of being new, there is a wealth of info out there, when I'm interested in a subject I hit that info hard and try and get a feel for the subject. There must be thousands of CNC build threads, dozens of 3D printing blogs, patents, manufacturing websites, youtube videos, personal websites, company websites, brochures.
If research means reading this thread and asking vague questions then something has gone very wrong. Heck what do we know, there might be an idea out there with your name on we would never have thought of. I'd like so see ideas being brought to the table even if they are bad ones. If I was in your position I would have been posting all the cool stuff I found daily.
With parts from Ebay I can pretty much build a fantastic machine for very little money (I got two ACservoed, ballscrewed linear actuators by Parker for $500, and one is huge), the issue is that as a group project it is useless to go that way. If a constraint is that the output of the project is something all can build then stock parts are a must. But you could say use standardised electronics and custom mechanicals with some stock elements (like the print head).
But really a bit early to be talking like this as so far absolutely nothing has happened except for a rise in air temperature
Graham
p.s. I am sorry for my ranty style in this thread, I'm just a bit frustrated.
Like I told you yesterday, once you get your feet wet, I think you will enjoy CNC.
In a previous post, I agreed with Graham on making a 3D printing machine for making casting patterns. I assume Graham wants this for the same reason as me, which is to make parts for CNC machines.
If you become actively involved in CNC, you will undoubtedly want the same abilities, because it will save you a lot of money when you need custom made parts and you can cast them yourself. Casting is also very cool stuff.
I would suggest that you focus on a 3D printer just to get your feet wet, and then all of us, as a group, can focus on creating one. Just my point of view.
Bruce
P.S. We can figure out the cost after we have designed
If the ultimate goal is to have metal parts, I wonder if it might be possible to skip some of the messy steps like sand-molds, and pouring hot metal?
Especially if you are looking for reasonably large parts, a bit of surface roughness should be acceptable. Might it be possible to automate a mig wire-welder and just build steel/aluminum parts directly?
I know that when I need to build up an area while welding manually, a series of "Dots" will often let me build up an area without melting away the substrate material, it should work for an automatic system as well.
A quick search on the internet shows that it has been "looked into", but I didn't find anything in production, The quick test I did with one of my 3 axis machines and an automatic feed hot-glue gun seems to indicate it might be a possibility for usable prototype parts.
I assume Graham wants this for the same reason as me, which is to make parts for CNC machines.
Not really, I want to be able to make parts that cannot be machined. The casting part is because it allows full strength materials, not just metal but resins and all kinds of things. For making CNC machines I just use more CNC machines.
Ken,
I have seen that done using a 6-axis robot arm and a mig welder, not really a cheap business. Doesn't need 6-axis but it will cost a fair bit in gas and welding wire!
I just tried to buy the Thing-O-Matic without the circuit boards with the hope that it would cost less and ship faster. Plus, I'd like to replace them with Prop powered boards.
But, they wouldn't go for it... They did say that if I bought it, I could sell the circuit boards easily. So, I think that's what I'll do...
Hi Ray, I decided to do the similar. I will order a MENDEL from the german reprep foundation and replace the electronics step by step. This will allow others to follow, if there is an advantage. We just have to proof, that the propeller does the job better. This is always the best argument.
ErNA
Looks like I can get the basic hardware for CNC/3D printer there and then get the extruder head from the Makerbot guys.
What do you think about the Zen setup?
If I understand right the $79 is just replacing some parts of the 7x7 CNC kit for a bigger axle distances.
A mechanical hardware kit for three axis including three stepper-motors for $330 is still a low price.
Now the most interesting question are:
- What is the maximum resolution?
- How big is the play in the axles if the direction is changed? (I assume that the axles are NOT build from play-FREE ballscrews at this price)
- What is the maximum feedrate without stalling the steppermotors?
- How does this feedrate go down if the infeed-motion goes up?
(stepper-motors with size NEMA 17 are quite small and the torque is limitied. Even if it is a 1.3A motor which is quite big for the size NEMA 17
- How does the precision of the mechanic go down (caused through bending of the frame )if bigger forces are applied when using bigger steppermotors?
I would estimate if I would like to mill aluminium with teh following parameters:
- milling-tool diameter 10mm
- a infeed-motion of 2mm
- feed-speed of 100 mm per minute the whole thing will bend that the
precision will go down to 0,5 mm. Which will be useless in most cases.
So what you have to to is using parameters like this:
- milling-tool diameter 3mm
- a infeed-motion of 0,5mm
- feed-speed of 20 mm per minute
to keep the forces low that the frame will not bend to much to keep a precision of 0,08 mm to 0,1 mm.
If you compare the two parametersets it is easy to see that the productivity goes down a lot with this machine.
I want to emphasise that this a RAW estimations from me as a guy who has only few experience with CNC-machines
Maybe the things are much better than I estimate. So this is why I'm really highly interested to get hard information.
The website of ZEN Toolworks does not mention anything about precision. I INTERPRET this that the precision is
quite low compared to CNC-machines at a pricelevel of $1500 to $2000.
So an old saying is still true here: Quality has its price
best regards
Stefan
Maybe Chris would like to tell his opinion about that.
It is always hard to give an appraisal on machines like this. It is of very basic construction so that it is easy to manufacture the parts (on another router) and cheap to sell so it probably not the stiffest of machines (it's plastic too). It seems to have standard threaded rod for the lead-screws with anti-backlash nuts. They are only anti-backlash up to the stiffness of the spring however and add friction to the screws decreasing available torque. The motors do seem rather small, Nema-23 might be better.
It really depends on what you expect from the machine, I'd not expect to do a great deal of precision machining but with a dremel or proxxon rotary tool you could make a lot of parts better than you ever managed with hand tools. If you tool VERY light passes with small carbide milling cutters you could also do some cutting of aluminium (panel fronts) but expect to die of boredom
As a 3D printer, once it get's all that height is gets even more floppy. The design of the long travel z-axis means that in order to make use of the travel the extruder will extend a long way down from the moving carriage. I'm not sure of the weight of a typical extruder but you could end up with a fair bit of flex. Some sensible use of reinforcement could probably fix this issue.
At the price as long as your expectations are not too high I think it is reasonable. That said you might like to check out the MDF made machines on the CNC zone for inspiration, you might be able to make something better and for less.
I was looking to learn as I went and so, other than getting the Nema17 with 62oz torque and of course the extruder, I was making my own. There are a few guys who will make you parts at a cost. So I wanted to start with a micro-mendel variant of my own. To that end I have done some hardware design and got a few parts from old roller blades and printers. SO far, I made my brackets to hold the rods from timber - lousy job but it will do until I can make some of the parts myself and replace the crappy timber joints. I posted the link to my thread. Also look in the same forum section for a writeup by a guy (cannot remember name, but his last post was just before Christmas. It is a great writeup, with all the problems he had.
I am sure the prop will do a better job. Just finalising some pcbs that will do the job for me. Not sure if I will get the time to do the Stepper Driver pcb this time around though. I am going to something similar to the Pulo? pcb, but a better IC variant (more robust). Too many other interesting projects to do
I emailed Zen and asked about the resolution. They responded that they claim 0.1 mm accuracy.
I think that's good enough for 3D printing since the Makerbot extruder tip bore hole is 0.5 mm.
Comments
I don't really care to know how they did it, because it's infringing to use with out a license, and... we need serious patent reform anyway. Software and business process patents are killing innovation big. Not cool.
Sorry, carry on.
That Google one is pretty cool. I did not even know Google had that up. Here is another one for you.
http://free.patentfetcher.com/
Bruce
Hmmmm.....
I think it is fair for a creator to own exclusive rights for 20 years. What's wrong with that? The only reform that I would like to see is in their processing time
Bruce
Regardless of the rights and wrongs they can provide inspiration.
Graham
Graham
I was thinking a central location to post the technology that is available. I believe that most likely these will be homebrew machines, instead of a commercial venture, but you never know. I was just saying that a lot of good information can be retrieved from a patent. I will provide them if anyone thinks it is worthwhile, otherwise I won't. Simple as that.
Bruce
I have no problems with patents, very helpful in understanding technology but we were drifting off.
Do whatever you want to do, I don't mind but I'm happy to search for myself personally.
Graham
I was thinking of way to collect and organize our discussions and any concensus or divergence.
I use google.code for the software project I am involved in (propforth) and I started one for a robotics club that did not pan out and I have since lost the url, but I believe this was on google groups.
Do you guys feel this thread is sufficient for capturing our plans as a "running discussion" or would something else be more suitable?
I tend to like using the thread as a running discussion, but have a separate repository for requirements, instructions, references, etc.
Perhaps the method of updating the first post in the thread with files to the "current state" would be sufficient? Your opinions please.
>> ...bad engineering...
To me, it isn't yet the engineering phase, this is still the research phase. This is the decision process. "Just spend more money" is not clear, we need to establish what to buy and how much it would cost to determine what the trade-offs are. I have found that "bad engineering" is the result of skimping on preparatory research. Please bear with me, I'm slow. I knew nothing about CNC and 3D printing four days ago, and now I have a good start on requirements AND there's even a couple local folks that are interested in helping with the build. This phase is still examining the trade offs and narrowing the requirements.
I see we have some numbers: if current best is .0001mm, and improving it 10x to .00001mm would be too difficult, then fine. Somehow I thought .01mm was mentioned as the current precision for something in the discussion; I wanted to examine the degree of difficulty to improve this. Sorry that this was unclear. 0.01mm is fine, if that precission is the same as other units, we can figure out a way to make the cost lower.
I see an important factor in printing precision is the nozzle size. I didn't realize that, I thought the stepper motor would be the main factor in determining precision.
The printing mechanism need not be commited to any particular medium, the design should accomodate motion suitable for our choice of head. Similarly, the CNC option can be limited to material softer than steel.
There have been some interest in the following:
- Interchangeable heads, both printing and CNC.
- CNC capability, meaning very rigid for a 3D printer, but maybe not so good for something as hard as steel. Possibly slower milling time as a result; Proxon was mentioned.
- Initially a 24" x 24" footprint (desktop size)
If we further say that the current portion of the project is focused on the hardware for the bed and motion, and the software is limited to (using a prop for) controlling the motors, does that give us a narro enough focus? [Still assuming that the PC software can handle covering CAD into the list of motion instructions, and the prop runs the motors, raises/lowers the extrude/cutter/pen head, and activates the extruder/cutter/pen]
With this set of trade offs, does this sound like a fixed gantry/moving platen arrangement would be feasible?
What kind of cost might we be looking at? Any early guesses?
You stated that "I want start by specifying a target for precision that is 10x better than the best currently available". This sounded like you had already made decisions instead of a thought experiment.
0.01mm is probably better tolerance than most DIY efforts manage. Spend more money means buy better and more accurate components, you can't do better than your worst components. Positional resolution comes down to the lead of your screws, if we assume a stepper motor it is best to assume that the smallest step will be a half step (even if micro stepping is employed) because this is the smallest step size with high torque. But a fine pitch screw may still have backlash or lead error (variation in the lead along its length) so you spend more money to get a better one. Of course if the machine flexes during machining or there is run-out in the spindle then you lose precision again.
If we are talking about using a PC in the first instance for motion control then the propeller is not even needed to control the motors, just buy cheap off the shelf stepper drivers run from the parallel port like most home CNCs.
I don't see any trade offs listed but a bridge router can of course have a 24"X24" size and interchangeable heads. If you feel that is what you might like then you have two goals, build a small router and build a print head of some sort. The former has been done ad infinitum so there is a lot of info online, the latter has been done but could be improved or perhaps new methods developed.
Graham
http://blog.makezine.com/archive/2010/10/homemade-high-resolution-dlp-3d-pri.html
The propeller could do it easily. A DLP projector shines UV onto the bottom of a glass dish of resin, this hardens a layer selectively, a build base is then lifted up and the process repeated, layer by layer in an inverted process.
Extremely fast, very easy (no real motion control) and high resolution. You could just output bmps to the projector from the propeller via an SD card.
But no milling functions!
I have some papers on this process deep in my vault, I always thought it looked like a cool process but could not afford a projector at the time, some years on they are fairly cheap. I'll try and dig the papers out.
Note that Zcorp now have a printer based on this technology which came from another company I forget the name of. However I have a feeling they don't do it inverted and that this inverted method is not patented.
Graham
LOL - I mis-read wiki, I thought it was a trademarked term. Shows what I know. I don't care what it is called as long as its consistent
This is fine as a target then. Of course this can be changed as specific motors etc are spec'd
Understood. Do you have any specifics so we can look at trade-offs? IE What is you favorite stepper motor for this type of application?
Understood, but the point of the project is to try controlling motion for CNC/3D printing with a propeller
Agreed.
I was hoping for someone with experience with a particular small router and/or particular head(s) to tell us the trade-offs of each so we could evaluate if it is desirable to buy one off the shelf or build one with X-improvements.
I was also hoping to track the consensus of the experience contributors, or at least record where opinions differ. Being a software guy, I have no basis for opinion on mechanical or electronics, so I have to ask lots of n00b questions.
Now THIS is totally awesome. In addition to CNC and 3D printing via extrusion etc, I would also like this option.
How about in addition to interchangeable heads, also design in an interchangeable bed so a glass sheet can be placed as the bed? Also, the "head" accepts an inverted base as in the post, the Z motion is provided by the head?
Even if the machine does not turn out to have the highest precision available, it certainly will be versatile, and definitely the coolest one on my block.
Oh, forgot to ask, anybody know how dangerous are the the chemicals in the UV exposure method? I don't want Minamata disease in my area.
Stepper motors are not very exciting beasts, they are all pretty much a muchness. But I would always go for 200 step, bipolar motors. If rated appropriately they step reliably (they don't loose steps randomly) and do their jobs well. You run them over the rated voltage with current limiting if you want performance, if you can stand low speed you can run them at the rated voltage with simple electronics. But given the cost of bipolar drives with current limiting these days I think it is best to buy the drives and get the propeller to produce step and direction signals.
If the "whole point" is to use the propeller then this is a complicated way to find a use for a propeller
No specific experience with small routers but I can spot a good one. Chris' looks rather nice, I won't try and state what kind it is in case I get the name wrong but it does have a gantry and it does move. The pros of the bridge type I suggest are that it is easier to make stiff, it is easier to move a table back and forth with a single screw or belt than a moving gantry (which will often have two screws linked by belts or some other less nice arrangements.) But it will have a smaller working area than the type Chris has built. For a 24"X24" machine you can almost get that in travel, for the bridge machine you will get more like 24"X12" (or visa versa I forget). But that sounds OK to me anyway.
I think if you look around on the web you can get an idea of the pros and cons yourself assuming you have some grasp of the physical world. You might like to collect a few images as points of discussion.
Graham
p.s. The resin is nothing all that special, you can get resin kits for making rubber stamps that must be something similar. But not cheap. Don't drink it mind.
Yes, the one I showed in the video posts earlier would be a gantry machine - it's mechanics are like that of a gantry crane. I agree with your assesment that a bridge machine is naturally more rigid than a gantry machine. A bridge mill, has the upper portion fixed, just like a bridge that we drive our cars over. I have built 3 styles of machines at home, Bed mill, Bridge, and Grantry. The Bridge mill and bed mills are the most rigid and probably the easiest to make.
As for the propeller requirement, I would think that working with MACH first makes the most sense. Even the machines I built at home for the specific purpose of designing a CNC control started out being controlled by MACH. It serves as a good baseline and benchmark for the control design.
I still think the OPs and other's interested in doing this project need to start by defining their budgets. There is no point at all doing anything on this project if the budget is unrealistic - and at this point in time it is VERY unrealistic. Just guessing, I would say we are looking at between $2000.00 and $3000.00 to make something useful.
Chris
A few more pros and cons on the configurations. The bedmill has one axis stacked on the other as well as the work table. You can imagine therefore a price in terms of accelerations and speeds for a given motor or a price in motor size to get the same performance. The column that supports the spindle must overhand a large amount for significant travel, this may make it less stiff, hence it is common to see wide but shallow beds on bedmills.
If we look at reprap, makerbot and uprint then you have a gantry system that moves and a table that raises and lowers. The uprint uses a single ballscrew and decent bearings on a cantilevered bed, the others tend to use multiple lead screws. The advantage is that you can get large x-axis travel easily but it would be harder to make it stiff enough for machining I suspect.
I don't know about the costs but that does not seem unreasonable. If I had a DLP projector however I could build that system from parts I have lying around. I may look into making one for my work as it would be useful.
Graham
Good point. Like all great pioneers have discovered, Now Is Always the Best Time to Give Up.
However, I only have dozen or so projects going at this time, so I will continue my research.
Maybe its time to clarify. Please indulge me for another post.
You guys with all the experience appear to have very clear ideas of what one can and can't do. This is very good, and finding you and this knowldege is the point of me asking. However, you also appear to expect a non experience experimenter to have the same understanding without your basis of experience, which is the driver of me asking. I appreciate your frustration with a n00b, please stay with me we are making good progress.
So we are talking budget, great. How much do the options you describe cost? What are the trade offs for each? These are the things I am trying to determine.
There needs to be some basis for a budget, as in what will my $$$ get me; I don't think its sensible start with a dollar amount and work backward when a project is in the research phase. Although I have seen businesses do it this way, their consistent failure is the basis for my opinion that the research phase is critical and must be explored and understood before setting the project in stone. (This is the term failure prone individuals haved used when they wish to skip following through on the research phase).
Incidently, I have a source of discarded equipement, from industrial microscopes to engines, and my plan is to harvest parts whenever possible.
The things that cannot be easily harvested, like the extruder head, I plan to purchase or othersie obtain later, so this is not included in the budget.
The mounting for the head need only to be designed to accept a unit of similar size and mass.
SCOPE: I am trying to limit the scope of the project to table that holds the work, the arm or bridge or gantry that holds the tool, and the motors and associated mechanicals that make it all move. I don't know if this is the right term, but I think of this project as the "motion" part of a CNC/3D printer.
We've already established that a prop will be used to control the motor, I would like to further assume that the prop is already purchased and is not included in the budget.
Also the software for CAD and any translation or conversion needed will be some popular existing software, so this is not included in the budget.
USE-CASE: Perhaps a use-case is usefull here: With the design I have in mind for a machine, the requestor makes a drawing and converts it into a stand formatted list of motion instructions.
The Requestor give this CAD development output to this machine on perhaps an SD card as a preosrted list of motion instructions. The machine (prop and mechanicals) only reads the instructions and performs them in the order given. It is the responsibility of the requestor to make the correct motions in the correct order.
Summary of SCOPE of research phase: Table and mounting for the tools. What is needed to make is move, and how strong does it have to be? What are the costs of the vaious options, or at minimum which are more expensive and which are less expensive and what are the (quantifiable) trade offs for each?
The anwers these questions will give an indication of what parts are needed, and the parts list will give us an indication of necessary cost for benefit achived, and THEN the trade offs can be determined. At that poit is where I would start scaling back requirements from millionths to thousandths to tenths and maybe back up to hundredths.
This is what I have in mind, however, I plan to accomplish the project for substantially less than $3000 (at least as far as my wife knows) since I don't plan to purchase all parts new retail.
Does this sound reasonable yet?
There are numerous factors that ultimately affect the cost of a machine. Take a screw for example, slot head zinc coated screws are reasonably cheap, but now compare that to a socket head black oxide cap screw and there is a huge difference. You can wrap up hundreds of dollars just in screws.
As I told someone once before, you are the painter (the creator), it is your canvas, now paint the picture. Your needs and desires for this machine will affect any price that we might suggest. And there are any number of other factors that will also affect the price, such as:
- How many parts can you make yourself?
- How many parts can you salvage?
- How many parts must be custom made?
- Can you use stock components for many of your parts?
- How accurate does it have to be?
- ETC...
To accurately answer these questions, you should first create drawings of what you want to build, and then create a parts list.However, I would guess that a decent "homebrew" 3D printing machine could be made for as little as $1500, but could easily reach $3000 - $5000 depending upon your needs and desires.
EDITED: Just to establish this point a little further, everyone has their own view. Take for example drive setup and the platform bed. While belt driven axises are much faster and cheaper to implement then screw driven axises, in my opinion they are much more difficult to design, but thats just me, I like belts, but I use screws. I also prefer a stationary platform bed. So you can see, we all have different opinions, and I am certain we all would attempt it in a different fashion. Every decision you make about the construction of the machine will ultimately affect the overall cost of the machine.
Bruce
I have no problems with "noobs" but as far as I can tell you have not as yet done much research on this subject and that is not because of being new, there is a wealth of info out there, when I'm interested in a subject I hit that info hard and try and get a feel for the subject. There must be thousands of CNC build threads, dozens of 3D printing blogs, patents, manufacturing websites, youtube videos, personal websites, company websites, brochures.
If research means reading this thread and asking vague questions then something has gone very wrong. Heck what do we know, there might be an idea out there with your name on we would never have thought of. I'd like so see ideas being brought to the table even if they are bad ones. If I was in your position I would have been posting all the cool stuff I found daily.
With parts from Ebay I can pretty much build a fantastic machine for very little money (I got two ACservoed, ballscrewed linear actuators by Parker for $500, and one is huge), the issue is that as a group project it is useless to go that way. If a constraint is that the output of the project is something all can build then stock parts are a must. But you could say use standardised electronics and custom mechanicals with some stock elements (like the print head).
But really a bit early to be talking like this as so far absolutely nothing has happened except for a rise in air temperature
Graham
p.s. I am sorry for my ranty style in this thread, I'm just a bit frustrated.
Like I told you yesterday, once you get your feet wet, I think you will enjoy CNC.
In a previous post, I agreed with Graham on making a 3D printing machine for making casting patterns. I assume Graham wants this for the same reason as me, which is to make parts for CNC machines.
If you become actively involved in CNC, you will undoubtedly want the same abilities, because it will save you a lot of money when you need custom made parts and you can cast them yourself. Casting is also very cool stuff.
I would suggest that you focus on a 3D printer just to get your feet wet, and then all of us, as a group, can focus on creating one. Just my point of view.
Bruce
P.S. We can figure out the cost after we have designed
Especially if you are looking for reasonably large parts, a bit of surface roughness should be acceptable. Might it be possible to automate a mig wire-welder and just build steel/aluminum parts directly?
I know that when I need to build up an area while welding manually, a series of "Dots" will often let me build up an area without melting away the substrate material, it should work for an automatic system as well.
A quick search on the internet shows that it has been "looked into", but I didn't find anything in production, The quick test I did with one of my 3 axis machines and an automatic feed hot-glue gun seems to indicate it might be a possibility for usable prototype parts.
Ken B.
Not really, I want to be able to make parts that cannot be machined. The casting part is because it allows full strength materials, not just metal but resins and all kinds of things. For making CNC machines I just use more CNC machines.
Ken,
I have seen that done using a 6-axis robot arm and a mig welder, not really a cheap business. Doesn't need 6-axis but it will cost a fair bit in gas and welding wire!
Graham
But, they wouldn't go for it... They did say that if I bought it, I could sell the circuit boards easily. So, I think that's what I'll do...
ErNA
http://www.amazon.com/Zen-Toolworks-CNC-Printer-Conversion/dp/B004I1B268
Looks like I can get the basic hardware for CNC/3D printer there and then get the extruder head from the Makerbot guys.
What do you think about the Zen setup?
If I understand right the $79 is just replacing some parts of the 7x7 CNC kit for a bigger axle distances.
A mechanical hardware kit for three axis including three stepper-motors for $330 is still a low price.
Now the most interesting question are:
- What is the maximum resolution?
- How big is the play in the axles if the direction is changed? (I assume that the axles are NOT build from play-FREE ballscrews at this price)
- What is the maximum feedrate without stalling the steppermotors?
- How does this feedrate go down if the infeed-motion goes up?
(stepper-motors with size NEMA 17 are quite small and the torque is limitied. Even if it is a 1.3A motor which is quite big for the size NEMA 17
- How does the precision of the mechanic go down (caused through bending of the frame )if bigger forces are applied when using bigger steppermotors?
I would estimate if I would like to mill aluminium with teh following parameters:
- milling-tool diameter 10mm
- a infeed-motion of 2mm
- feed-speed of 100 mm per minute the whole thing will bend that the
precision will go down to 0,5 mm. Which will be useless in most cases.
So what you have to to is using parameters like this:
- milling-tool diameter 3mm
- a infeed-motion of 0,5mm
- feed-speed of 20 mm per minute
to keep the forces low that the frame will not bend to much to keep a precision of 0,08 mm to 0,1 mm.
If you compare the two parametersets it is easy to see that the productivity goes down a lot with this machine.
I want to emphasise that this a RAW estimations from me as a guy who has only few experience with CNC-machines
Maybe the things are much better than I estimate. So this is why I'm really highly interested to get hard information.
The website of ZEN Toolworks does not mention anything about precision. I INTERPRET this that the precision is
quite low compared to CNC-machines at a pricelevel of $1500 to $2000.
So an old saying is still true here: Quality has its price
best regards
Stefan
Maybe Chris would like to tell his opinion about that.
that
It really depends on what you expect from the machine, I'd not expect to do a great deal of precision machining but with a dremel or proxxon rotary tool you could make a lot of parts better than you ever managed with hand tools. If you tool VERY light passes with small carbide milling cutters you could also do some cutting of aluminium (panel fronts) but expect to die of boredom
As a 3D printer, once it get's all that height is gets even more floppy. The design of the long travel z-axis means that in order to make use of the travel the extruder will extend a long way down from the moving carriage. I'm not sure of the weight of a typical extruder but you could end up with a fair bit of flex. Some sensible use of reinforcement could probably fix this issue.
At the price as long as your expectations are not too high I think it is reasonable. That said you might like to check out the MDF made machines on the CNC zone for inspiration, you might be able to make something better and for less.
Graham
I am sure the prop will do a better job. Just finalising some pcbs that will do the job for me. Not sure if I will get the time to do the Stepper Driver pcb this time around though. I am going to something similar to the Pulo? pcb, but a better IC variant (more robust). Too many other interesting projects to do
I think that's good enough for 3D printing since the Makerbot extruder tip bore hole is 0.5 mm.