Hello all, Is it possible to CNC old parts from a clear picture? I'm thinking things that are mostly flat with fewer layers unlike an intake manifold with multiple layers. I know there are planers that will take a photo and transfer the image to a wooden panel. Is the same true for a CNC machine?
wood is easier,take the pic and multiply it by scale (1:24 =24 times the size in the pic for 1:1) CNC need actual dimensions,or a print to work with
If you have the hard part some prototyping and R&D machine shops can create duplicates using coordinate measuring techniques and CNC machining. It is an expensive process. If you are working from a photograph it would be better to start with the critical dimensions and design a new piece with the right look.
laser scan or inspection arm could get a solid model for cnc program.We use laser and inspection arm quite often to gather data we dont have.Would say it would probably cost a minimum. of $500.00 to do a simple part.
Create a solid model in Autocad then send the file to a prototype service. They can make stuff out of a lot of different materials these days. You can also generate tools path from your solid model and send it out for CNC.
Ain't that the truth... I work for a prototype CNC shop and the parts I machine can cost a pretty penny....
That sounds a lot like an approach to generating architectural base drawings from old blueprints I find a**** young CAD draughtspeople in lots of architectural offices - and try to stamp out wherever I find it. They would scan a blueprint and import it into the CAD programme as a raster image, and then trace over it without orthagonal constraint or any attention to dimensions. The result is an irrational mess that I can't work with when do***enting subsequent construction. The proper way to prepare this sort of base do***entation is to study the original and try to get into the original designer's head, especially regarding dimensional decisions. Everyone will start with a base decision, and refine the design from there. The later the dimension is decided in the process, the more likely it will be fractional or otherwise complex; the earlier it's decided, the more likely it will be in whole units or sensible multiples. This doesn't always help with very complex designs, as things get revised and revised again throughout the process. Also keep in mind the measurement system the original designer was likely to have used. Something about 2' long and known to have been made in France is likely to measure 600mm on the dot. Then again, it might not. If you can't get to the original artifact, get as much information as you can about other stuff it has/had to interface with (what does it bolt up to?) and reverse-engineer from there. If you're working from photographs you will have to understand perspective very well indeed. It helps if you have access to several photographs taken from different angles. The greater the variety of information you have to work from, the closer to the original you can get, even if that might not be all that close in the end. Alternatively you can use the trace-over-scan method to produce a .dxf and have something vaguely inspired by the subject of the photograph. It depends on what you're trying to achieve.
Anything's possible. This guy recreated, almost entirely from scratch, an entire 1919 Excelsior board-track racing motorcycle, starting with nothing more than two corroded engine case halves and an 85-year-old B&W photo that showed only the right side. Doc.
With a few dimensions, I can input any part into a 3D modeling CAD program called SolidWorks. I am an Industrial Designer and have been doing this for over 20 years. I've made parts for fellow Hambers and Inliners for years. Send me a PM.
The process is known as raster to vector conversion to convert bitmaps, jpegs, etc. to machinable lines and arcs. Most higher end CAM systems have this function built in.
I am an Industrial Designer as well and use Solidworks. Sometimes a picture can work if you have at least one dimension and its a 2d part. I have taken pictures of parts before and pasted them into Solidworks and drawn from the picture. It was fast and worked well. Just have to note one dimension and scale from there, if you have some critical dimensions it can get a little trickier. Other than that you would need to reverse engineer the parts you have that are complicated. We use Faro Arms, inspection tables with height gaues, dial indicators, mics, calipers etc. From their I open the part in a CAM program and tell the machine how to cut it. From there the machine is setup with material and once positions are set at the machine and tools loaded, a part can be made.
Anything is possible nowadays . . . just depends if you have the coin to spend http://www.jaylenosgarage.com/extras/articles/jay-lenos-3d-printer-replaces-rusty-old-parts-1/ DLMS is a great alternative if you want to do a one-off prototype as well considering the sunk cost of programming, tooling and machine setup time for a CNC deal
I normally make paper templates, I have my screen calibrated 1:1 and use Rhino 3D to make lines. Then I CNC plasma. Do it all the time. I'm a 3D Surface Modeler by trade. I have also 3D Printed parts like in the video above. You can also 3D Print molds and cast off that. It's all about $$$.
All, This is really great info thanks for all the responses. I have been playing with the idea for a long time and wanted to know if it was possible. I don't have a good unobtanium part yet but hopefully I will soon. Thanks again!
The problem that I see with this is that a lot of the unobtainium parts that spring to mind are cast.
I've been reverse engineering parts for 30+ years. A paper describing the process used is available for free download from my company website [www.beytagear.com]. The parts in the paper were for a 1990s Formula 1 car. Low volume parts are well into the $1500+ per item range. On one of the weekend shows they made a prototype rear end cover with a 3-d printer. Uses weed wacker like plastic string to print out an actual part from a 3-d CAD file. This will eventually drive costs down for low volume aluminum castings; the plastic model will replace the wood/aluminum pattern. If shrink allowance is needed it can be scaled into the 3-d model. Castings will still need machining. Chuck Schultz Winfield, Illinois
Hadn't thought of using 3-d printing that way, although its kind of a natural. That has a lot of potential.
That's the obvious and sensible way. There is a lot of talk about sintered metals these days, which seems a bit limiting. I think some people are getting starry-eyed about visions of Star Trek replicators. No use making a fetish of the thing.
I've heard on a recent TV show that people are making semi-automatic gun parts with prints and sketches online and using the 3D printing/plastic method to make exact replicas of parts not produced and actually using them to ***emble a working gun.. pretty crazy to think what else could be made with this method...
I was contacted today by a start-up company that plans to introduce a 3-d scanner that would build a 3-d model suitable for use in 3-d printing. It wouldn't help with stuff that has lots of internal details but for trim stuff it would be great. Now we need a tech thread on home built investment casting.
http://www.jalopyjournal.com/forum/showthread.php?t=661477 http://www.jalopyjournal.com/forum/showthread.php?t=656069 http://www.jalopyjournal.com/forum/showthread.php?t=636663
Hi Doc....Paul Brodie is a wizard building this machine from scratch...there not cheap..like...$130,000....I'm into this type of bike and there is a lot of intrest in this era of bike board track racing.....
The Jalopy Journal
