What's new
What's new

optical pattern cutter

swellwelder

Stainless
Joined
Sep 21, 2002
Location
Valley City, ND USA
Has anyone built, or know of plans to build an optical pattern cutter? I have a pantograph head that kinda, sorta works, but it would be a whole lot more useful with something that could just use basic drawn or scanned patterns. Of course, the cost would be nice if under $500 or so. I really have no idea of what makes these things work, I assume a computer would be involved, but anyone got any ideas?

Dale Nelson
 
Has anyone built, or know of plans to build an optical pattern cutter? I have a pantograph head that kinda, sorta works, but it would be a whole lot more useful with something that could just use basic drawn or scanned patterns. Of course, the cost would be nice if under $500 or so. I really have no idea of what makes these things work, I assume a computer would be involved, but anyone got any ideas?

Dale Nelson
 
Dale, I've got an Airco optical tracer burning machine, and from looking inside the tracer, its not something you'd want to try to build from scratch. The actual tracing mechanism uses a projection lamp and a motor driven spinning mirror to somehow follow the line and keep itself centered on the line. My tracer itself is a Westinghouse, and when they quit making it Koike Aronson bought the rights and continues to make it, so I guess its a good design as tracers go, having been in production for about 25 years or so.

Tracers have several downsides. Scanned images are pretty much out of the question. The tracer has to have a crisp and relatively wide black line on a clean and opaque sheet of paper in order to trace properly. Get a smudge on the paper in the area of the line, and that drawing is no good. Kerf compensation is done by rotating the lens. So, if your blank has a hole in it, you have to run the OD with one lens setting and then the ID with the lens reset to move the head opposite. You have to be real careful with lead-in lines, because putting a purpendicular lead in means you have no idea which way the tracer is going to drive the torch once it intersects the actual drawing line. The problem here is that running clockwise reverses the direction of the offset as compared to running the same pattern counterclockwise with the same lens setting. The machine has to be attended most all the time, since you have to blow thru and start the motion based on when you think the preheat is right. Also, for multiple parts, you have to either draw multiples or move the drawing to a new place on the drawing table for each part. Last but not least is the overall size of the machine. Mine has a cantelieverd burning arm, so the area for tracing takes up an amount of space equal to the area for burning. A 4x8 machine takes up about 10'X10' space.

If I was looking to build something, I'd build a CNC burning table. With the components that are out there today, this can be done without being an electronics wizard. No drawings on paper to have to store flat in a clean place. No requirement of multiple drawings or moving drawings to cut multiple parts. The control can turn the burning oxygen off and on, and by programming a dwell prior to starting motion the machine can pierce, burn the part, oxy off, move to next start, pierce, etc etc without you having to stand there and watch it. By building a bridge type table you have the same capacity as a cantilever machine in a much smaller space. A bridge machine also allows you to use a plasma torch for thinner stuff as long as you plan for adequate speeds. A cantilever isnt suitable to use with plasma, because it will be too shaky on change of direction with the large overhang at the high travel speeds. The CNC machine can use normal G code cutter comp for kerf allowance without all the complications of adjusting the lens and so forth. Overall, the info is readily available for building a CNC table, and the parts are cheap and available too. A CNC table would do anything an optical tracer table will do, plus a lot more, and do it faster and more accurately.
 
Dale, I've got an Airco optical tracer burning machine, and from looking inside the tracer, its not something you'd want to try to build from scratch. The actual tracing mechanism uses a projection lamp and a motor driven spinning mirror to somehow follow the line and keep itself centered on the line. My tracer itself is a Westinghouse, and when they quit making it Koike Aronson bought the rights and continues to make it, so I guess its a good design as tracers go, having been in production for about 25 years or so.

Tracers have several downsides. Scanned images are pretty much out of the question. The tracer has to have a crisp and relatively wide black line on a clean and opaque sheet of paper in order to trace properly. Get a smudge on the paper in the area of the line, and that drawing is no good. Kerf compensation is done by rotating the lens. So, if your blank has a hole in it, you have to run the OD with one lens setting and then the ID with the lens reset to move the head opposite. You have to be real careful with lead-in lines, because putting a purpendicular lead in means you have no idea which way the tracer is going to drive the torch once it intersects the actual drawing line. The problem here is that running clockwise reverses the direction of the offset as compared to running the same pattern counterclockwise with the same lens setting. The machine has to be attended most all the time, since you have to blow thru and start the motion based on when you think the preheat is right. Also, for multiple parts, you have to either draw multiples or move the drawing to a new place on the drawing table for each part. Last but not least is the overall size of the machine. Mine has a cantelieverd burning arm, so the area for tracing takes up an amount of space equal to the area for burning. A 4x8 machine takes up about 10'X10' space.

If I was looking to build something, I'd build a CNC burning table. With the components that are out there today, this can be done without being an electronics wizard. No drawings on paper to have to store flat in a clean place. No requirement of multiple drawings or moving drawings to cut multiple parts. The control can turn the burning oxygen off and on, and by programming a dwell prior to starting motion the machine can pierce, burn the part, oxy off, move to next start, pierce, etc etc without you having to stand there and watch it. By building a bridge type table you have the same capacity as a cantilever machine in a much smaller space. A bridge machine also allows you to use a plasma torch for thinner stuff as long as you plan for adequate speeds. A cantilever isnt suitable to use with plasma, because it will be too shaky on change of direction with the large overhang at the high travel speeds. The CNC machine can use normal G code cutter comp for kerf allowance without all the complications of adjusting the lens and so forth. Overall, the info is readily available for building a CNC table, and the parts are cheap and available too. A CNC table would do anything an optical tracer table will do, plus a lot more, and do it faster and more accurately.
 
Munch

With a cnc burning table, how is the pattern(circle, square, whatever) generated? I would assume some type of CAD design? This sounds like what I was thinking of in my original post, still wondering if the parts can be found cheap(less than $500 or so)? I saw a burn table at a recent auction that used patterns drawm on a standard piece of paper and placed under some kind of optics? Not real sure how it worked, this was a $60,000 piece of equipment, so it probably had more to it than met my eye! I am a little familiar with the type you have, the company I worked for in the mid 80's had a burn table like yours, special ink on special paper, drawn on a drafting table. As I sit here writing, am starting to think it mighr pay to just watch ebay for one in my price range!

Dale
 
Munch

With a cnc burning table, how is the pattern(circle, square, whatever) generated? I would assume some type of CAD design? This sounds like what I was thinking of in my original post, still wondering if the parts can be found cheap(less than $500 or so)? I saw a burn table at a recent auction that used patterns drawm on a standard piece of paper and placed under some kind of optics? Not real sure how it worked, this was a $60,000 piece of equipment, so it probably had more to it than met my eye! I am a little familiar with the type you have, the company I worked for in the mid 80's had a burn table like yours, special ink on special paper, drawn on a drafting table. As I sit here writing, am starting to think it mighr pay to just watch ebay for one in my price range!

Dale
 
I have a commercially manufactured optical trace machine- I bought it new in about 1992 from C&G systems, which is a division of thermal dynamics. They mainly make CNC units, but they still make the optical units as well.
A new optical trace unit, with table, runs around 10 grand. I am not sure how you make one of something for $500 that a factory making lots of em sells for 20 times that- I guess if you get everything for free, and your time is worth nothing.

Anyway, as metalmuncher says, virtually every optical trace unit ever made uses the westinghouse design electric eye. You might find one of these used- but I am sure, new, they are a grand or so.
Then you need an electronic controller. Not a computer, but a custom made controller that converts the signals from the electric eye into voltages that control the motors that run the torch in the x and y axis. On my machine, this is a circuit board about a foot square- a few IC's, and a bunch of other miscellany. Then you need the mechanicals- racks and gears, frames and drive motors. Mine also has an electronic torch height feedback system, and a third axis motor- I consider this essential for serious cutting, as the metal warps from the heat of cutting, and will throw the torch off registration if you dont have it.
Add it all up, and its pretty complicated, and, since they are essentially obsolete in industry, a lot of the parts are not readily available.
I have little problems with patterns- I can draw a pattern by hand much quicker than I can debug an autocad file, and I use 4 foot wide butcher paper, at something like $50 per 500ft roll. I buy Sharpies by the dozen, and when I have a mistake, or even some dirt on the pattern, I find white-out works great.
But to build one from scratch would be a lot of work.
I would recommend either trying to find one used, cheap, or building a CNC version instead- the parts are actually cheaper, and easier to find in this digital age. You can go to www.CNCZone.com and find a bunch of guys who have done it- my guess is parts are closer to a grand, and build times vary. There are companies that sell kits as well.
 
I have a commercially manufactured optical trace machine- I bought it new in about 1992 from C&G systems, which is a division of thermal dynamics. They mainly make CNC units, but they still make the optical units as well.
A new optical trace unit, with table, runs around 10 grand. I am not sure how you make one of something for $500 that a factory making lots of em sells for 20 times that- I guess if you get everything for free, and your time is worth nothing.

Anyway, as metalmuncher says, virtually every optical trace unit ever made uses the westinghouse design electric eye. You might find one of these used- but I am sure, new, they are a grand or so.
Then you need an electronic controller. Not a computer, but a custom made controller that converts the signals from the electric eye into voltages that control the motors that run the torch in the x and y axis. On my machine, this is a circuit board about a foot square- a few IC's, and a bunch of other miscellany. Then you need the mechanicals- racks and gears, frames and drive motors. Mine also has an electronic torch height feedback system, and a third axis motor- I consider this essential for serious cutting, as the metal warps from the heat of cutting, and will throw the torch off registration if you dont have it.
Add it all up, and its pretty complicated, and, since they are essentially obsolete in industry, a lot of the parts are not readily available.
I have little problems with patterns- I can draw a pattern by hand much quicker than I can debug an autocad file, and I use 4 foot wide butcher paper, at something like $50 per 500ft roll. I buy Sharpies by the dozen, and when I have a mistake, or even some dirt on the pattern, I find white-out works great.
But to build one from scratch would be a lot of work.
I would recommend either trying to find one used, cheap, or building a CNC version instead- the parts are actually cheaper, and easier to find in this digital age. You can go to www.CNCZone.com and find a bunch of guys who have done it- my guess is parts are closer to a grand, and build times vary. There are companies that sell kits as well.
 
Mechanical pantograph tracer units can be very effective. Most are designed with a magnetic stylus in line with the cutting torch that follows a 3/32" to 1/8" steel pattern you have to make. The magnet is round, driven by a small electric motor driven speed reducer. The stylus rotates at a rate that drives around torch the at the feed rate to make it work properly. There's a speed control to adjust the stylus RPM.

Wood and aluminum are much easier to work to accurate but irregular shapes by hand. I've run one of these many times against wood or aluminum patterns using anything from a pencil eraser to a double cut carbide burr as a stylus. The non-mag stylus needs friction or teeth to bite to provide traction. I had to hold the stylus against the pattern by hand but otherwise it works the same as with a steel pattern.

The tricky part is making the pattern the right amount undersized for OD cuts and undersized for ID cuts but this is more careful layout than genius.

The beauty of the mechanical pantograph pattern cutter is that a handy home shop machinist can make it all using his oxy/acetyl cutting torch as the business end.
 
Mechanical pantograph tracer units can be very effective. Most are designed with a magnetic stylus in line with the cutting torch that follows a 3/32" to 1/8" steel pattern you have to make. The magnet is round, driven by a small electric motor driven speed reducer. The stylus rotates at a rate that drives around torch the at the feed rate to make it work properly. There's a speed control to adjust the stylus RPM.

Wood and aluminum are much easier to work to accurate but irregular shapes by hand. I've run one of these many times against wood or aluminum patterns using anything from a pencil eraser to a double cut carbide burr as a stylus. The non-mag stylus needs friction or teeth to bite to provide traction. I had to hold the stylus against the pattern by hand but otherwise it works the same as with a steel pattern.

The tricky part is making the pattern the right amount undersized for OD cuts and undersized for ID cuts but this is more careful layout than genius.

The beauty of the mechanical pantograph pattern cutter is that a handy home shop machinist can make it all using his oxy/acetyl cutting torch as the business end.
 
The old optical tracers are a major pain to construct (analog electro-opto-mechanical black magik :eek: ), don't work very well, and are almost impossible to maintain.

I would go CNC on this. The cost will depend on your scrounging abilities, and how much stuff you have on hand (torch, steel tubing, bearings, etc) - but if you're good, lucky, and have a parts collection to work from, it might come in at under $200. The control can be an old PC (Pentuim 100-ish or better) with FreeDOS, TurboCNC, and some Xylotex stepper drivers. The motors don't need to be exceptionally large (you aren't driving a mill with it) - 60 to 80 oz/in 200-step-per-rev motors is probably OK. Timing belt drive is probably ok (get a long timing belt, like an XL series, cut it, attach one end to the carriage, and the other end to a steel wire, run the belt over a timing pulley on the motor, and the wire over a plain pulley to the other end of the carriage)

A Xylotex card for 3 motors is $145, motors can sometimes be had for free out of old equipment (copiers, older/heavilly-built laser printers, and similar machines often have suitable steppers in them) - just make sure that they are rated for under 2.5 amps, or you won't get the full torque out of them.

Just use any program that can output DXF, it doesn't have to be AutoCAD.

Oh, and the other thing I always recommend is if practical (i.e., the machine is not out in the shop 3 miles away from your CAD PC that's at your house, or something like that), pull ethernet to the machine - it beats sneakernet anyday, and old 10baseT stuff (which is all the speed this application really needs) can often be had for the asking... If you need FTP software for DOS, just email me - I've got a collection that I assembled from around the web that handles most older network cards, and is freeware (I can be reached at yahoo.com - my screen name here is the same as my address is there).

Links:

TurboCNC
http://www.dakeng.com/

Xylotex
http://www.xylotex.com/

FreeDOS:
http://www.freedos.org/

A 10 second google search turned up this (looks like freeware or shareware - it's got a download link, and no price listed?):
http://home.carolina.rr.com/gcodemcode/gcode.htm
 
The old optical tracers are a major pain to construct (analog electro-opto-mechanical black magik :eek: ), don't work very well, and are almost impossible to maintain.

I would go CNC on this. The cost will depend on your scrounging abilities, and how much stuff you have on hand (torch, steel tubing, bearings, etc) - but if you're good, lucky, and have a parts collection to work from, it might come in at under $200. The control can be an old PC (Pentuim 100-ish or better) with FreeDOS, TurboCNC, and some Xylotex stepper drivers. The motors don't need to be exceptionally large (you aren't driving a mill with it) - 60 to 80 oz/in 200-step-per-rev motors is probably OK. Timing belt drive is probably ok (get a long timing belt, like an XL series, cut it, attach one end to the carriage, and the other end to a steel wire, run the belt over a timing pulley on the motor, and the wire over a plain pulley to the other end of the carriage)

A Xylotex card for 3 motors is $145, motors can sometimes be had for free out of old equipment (copiers, older/heavilly-built laser printers, and similar machines often have suitable steppers in them) - just make sure that they are rated for under 2.5 amps, or you won't get the full torque out of them.

Just use any program that can output DXF, it doesn't have to be AutoCAD.

Oh, and the other thing I always recommend is if practical (i.e., the machine is not out in the shop 3 miles away from your CAD PC that's at your house, or something like that), pull ethernet to the machine - it beats sneakernet anyday, and old 10baseT stuff (which is all the speed this application really needs) can often be had for the asking... If you need FTP software for DOS, just email me - I've got a collection that I assembled from around the web that handles most older network cards, and is freeware (I can be reached at yahoo.com - my screen name here is the same as my address is there).

Links:

TurboCNC
http://www.dakeng.com/

Xylotex
http://www.xylotex.com/

FreeDOS:
http://www.freedos.org/

A 10 second google search turned up this (looks like freeware or shareware - it's got a download link, and no price listed?):
http://home.carolina.rr.com/gcodemcode/gcode.htm
 
Swellwelder,

Very interesting subject: the machine Ries refers to was purchased from C & G Systems, and in 1992 I was the owner of the company, (the "C" in C&G) and I designed the machine which he bought!

If you want a small, low cost machine, get a magnetic template follower from one of the small machine makers; these are the cheapest. If you need a larger size, you might be able to find a used machine somewhere for a low price.

A "pantograph" is a kind of non-descript term which doesn't tell us a lot about exactly how the machine is built; if you can describe it in some more detail, perhaps I can suggest something. Other interesting info would be whether you do plasma or oxyfuel cutting, the max size of a single piece, the thickness, etc. From your comments, I gather that you may not use this machine a lot??

Regards,
Jack C. Cain
 
Swellwelder,

Very interesting subject: the machine Ries refers to was purchased from C & G Systems, and in 1992 I was the owner of the company, (the "C" in C&G) and I designed the machine which he bought!

If you want a small, low cost machine, get a magnetic template follower from one of the small machine makers; these are the cheapest. If you need a larger size, you might be able to find a used machine somewhere for a low price.

A "pantograph" is a kind of non-descript term which doesn't tell us a lot about exactly how the machine is built; if you can describe it in some more detail, perhaps I can suggest something. Other interesting info would be whether you do plasma or oxyfuel cutting, the max size of a single piece, the thickness, etc. From your comments, I gather that you may not use this machine a lot??

Regards,
Jack C. Cain
 
Ries,
Thanks for the butcher paper suggestion. Never thought about that and coulda bought a lifetime supply for what I've spent at a local artists supply store on paper.

Depending on what you cut with yours, you might want to consider adding a water spray head if warping is a problem. Mine has a ring that mounts around the torch just above the tip retaining nut, with a series of holes drilled in it that surround the cut with a bunch of little water streams. Its got a solenoid that is activated when the cutting oxygen solenoid activates, and a needle valve to adjust flow. Most of the burning I do is in the 1/2" to 2" range, and warping isnt a problem. But, I've burned quite a bit of 3/16 and 1/4 T-1 plate, and the water works great on that. Without it, thin T-1 will sit there still as a mouse and then suddenly jump up maybe half an inch almost instantly and slam into the tip. With water, the material won't move around, and the cutting oxy keeps the water blown away from the cut itself just far enough that there's no problem with it quenching the burned edge. Water is also helpful when burning small parts from thicker stock where the part you're cutting out wants to get red hot and the burned edge finish begins to deteriorate. De-slagging small parts that have been water cooled is as easy as it is on larger machine cut parts, whereas I find if I let them get really hot then the slag is actually a mix of slag and molten steel, which can be difficult to remove.
 
Ries,
Thanks for the butcher paper suggestion. Never thought about that and coulda bought a lifetime supply for what I've spent at a local artists supply store on paper.

Depending on what you cut with yours, you might want to consider adding a water spray head if warping is a problem. Mine has a ring that mounts around the torch just above the tip retaining nut, with a series of holes drilled in it that surround the cut with a bunch of little water streams. Its got a solenoid that is activated when the cutting oxygen solenoid activates, and a needle valve to adjust flow. Most of the burning I do is in the 1/2" to 2" range, and warping isnt a problem. But, I've burned quite a bit of 3/16 and 1/4 T-1 plate, and the water works great on that. Without it, thin T-1 will sit there still as a mouse and then suddenly jump up maybe half an inch almost instantly and slam into the tip. With water, the material won't move around, and the cutting oxy keeps the water blown away from the cut itself just far enough that there's no problem with it quenching the burned edge. Water is also helpful when burning small parts from thicker stock where the part you're cutting out wants to get red hot and the burned edge finish begins to deteriorate. De-slagging small parts that have been water cooled is as easy as it is on larger machine cut parts, whereas I find if I let them get really hot then the slag is actually a mix of slag and molten steel, which can be difficult to remove.
 
In my employer's shop we have one of those wonderful flame cutters. It is the big lump of trash covered in dust. As much smarter men have pointed out; it needs babysitting and it gives marginal results. I grew tired of printing full size blueprints with wide lines. I had to print a different sheet for each cutout and hope it didn't screw up our only piece of plate. The cut and weld guys cussed me and the cutter.

When we need plates burned out, I email the .dwg files to Accro Tool in the woods of Penn's land. They give them to either the laser cutter robots or let the water jet do the cutting. I get perfect parts that look like sandblasted milled edges. There is no warp, no slag, 0.005" accuracy in thick slabs, and it is cheaper than that damn torch.
My boss would probably give me a pass and let me haul it away. He gave me a pass for a 24 by thirty inch milling machine to make printed circuits and other things. We just paid twenty grand to have the OEM company rebuild it. It gave marginal results, the parts sucked, and the total cost was no bargain. I have a guy in Chicago, Ontime Circuits, that will make perfect circuit boards, overnight. I email the gerber file and the parts arrive next day UPS.
My hillbilly advice is to skip the optical stuff and go directly from computer to cutter. If you can plot the part, just let the cutter follow along and torch the parts. My cheapskate advice is to farm it out and save some money.
 
In my employer's shop we have one of those wonderful flame cutters. It is the big lump of trash covered in dust. As much smarter men have pointed out; it needs babysitting and it gives marginal results. I grew tired of printing full size blueprints with wide lines. I had to print a different sheet for each cutout and hope it didn't screw up our only piece of plate. The cut and weld guys cussed me and the cutter.

When we need plates burned out, I email the .dwg files to Accro Tool in the woods of Penn's land. They give them to either the laser cutter robots or let the water jet do the cutting. I get perfect parts that look like sandblasted milled edges. There is no warp, no slag, 0.005" accuracy in thick slabs, and it is cheaper than that damn torch.
My boss would probably give me a pass and let me haul it away. He gave me a pass for a 24 by thirty inch milling machine to make printed circuits and other things. We just paid twenty grand to have the OEM company rebuild it. It gave marginal results, the parts sucked, and the total cost was no bargain. I have a guy in Chicago, Ontime Circuits, that will make perfect circuit boards, overnight. I email the gerber file and the parts arrive next day UPS.
My hillbilly advice is to skip the optical stuff and go directly from computer to cutter. If you can plot the part, just let the cutter follow along and torch the parts. My cheapskate advice is to farm it out and save some money.
 








 
Back
Top