Design thoughts for a 4x8

[Mauser]

So I've been reading and reading, almost to the point where my initial rabid enthusiasm from a month ago has slipped into paralysis. Trying to find the "Best way" to do something is getting more and more tempered by reality. There is no ONE best way. But I HAVE had a few unique ideas I'd like to bounce off the community and get a little feedback on.

1) Timing Belt drive.

I have seen on YouTube some interesting examples of a linear belt drive using a stepper motor and two lengths of timing belt meshed together. The advantage of using the meshed belts that unlike using a single belt looped over the stepper motor, there is only a very short free loop, limiting the possibility of belt stretch. One belt is glued down in a recess in the rail, the other is laid down on top of it, fixed at the ends, with a hump in the middle that goes over the drive pulley. It is held in engagement with the fixed belt by a pair of roller bearings on either side of the hump. The result is kind of like chasing down a wrinkle in a hallway carpet runner.

My twist on the idea is that instead of using two very long belts, that the moving portion could be a normal belt, turned inside out. The result being like a cross between a rack and pinion and a tank tread.

2) Keeping the Gantry square

On a big machine, keeping the gantry square is an issue. Often this is managed by having a drive shaft going across the whole gantry, or dual stepper drives, or a lead-screw under the table in the center.

My thought is inspired by the mechanism from an old Parallel Rule (Those of you who learned drafting on paper rather than computers may remember them.) So picture a setup using a loop of thin (say 1/16") aircraft cable nicely tensioned around some of those grooved bearings to link the left and right sides of the gantry so that if one side moves, the other side MUST move. Has anyone tried this and can it work?

3) Speaking of those grooved bearings...

I started off looking at ideas with the skate bearings and angle iron, then grooved bearings and the thin edge of angle iron, CRP's guides, and so on, but I think I really like the grooved bearings and the actual rails they are DESIGNED to roll on. Can anyone point me at the best price/source? I get a bit annoyed at looking at catalogs with no prices in them. Makes it hard to budget.

4) Router-mount overhang.

One of the earliest sites I visited was cncroutersource.com, which had a lot of really good information, but alas has a LOT of incomplete pages. But one that was there discussed moment arms in your typical machine. Since I've been planning to build the machine from wood, primarily Russian Birch Ply, I was very concerned with rigidity. I could see LONG moment arms caused by both the router hanging way out there at the end of an L-shaped Z-axis carrier, suspended under a very high Y-axis in the vertical plane.

I wanted to counter this by first, making the "wheelbase" of the x-axis carriage very long (almost 2' in some variants I've been contemplating), and second by having the Y-axis be in a horizontal plane, with the Z-axis carrier being between the rails. it would REALLY shorten up that moment arm.

But I can see some big problems with this approach, including that I'd basically be making TWO y-axis gantries to support a box in the middle, doubling my weight and the number of bearings and length of rail. I've considered making what I was calling the "front rail" considerably less substantial, using it to counter twisting moment, but not really carrying a lot of weight.

The primary mission of this machine is cutting shapes out of plywood, so the clearance will stay pretty low, also shortening the moment arm. But am I obsessing unnecessarily?

5) And on the topic of the Y-axis...

Triangular torsion boxes. Not much to say there, but that's what I'm picturing. A vertical member for the rails to be attached to, a horizontal member for rigidity, a bunch of triangular webs, and another piece of wood closing up the back.

Anyway, once I nail down some of these ideas, then it'll be time to noodle around in SketchUp to nail down the real dimensions, so I'd really like some feedback.

[ger21]

So I've been reading and reading, almost to the point where my initial rabid enthusiasm from a month ago has slipped into paralysis. Trying to find the "Best way" to do something is getting more and more tempered by reality. There is no ONE best way. But I HAVE had a few unique ideas I'd like to bounce off the community and get a little feedback on.

1) Timing Belt drive.

I have seen on YouTube some interesting examples of a linear belt drive using a stepper motor and two lengths of timing belt meshed together. The advantage of using the meshed belts that unlike using a single belt looped over the stepper motor, there is only a very short free loop, limiting the possibility of belt stretch. One belt is glued down in a recess in the rail, the other is laid down on top of it, fixed at the ends, with a hump in the middle that goes over the drive pulley. It is held in engagement with the fixed belt by a pair of roller bearings on either side of the hump. The result is kind of like chasing down a wrinkle in a hallway carpet runner.

My twist on the idea is that instead of using two very long belts, that the moving portion could be a normal belt, turned inside out. The result being like a cross between a rack and pinion and a tank tread.

Best belt drive ever! (If I do say so myself) - CNCzone.com-The Largest Machinist Community on the net!

I've built a prototype with a stepper (haven't tried it yet), and someone else has one working. You're tank tread idea is discussed there as well.

So I've been reading and reading, almost to the point where my initial rabid enthusiasm from a month ago has slipped into paralysis. Trying to find the "Best way" to do something is getting more and more tempered by reality. There is no ONE best way. But I HAVE had a few unique ideas I'd like to bounce off the community and get a little feedback on.

1) Timing Belt drive.

I have seen on YouTube some interesting examples of a linear belt drive using a stepper motor and two lengths of timing belt meshed together. The advantage of using the meshed belts that unlike using a single belt looped over the stepper motor, there is only a very short free loop, limiting the possibility of belt stretch. One belt is glued down in a recess in the rail, the other is laid down on top of it, fixed at the ends, with a hump in the middle that goes over the drive pulley. It is held in engagement with the fixed belt by a pair of roller bearings on either side of the hump. The result is kind of like chasing down a wrinkle in a hallway carpet runner.

My twist on the idea is that instead of using two very long belts, that the moving portion could be a normal belt, turned inside out. The result being like a cross between a rack and pinion and a tank tread.

2) Keeping the Gantry square

On a big machine, keeping the gantry square is an issue. Often this is managed by having a drive shaft going across the whole gantry, or dual stepper drives, or a lead-screw under the table in the center.

My thought is inspired by the mechanism from an old Parallel Rule (Those of you who learned drafting on paper rather than computers may remember them.) So picture a setup using a loop of thin (say 1/16") aircraft cable nicely tensioned around some of those grooved bearings to link the left and right sides of the gantry so that if one side moves, the other side MUST move. Has anyone tried this and can it work?

http://www.cnczone.com/forums/diy-cn...ock_solid.html

3) Speaking of those grooved bearings...

I started off looking at ideas with the skate bearings and angle iron, then grooved bearings and the thin edge of angle iron, CRP's guides, and so on, but I think I really like the grooved bearings and the actual rails they are DESIGNED to roll on. Can anyone point me at the best price/source? I get a bit annoyed at looking at catalogs with no prices in them. Makes it hard to budget.

Try Modern Linear or Superior Bearing. Don't be surprised if the shipping is $20 or more, so order everything at once.


[quote]
4) Router-mount overhang.

One of the earliest sites I visited was cncroutersource.com, which had a lot of really good information, but alas has a LOT of incomplete pages. But one that was there discussed moment arms in your typical machine. Since I've been planning to build the machine from wood, primarily Russian Birch Ply, I was very concerned with rigidity. I could see LONG moment arms caused by both the router hanging way out there at the end of an L-shaped Z-axis carrier, suspended under a very high Y-axis in the vertical plane.

I wanted to counter this by first, making the "wheelbase" of the x-axis carriage very long (almost 2' in some variants I've been contemplating), and second by having the Y-axis be in a horizontal plane, with the Z-axis carrier being between the rails. it would REALLY shorten up that moment arm.

But I can see some big problems with this approach, including that I'd basically be making TWO y-axis gantries to support a box in the middle, doubling my weight and the number of bearings and length of rail. I've considered making what I was calling the "front rail" considerably less substantial, using it to counter twisting moment, but not really carrying a lot of weight.

The primary mission of this machine is cutting shapes out of plywood, so the clearance will stay pretty low, also shortening the moment arm. But am I obsessing unnecessarily?
[quote/]

I've got a pretty compact design made from baltic birch ply.
Z AXIS - Moving bearings or moving rails ?

But both the Shopbot and Mechmate use the double gantry approach, with the spindle in the middle.


[quote]
5) And on the topic of the Y-axis...

Triangular torsion boxes. Not much to say there, but that's what I'm picturing. A vertical member for the rails to be attached to, a horizontal member for rigidity, a bunch of triangular webs, and another piece of wood closing up the back.[quote/]


Harder to build, harder to mount??

[Mauser]

1) Yes, that's probably the exact video that inspired me. I haven't seen anyone implement it yet on a CNC machine. Hopefully XL belts mesh nicely enough.

2) Yay! Validation. The big thing I was worried about was not in evidence, that is, nobody complained about excess resistance in their machines due to the tension. Even the variations I'd been thinking of were considered.

3) There are also some local industrial motion suppliers around here, but they make up for the lack of shipping with higher prices. The one catalog I was looking through didn't have sections of rail long enough to do the whole thing in one piece. Obviously this means some extra work to make sure the joints are perfectly even and straight.

4) I recently saw pictures of the Mechmate for the first time. One issue with the double-rail gantry is that basically your X-axis needs to be the length of your work area PLUS the width of the gantry, but since I'm looking at a table top made of two pieces of 5x5' Birch ply, and a 4x8' work area, I could make that sucker 2 feet wide! The advantage is it turns the moment arm from a class 1 to a class 3 lever. But it also complicates the Z-carriage design to take advantage of that support. I'm still thinking about that part.

I was tempted to comment in that thread when people started over-counting their travel needs. They only need to go from Z-Min with the shortest tool they'll ever use (short tool touching spoil board, not the collet, not the face of the router or the carriage) to Z-max with the longest tool they'll ever use (Maximum tool length + Maximum stock size + margin, which need not equal putting the tool even with the bottom of the gantry.) What matters most is minimizing the distance from the bit to the nearest point of support, whether it's the rails or the bearings that travel with the carriage. I saw that finite element analysis in the Bamboo router thread. It might be interesting to see what it would look like under my idea.

5) I guess it might make more sense once I draw it up. Basically the "Hypotenuse" piece would end where you need to bolt things down to the X carriage.

(Oh, and I don't recall setting "Need Help" in the title, but I guess it does help get the juices flowing again.)

[vid1900]

I really like the grooved bearings and the actual rails they are DESIGNED to roll on. Can anyone point me at the best price/source?.

Superior Bearing Company


The GW3 wheels were $16 last time I bought them and the V-rail caps were $11 a running foot.

[Mauser]

Superior Bearing Company


The GW3 wheels were $16 last time I bought them and the V-rail caps were $11 a running foot.

Interesting. That rail adds up fast when you realize you need four times the length of each axis....60 feet for X+Y alone! Now I know why people run on angle iron!

A friend wanted me to add point 6

6: Has anyone tried fabric cutting using say, a rotary wheel cutter and software to rotate it about the vertical axis, keeping it tangent to the direction of travel? Or just a Laser?

And I came up with 7.

7: If you're cutting parts completely free from the plywood, how do you make the "Tabs" or "Sprue"? Is that automatically a part of the tool path? Are there alternative part hold down techniques people use? I was briefly toying with the idea of a sort of "Rolling pin" covered with foam attached to the gantry that rode on top of the ply holding any parts down to the table as they were cut free (Which would work best if the final cuts were pure Y). I can just imagine how tricky it might be pulling a skeletonized sheet of ply off the work table with all your parts held on by a hair.

[louieatienza]

6: There was another person here who was in the planning stages of building a tangental knife cutting machine, but haven't seen progress... But it might be also done with a free-swiveling head, and having toolpaths that extend before and after the cut line... Sailmakers and others have been using lasers for fabrics for a long tme now. Just watched a helicopter factory on TV doing so with kevlar for propeller blades.

7: The tabs are created in CAM and can either be done automatically or placed manually.

CR Onsrud has a roller option for their machines for holding sheet down as you describe and even multiple sheet cutting. Vacuum fixturing is another option, as well as "onion skinning."

1: I've only seen one DIY with a completed meshing belt drive. Haven't seen it at full speed however.

4: It depends on the type of work you're doing. For me, I routinely cut 1/8" aluminum sheet, using a stub endmill, where the collet is about 1/8" above the material, so for all practical purposes that's collet-to-table. I also do work where I need clearance for any fixtures I use, and many times I could use more z clearance. For example, mounting a CNC machine clamp to my machine. Or another example: milling dhallow features on thicker stock.

2: Other than the resistance of the bearings themselves there should be little resistance due to the "rolling" nature of the mechanisms (check Rolamite thread)

Really, however, the forces for cutting plywood is not as great as say hardwoods or metal. And you can reduce that further with tooling and spindle choice.

[ger21]

Mach3 has tangential capabilities built in.

Vacuum would be far preferable to tabs.

Personally, when cutting plywood and MDF, I leave a .005-.008 skin on the bottom, pop the parts out by hand, and clean up the edges with sandpaper.

[louieatienza]

Mach3 has tangential capabilities built in.

Vacuum would be far preferable to tabs.

Personally, when cutting plywood and MDF, I leave a .005-.008 skin on the bottom, pop the parts out by hand, and clean up the edges with sandpaper.

Gerry never realized that. Given a vector file, does Mach3 automatically change the cutter angle to follow the lines?

I do the "onion skin" technique as well. Another thing I do with it is swap to a smaller bit to do the final "cutout". It rarely if ever moves the piece when cut through...

[jmoshier71]

A couple of thoughts:

1. I considered using the old block and tackle mehtod for keepign the gantry square. This has 2 issues that I can see 1) the system will not be as accurate as you think. The stretch in the cabel will easily add .030" of play over a table that big, 2) the drag caused by all of the mechanism will add to the load that the motor has to overcome.

Look at misumi. They have timign belts, bearings, sprockets, v-wheels, tracks, etc. Most of this is metric, but it is a great resource for these types of projects. They show pricing which also helps.

[ger21]

Gerry never realized that. Given a vector file, does Mach3 automatically change the cutter angle to follow the lines?

Yes, the A axis will rotate to follow the path.

[Mauser]

Yes, the A axis will rotate to follow the path.

Interesting. Although if I were to hook up a rotary cutter (think a razor-sharp Pizza cutter) what would I use to orient it? Mere steppers wouldn't necessarily zero it out. A positional encoder like you would have on a servomotor might keep it oriented, but can Mach interpret the input? The drag cutter idea (like a castor wheel) could handle orientation, but introduces problems of its own. Fabric stretches, for one thing.

jmoshier71: On the other thread discussing the use of cables, people reported the resistance being very low with ball bearing pulleys, and the rigidity increased, although you do raise a good point in favor of keeping the runs short to limit stretch. Most appear to use turnbuckles to keep the tension high.

louieatienza: I'm wondering what kind of force we're actually talking about when routing say, 3/4" hardwood veneer ply (my weapon of choice is usually a 1/2" straight carbide plunge router bit - I make lots of slots). That could be a factor in my structural overengineering. On the other hand, based on the drive belt thread, I might be looking at too light a belt. I've been considering having a drive shaft for the X-axis on the gantry, and running belt drives on both sides off of it.

I'd have to consult with my seamstress friend about the viability of a laser with the fabrics she uses. I wouldn't want to set her shop on fire! (or generate toxic fumes).

[louieatienza]

The "C" axis likely has a home switch to position at 0 degrees. If you have a spindle that would limit the amount of rotation due to the wires.

As for forces for cutting plywood, it depends... Baltic birch is pretty dense with twice the layers of regular plywood, and it's a harder wood than regular plywood. If I had to estimate, manually I'd probably need about 50-60 pounds of force to push a handheld with a 1/2" bit through plywood in one pass. Even my wimpy CNC setup is capable of 3-5 times that force. And with the free-cutting nature of CNC bits that reduces the forces needed significantly. The more HP spindle you have the better.

I've seen a few builds that run both sides of the gantry via jack shaft. If done right belts can be a very viable option.

A long time ago I worked for a sign company that had a 30W laser machine, which we used for cutting acrylics, etching lacquered brass and wood. It required us to duct all the fumes out. Table surface would be key, as the heat of the laser could collect under what you're cutting (we raised pieces off the table surface to prevent bottom burning.) You must be doubly sure the beam will not reflect off a part of a machine into someone or something flammable; our machine had an enclosure. There is also a risk of exposure to radiation; you DON'T want a laser beam touching your skin!