New Machine Build- 16' x 5.5' Fabric Cutter/Plotter

[Pyronaught]

This is quite an impressive machine! I am looking at building something like this myself. I have a few questions:

1. stepper size and rating
2. chain (looks like ANSI 35 but I can't really tell)
3. it looks like a 9 tooth drive sprocket, is that right?

The answers to these questions would be really helpful to be able to roughly replicate the gantry setup you've got. I'm looking at doing a rotary cutter so the 1300ipm rapids you talked about are really desirable.

I can see how you got the Y rails parallel, use precisely cut t-slot sections and perhaps no adjustment is necessary. But how did you ensure that your X rails were parallel? It seems like being really careful on the woodworking part can only get you so far. I saw that you used aluminum angle to mount the makerslide to the table. Did you just drill oversized holes so you could slide things around and use an 8ft level to align all the sections to one another? Does the lack of aggressive gusseting on the gantry riser help with the parallel tolerance issue? What I mean is that it looks like it could flex a bit which would mean that instead of needing to be parallel to a few thou you could get away with a 0.020 or 0.030 and still be OK.

Thanks for any help you can give!

The chain is 1/4" pitch standard roller chain. The drive sprocket was 9 tooth, but don't buy them from buildyourcnc.com because both the sprocket and idlers they sell are total crap quality and will cause problems. I forget where I finally wound up buying good ones though, I'd have to look it up. Motors were NEMA 23, except the motor that spins the tool head which was NEMA 17. I should have just put a NEMA 23 on that one too because when the cutter wheel digs into the table surface the motor can't turn it sometimes and it loses steps.

The rails are not mounted against the edge of the table, so they are not effected by inaccuracies in the table edge. They sit out away from it by maybe 1/4". I used a metal spacer bar to get the rails parallel. I believe I installed one side first and then used the spacer bar to align the other side before screwing them down. I also ran the gantry back and forth to make sure there was no resistance or binding as a double check for the parallelism before finally screwing down the other side. It is one continuous rail on each side, so I could clamp the ends to keep it from flipping over while aligning it. The mounting holes are not really oversized, so there is not much play.

Sorry for the slow response, I didn't have this thread setup to notify me when people post here.

[Pyronaught]

I've also changed the way I'm making the vacuum boxes. I had to modify my process for vacuuming one sheet of film over top of another sheet of film, with the vacuum sucking each of them down individually. This involved routing a vacuum channel around the outside of the work area, applying the first layer of film over top and turning on the vacuum, then using the machine to punch holes at intervals around the channel in order to pass the vacuum to the second sheet. So since each different job requires a different channel pattern, I had to rebuild the vacuum boxes with removable top surfaces. It turns out you really don't need anything but a screw in each corner, since the vacuum itself will suck the board down and hardboard lays pretty flat on its own anyway.

So when rebuilding the vacuum boxes, I just cut the support strips and then grouped them all together and cut notches through all of them at once with a dado blade. The boards were arranged into two staggered groups so that the notches would be staggered like you see in the image. The boards were then glued to the table one at a time using weights to hold them flat till the glue dried. I could glue two or three strips a day like this and get a whole panel done by the end of the week. This was much faster than the way I was building vacuum boxes initially, and the resulting table surface is more level as well.

[Pyronaught]

Pyro,

Great project! It actually inspired me to build my own CNC fabric cutter (I make super light hiking gear as a hobby) . Mine is built from some salvaged Isel slides and 80/20 drops. I'm attaching some pics of my build below. I have some questions about your post processor and sent you a PM. Cheers!

Nice design, metal table is really the way to go if you can afford it. After a year of sitting in my garage with no climate control, my wood table did shift a bit and lose some of it's levelness. I shimmed the surface to re-level it, but what a hassle. If I had it to do again I would probably build it more similar to yours, and also possibly go with steel wheels on a steel rail mounted similar to how your slides are (but still chain drive). The reason is that the plastic wheels wear too much, but the maker slide method is still probably the cheapest solution there is, and it's lighter weight allows for faster speeds. The edge mounted rails like what I did do save some width on the table though, and space was really confined where my table is located so I pretty much had to do it that way.

[brownfox]

Nice work, Pyronaught! I'm very impressed.
I'm involved in the same business in Russia, that is, making balloons and other works of plastic, foil, cardboard, foam, etc. My goal is to make a machine like yours. Currently i'm retrofitting a very old pen plotter. By this moment, I get it move under Mach3, now working on a toolhead. Can you please give us more details on how did you make it?
Specifically -
1. What is the transmission from the stepper to the tools? Gears? Timing belt?
2. What is the coupling between the transmission and the piston rod? How do you get it slide and rotate at the same time?
3. How did you make the hot wheel?
4. Do you have a home switch on the tangential drive? How do you make an initial alignment of the knife?
Thank you and advance and best wishes.

[Pyronaught]

The transmission between stepper and the rotating tool holders is spur gears. The stepper has a pinion gear on it and then each rotating tool holder has a large gear driven by this central pinion gear. Both shafts are driven by a single motor since only one of them would ever be deployed at a time.

The piston rod runs through the rotating shaft and is made to rotate along with the shaft by cutting a slot down the full length of the piston and then having a pin sticking through the rotating shaft that keys into the slot and forces the piston to rotate with the shaft while still allowing it to move up and down.

A more detailed account of the development of the hot wheel can be found here: Metalized Film Envelopes - RC Groups

I do not use a home switch on the tangential drive, I just align them by sight each time the machine is powered up. The problem with using a home switch is that you often need to rotate more than 360 degrees, so the home switch would limit you to less than 360 unless you had a way to disable the switch when the machine is in use. I might add this feature later, along with a second home switch on the other gantry rail so that both sides of the gantry can be homed independently. With only one home switch on the gantry it is easy for it to get out of alignment and still be racked even after homing it.

[brownfox]

Thank you very much for your quick reply.
I understand the principle of sliding gears in general. Can you make a closeup photo? Well, I don't ask you to disassemble the unit for that Anyway, things may be different when I get the cylinders.
Concerning home switches, I don't see a great problem here. If a switch is programmed as a home switch, not a limit switch, it doesn't affect regular movement. If you are short in input pins (as it usually is with LPT port), you can tie them together - Mach3 is smart enough to perform homing one axis a a time. Of course, it's very desirable to have separate home switches for dual drive systems - thus only 2 input pins are required.
Per my previous experience, the best home switches are optical with Schmitt trigger (for example, HOA2001-001) - this ensures excellent repeatability. For size-critical places, a sensor from an old mouse may be used.

[Pyronaught]

In this picture, on the right side tool you can see a nut on the piston shaft. That nut is holding a small bolt that protrudes into the keyed channel on the piston so that when the sleeve turns the bolt will cause the piston to turn with it. http://www.cnczone.com/forums/attach...0&d=1383848059

The picture below shows the CAD image of the piston, except I only used one channel instead of the two seen in the image. One is all you need. The trick with this setup is you have to have the center hole where the pneumatic actuator attaches be dead-on center. Even the slightest bit of off-center will cause the piston inside the actuator to bind and cause the piston to get stuck. This is the area that was giving me problems, as no matter how careful I was with the lathe I could not produce an absolutely perfect center hole. I don't know if maybe my lathe centers were off a tiny amount or if maybe the actuator itself had some kind of inaccuracy to it but in the end I had to couple these two components loosely with enough play to keep them from binding. The alternative to this setup is to have the entire motor assembly ride up and down with the tool, which I've seen in several designs. My first tool head was like that, but with two tool heads being driven by one motor you can't easily do that and it's also more complicated.

[Pyronaught]

Even if the home switch on the rotating shaft is just programmed to be a home-only switch, you still have to have a setup that can be actuated from either direction. Since you will want to be able to do more than 360 (I like the ability to go about 45 degrees past one revolution), the home switch needs to be able to be toggled from either direction instead of the usual single direction. There would have to be a nub hitting a double sided ramp or something like that where the trigger could hit the switch and keep going without interfering with the motion, then hit it again on the way back.

One thing I like about working with the rails derived from 80-20 profiles is that you have a T-slot running the entire length of the rail. Thus you can put your limit switches on sliding plates and adjust them to shorten the working area of the table as needed. A lot of times I'm only using half the table for cutting and the other half turns into a workbench. The last thing you want is for the machine to go plowing into a bunch of stuff sitting on the table, so it is a simple matter to change the switch positions and keep the gantry from ever accidentally going where it shouldn't.

[pippin88]

Or a hall effect sensor

[Pyronaught]

Or a hall effect sensor

I was thinking one of those optical IR emitter/detector pair type sensors, then just put an arm on one of the gears that swings through the gap. Clearance around the edge of the gear is pretty tight though (see below).

[brownfox]

Optical sensor - see how it looks at my router (Z axis). Work well for a few years, the repeatability error is well below 0.1 mm. As for hall effect sensors, I don't believe they provide the required precision. Same for inductive sensors.

[Pyronaught]

Yep, that's the IR emitter/detector pair I was talking about. Back in the 80's before they made those one piece molded pairs you used to have to buy the IR LED and the transistor separately and make a little PC board that held them pointed at each other like that. This would work perfect for the A axis home switch though, I'll have to add one when I get the time.

[elfrench]

Would a thin layer of MDF (.25") make a better surface for the vacuum table?

[Pyronaught]

Would a thin layer of MDF (.25") make a better surface for the vacuum table?

That is actually what I'm using in the pictures here, but I had to upgrade it to 3/4" MDF because the thin stuff was causing problems when heat sealing seams. Even with how close together the underlying support is, I still had issues with 3/16" MDF flexing and causing problems with seam quality.

[WMarineCanvas]

This post has gotten my interested in building a plotter/cutter table for my marine canvas shop. I see that there is an absolute wealth of information here and I think the table build sounds pretty straight-forward but where should one start as far as the mechanics of the gantry and tool setup go? Also, would a build like this be compatible with existing plotter/cutter software or is that stuff completely proprietary and essentially a "closed box" which would require a custom build to operate on custom code? Thanks!

[Pyronaught]

The gantry is pretty straight forward, but the tool head details is the hard part. Especially if heat is involved like what I'm doing. Looking at the pictures here should give you some ideas though.

You can mostly use off-the-shelf software to do this, depending on what you are doing. I did write a custom post-processor to do things like remove all the unnecessary G code and all Z-axis references from the G1 commands in the G code files generated by my CAM software. Originally I was adding my own angle commands to implement the tangential cutting, but ran into problems with that and am now just using the tangential features built into the Mach3 software. The only other custom thing I added was the ability to auto-generate registration lines, which you are probably using if you are making sails. Otherwise you would have to manually add the registration line marks in your CAM program or the original CAD file.

I'm not familiar with any commercial fabric cutting software so I don't know how programmable they are. I'm sure some are customized for one brand of machine made by the same manufacturer, but there might be more generic solutions out there not tied to any one machine. I messed around with Sheet CAM for a while but had better luck with CAM BAM so that's what I use. I believe Sheet CAM has better nesting features though.

[tropostudio]

I hadn't checked this thread for almost a year. Cool to see all the new information, and this is really the best source for any info on a DIY CNC fabric cutter. Definitely a dearth of info on these machines, and frankly the costs of commercial machines seem really high for how basic they are. Supply and demand and economies of volume production, I guess.

After looking at your posts on rotary heat sealing in the RC Groups forum, I have a few questions:

1) What is the supply voltage for the cartridge heater? Does it run off an independent power supply, and how do you control heating?
2) How do you insure heat transmission between the cartridge heater and the graphite impregnated bearing? Thermally conductive grease? Does the bearing rotation wear on the heater cartridge?
3) Do you think this basic design would work for a rotary hot knife?

[pdidefense]

Hey guys! I am looking at building a CNC Fabric cutter to process various grades of Kevlar - I don't need anything bigger than 3ft X 3ft. Could anyone lead me in the right direction for a DIY tutorial? How much are you guys spending on building these units? Thanks for all the help in advance!

[cyclotron]

here is mine finally working! You were a big inspiration Pyro!

[Mtnclimb007]

Cyclotron, That is awesome! Would you be willing to share some details on how to build?