[Cjmovie]

Hey,

I'm part of an engineering club at my school. We're interested in building a CNC system to fabricate parts. Currently, we rely on donations from a company that has a waterjet machine (which is really, really nice) but we can't depend on them anymore. The student who's dad worked on the machine has graduated and they can't really fit us in except once a year as it currently stands (the machine is almost always tied up, especially when they get work cutting 2ft thick granite...)

We'd like to build a CNC capable of cutting, at the very least, MDF wood. It would be a huge plus if we could also cut 1/8" machinable aluminum sheeting. The largest parts we manufacture (and commonly do make) are around 60" by 40", so this is the size we'd need for the routing area.

The main catch here is our limited budget. We already have a router and bits that can be used (donated). It's fairly powerful, and we know for a fact it can cut all the materials we've listed above. We also have access to a large amount of cheap or even free 80/20 (1010 and 1515 sizes). In terms of simple electronics we also have most of what I believe we need (wiring, limit switches, anything simple and standard). The more complex controllers we don't have.
We have access to a fairly large amount of plywood and 2"x4" lumber. We have a nearly unlimited supply of nuts, bolts, etc.

As far as machining goes, we have a fairly large amount of power tools. Electric drills and drill press, band saw, table saw, etc.

Our budget is around $800. Is it possible? If not, what could make it possible?

I've looked into a fairly large number of ideas, and can't seem to figure out everything we'd need. I also attempted to design my own system, but soon found that end-supported shafts don't work too well on systems this size. At the same time, supported shafting systems are prohibitively expensive - more than our budget for a single axis alone.

[Weldtutor]

Our budget is around $800. what could make it possible?
At the same time, supported shafting systems are prohibitively expensive - more than our budget for a single axis alone.

Belt drive or chain drive systems would be a good alternative drive system to consider to get within your budget for a router system.
Your sources for motors & electronics might also have the necessary timing belts & pulleys to use for drives.

Here are some links & there are others you can search for.

http://www.cnczone.com/forums/showthread.php?t=39813

http://www.cnczone.com/forums/showthread.php?t=16634

And HERE

[jsheerin]

I'd suggest looking at Ahren's parts at www.cncrouterparts.com. I'm not sure if you'd get it done for $800 though... The 1515 will possibly have more deflection than you'd want. If you have a lot, I guess you could make some type of truss structure for your gantry and maybe something similar for your bed. Also, if you support the bed at more locations than just the corners, deflection will go way down. Personally, I used 1530 and larger pieces building my 4'x4' router.

[BobF]

Look at Joes 4 x 4 hybrid. I believe that it could be scaled up to the 60" x 40" if you used rack and pinion or belt for the long axis.

[Hirudin]

Have you checked out "linearmotionbearings2008" on eBay? They sell shafting, bearings, spindles, and ballscrews. $800 is probably still a bit low probably, but at least you could get all 3 axes.

[jcc3inc]

Sir,

If you will email me directly, I'll send you pix of a machine that I made. It is 32 x 36 and made from easily gotten parts.

[email protected]

Regards,
Jack C.

[rowbare]

You might also want to look at building a Mechmate. It would probably go beyond your budget unless you are really good at scrounging and getting material donated.

If you look through the forums there you can find people who have done group buys and had the gantry parts cut and bent (or perhaps this could be your annual water jet job).

http://www.mechmate.com/

[Cjmovie]

Hey again everyone.

Thanks for the input. From what you've shown, it does seem impossible to get a machine that's really worth the effort at that cost. We're going to be talking to the people higher up and see if we can get some more money for the project. In the mean time, we've spent time doing more research and also attempting design of our own system.



In the picture shown, the Y-Axis rail is going to be changed to vertical mounting if we go with it. The current way it's mounted was just what we went with at first, and haven't updated it yet. The dimensions edge to edge are 4' x 8', and the Z-axis travel is around 5". Acme threaded rods are 1/2" diameter. X-Axis has two steppers controlling motion, one on each side. Y-Axis and Z-Axis each have one.

The linear motion bearing system is the one for purchase here:
http://www.cncrouterparts.com/store.html
The rails for the X-Axis are 2" wide, with only one end hanging off. Y and Z axis are 4" wide.

The way it's currently set up in our drawing we're not 100% sure in the X-Axis. Do we need to have both sides doubled up, or is it OK to run with one bearing on each side like it is currently?

Also, about the electronics. The CNC control system at hobbycnc.com for 4 motors is a DIY kit, and none of us are capable of soldering components that precisely. We'd instead like to see about a solder-free solution. It seems the ones there don't require a VFD, as it's built in - is this true? We were looking at as a replacement a Gecko controller, the G540. In addition to this and our 4 stepper motors, would we need to buy anything else? (I understand there are resistors required? Is the VFD also integrated here?). I do know we'd need to buy/build a power supply of ample rating for either system, as that's not included.

Thanks for your input!

[jsheerin]

A VFD is not included with any the setups you mention. A VFD is commonly used to run a 3 phase motor / spindle. Is this what you're planning on getting? This is more expensive than using a router which you said you already had... If you want something to turn your router on and off, you need a relay and possibly a power supply for it. Add maybe $50 tops to your cost for that.

I use a G540 on my 4'x4' router with 1/2-10 5 start acme screws and Dumpster CNC zero backlash nuts. It works great for me. However, you will be limited in your top speed by screw whip. You will be even more limited if you don't use the multi start screw. I get around 250ipm with ~60" screws. At that speed they're starting to whip but not too bad. At 8' you will be limited to a lower speed. I'd look into rack and pinion, a bigger diameter lead screw (more expensive, less options for cheap zero backlash nuts available), or a rotating nut setup (a bit more expensive, but a lot more complicated to design and build) if you want fast movement, probably in that order.

[Cjmovie]

A VFD is not included with any the setups you mention. A VFD is commonly used to run a 3 phase motor / spindle. Is this what you're planning on getting? This is more expensive than using a router which you said you already had... If you want something to turn your router on and off, you need a relay and possibly a power supply for it. Add maybe $50 tops to your cost for that.

I use a G540 on my 4'x4' router with 1/2-10 5 start acme screws and Dumpster CNC zero backlash nuts. It works great for me. However, you will be limited in your top speed by screw whip. You will be even more limited if you don't use the multi start screw. I get around 250ipm with ~60" screws. At that speed they're starting to whip but not too bad. At 8' you will be limited to a lower speed. I'd look into rack and pinion, a bigger diameter lead screw (more expensive, less options for cheap zero backlash nuts available), or a rotating nut setup (a bit more expensive, but a lot more complicated to design and build) if you want fast movement, probably in that order.

What diameter lead screw would be needed, do you think? We aren't as concerned about speed as we are cutting area. Could we get away with 5/8" maybe? 3/4"? I'm not sure how to design for a rack/pinion system. In addition, the place we work in is rather dirty, and it seems that kind of system would be more susceptible to dirt and etc. clogging it?

Also, I was unsure on the necessity for a VFD because the buildyourcnc.com website has a kit for sale including both a controller and VFD's, making me think they were necessary. Our router would be manually controlled, so no need for one on that. Although the Gecko says it has two GP outputs, so I imagine we'll find a fun use for those.

[jsheerin]

Here's some information on critical speeds of screws to get you going:
http://www.nookindustries.com/acme/AcmeCharts.cfm
There's lots more on this site (and Nook's) that you can read.

Really, any system is susceptible to getting clogged with dirt. Considering there will be dust or shavings flying off your router, you will get it dirty on your own. I'd highly recommend thinking about protecting your linear motion components during your design stage. I am not a rack and pinion expert, but I think some designs put the rack facing down so debris won't accumulate in the rack. Ahren has a thread showing his rack design somewhere on this page that you could check out. Lots of other examples as well if you search.

I'm using one of the outputs on my G540 to control turning on and off the router with a relay. Works great.

[Cjmovie]

However, you will be limited in your top speed by screw whip. You will be even more limited if you don't use the multi start screw. I get around 250ipm with ~60" screws. At that speed they're starting to whip but not too bad. At 8' you will be limited to a lower speed. I'd look into rack and pinion, a bigger diameter lead screw (more expensive, less options for cheap zero backlash nuts available), or a rotating nut setup (a bit more expensive, but a lot more complicated to design and build) if you want fast movement, probably in that order.

I've seen a bunch of other posts through searching that say a screw of this size (X-axis) is a big no-no. I've also seen mention of belt drives, and it seems either that or a rack/pinion setup would be really good for the X axis (maybe even the Y axis as well?).

We have a large amount of timing belt sitting around at our shop. We used it to build a forklift style lifting mechanism (4 stages originally, could lift up to around 16' of 8020 from a starting height of 5') so I imagine it's fairly strong stuff, at least strong enough for a CNC build. I'm just not sure about tolerances etc. that may be important for this kind of build. We also have large amounts of pulleys/idlers that match the belt. I'd say we have a good 20' of it sitting around in one piece (it's all non-continuous, except for very small pieces we could maybe use for reductions?). I've searched a lot but I can't seem to get a good handle on how it'd work for us, though. What kind of speeds, what power steppers to use (I imagine just weight of gantry*u*wanted acceleration*torque at 0-X rpm*reductions?). What kind of accuracy can I expect? Any links to good plans?

[Weldtutor]

We have a large amount of timing belt sitting around at our shop. We also have large amounts of pulleys/idlers that match the belt. What kind of accuracy can I expect? Any links to good plans?

This is a LINK to a commercial belt drive system that will answer some of your questions.

It sounds as if you have most of the parts needed to make a much simpler, but workable drive.

A search here on CNC Zone for "belt drive" will give over 100 threads to read.

This is a LINK to an interesting belt driven router.

[Cjmovie]

This is a LINK to a commercial belt drive system that will answer some of your questions.

It sounds as if you have most of the parts needed to make a much simpler, but workable drive.

A search here on CNC Zone for "belt drive" will give over 100 threads to read.

This is a LINK to an interesting belt driven router.

Those are running servo motors though, right? So I assume our setup would be a good bit slower, taking into consideration both a load of 50lb+ and the use of steppers. How accurate would a home-brew system be, though? From stretch in the belt, etc. could we still manage accuracy to 1/100"? What about various methods of using the belt (I've seen at least two types of system). I've searched the forums here but can't find anything in terms of a good starting point, things to avoid, etc... Does anyone know of any handy topics in particular?

[Med-Pac]

Hey CJ,

Here's our build, we've been cutting on it for about 2 years now. I'd go with the rack and pinion. Cheaper and easier to design. Here's a few pics of ours, hopefully a picture truly is worth a thousand words becasue I hate typing. I believe there is still a thread on our build, but you may have to go back quite a ways in the thread posts.

Hope this helps ya

Mike

[BobF]

Some of the builders of Joes 4 X4 are using 1/2-8 8-start screws. This gives one revlolution per inch. Most of the rest are using 1/2-10 5-start screws giving 2 revolutions per inch.
For belt to achieve the same you need to be in the 1-2 revolution per inch range, so the motor drive pulley would need a circumference of 1/2-1 inches. That would be hard to find probably, so you will need to gear down some. A 2 inch diameter pulley would need a 2 x pi reduction to give 1 revolution per inch.

[NEATman]

If you have plenty of belt material on hand, you should strongly consider this design:
http://www.cnczone.com/forums/showthread.php?t=59570
It has the advantages of rack and pinion but less expensive, is stiffer than a long springy belt, and has no whip issues like acme or ball screws in long lengths.

Keith

[Cjmovie]

We've actually decided to make it 4x4 and use acme leadscrews for the moment, and switch to 4x8 with rack and pinion once ahren finishes his design.

Here's a little update for anyone that's interested:
http://d5robotics.org/Content/View.php?id=0

[RyanBrown]

Also heres some pictures for those lazy at heart and need the thumbnails :-p