Finishing up my second CNC

[Chemeleon]

About two months later than I'd originally intended, but I'm finally pretty much done with this new CNC router. My first one was an all MDF build (heck of a lot cheaper than metal, and could always reuse the movement parts on a metal build), about 3'x4' cut area. Definitely a usable machine, but the flex caused by it being mdf, and using lower quality linear slides (skate bearings and angle iron) limited its speed and accuracy. So a few months back, I decided I wanted a second machine, designed to be as stiff and fast as I could make it, but sized so I could put it on my desk in my office so I don't have to run out to the shop to cnc things.

Once I decided to make it, I also decided I would document everything and publish a guide on building it. I need to finish that off still, but its about 2/3 written, 20 pages of pure text already, and I've got probably 100+ photos from the build to add to all the instructions, plus formatting, etc, so I suspect my "guide" is going to be more of a book

All that's left now is to wire up the electronics, and let it cut the final aluminum router mounts to replace the temp wood ones - I didn't feel like hand routing a 3.5" diameter hole in 3/4" aluminum plate . Should have it up and running tomorrow afternoon.

As to specs - framing is 3"x3" 15 series 8020, with 1/2 and 3/4" aluminum plate for the gantry carriage, chinese 15mm profile rails and bearing blocks, 425oz nema23 stepper motors, and 1/2" acme leadscrews with 0.5" lead. Cut area is just a hair over 12"x12" with 3-4" of Z clearance.

 

[jaustin]

Looks pretty good,
I started my learning into cnc with a carvewright, now I have converted a Little Machine Shop mini mill into a cnc mill.
My problem is learning the software. i got the basic down now just have to find time to get back messing with it.

i am using CamBam and Mach3

 

[Windy]

That's great, I think our next build will be a portable unit for small parts like this. We're still working on finishing up our 5x10 build.

Finishing up CNC'ing out the left/right gantry sides



Gantry assembled, this is almost 7' long and well over 100lbs


Table finally finished welding and ready for painting, the table weight alone is over 1000lbs, been using a skid steer to move it around.



Finished rendering


Gantry rendering



Keep us updated - Windy

 

[Chemeleon]

nice looking machine I'd considered going green as well, but decided I didn't want to spend the extra on custom extrusion colors. My original goal for this little machine was basically just a dev machine - I write my own CAM software, wanted something on my desk for testing code output. Now though, I think its primary purpose will be as a production machine (hopefully anyway). Awhile back, I wrote a specialized little cam app that'll generate code for machining dovetail jigs, as well as setup a website where you can layout and order a custom jig design - tail/pin spacing and size, bearing diameter, dovetail bit angle, etc. Plan is once this machine is running and I've done some tests on its capability, to try selling those jigs, machining them with it. Seems like they ought to sell well though - every woodworker it seems buys at least one jig, often multiple, and mine give you the flexibility of the adjustable jigs, with the simplicity of the keller/charles neil style ones, at a much lower price

 

[Carl Fisher]

Yes please on the write up! I'm starting mine next month and am still trying to plan out my purchase list.

 

[Chemeleon]

Its alive, alive! Muahahaha! <cue thunder>


Though, it did lose a little magic smoke - turned out the one stepper driver was bad. Didn't seem to be obeying the microstepping settings to start with, then after a few minutes, started to smoke. For now, just borrowed one from the other machine.

So far, have it running at 200ipm and accel of 75. The accel in particular is a *huge* step up over my first machine, where I could only do 10-15 before cut quality suffered. At 75, its essentially instantly to full speed it seems, which speeds things with lots of direction changes like 2.5d carving up a ton, and I'm getting zero noticeable vibration or resonance in wood so far. The real test will be some brass and aluminum I need to cut tomorrow.

Video of its first test, no fastforwarding:
[video=youtube;6eQPMv10RR0]http://www.youtube.com/watch?v=6eQPMv10RR0[/video]

 

[CrealBilly]

I must have one... i could rough out semi hollow carved top and back guitar bodies with that. In a fraction of the time it takes by hand.

 

[Carl Fisher]

That thing is FLYING!

What stepper motors did you use and what lead screws?

 

[Chemeleon]

Carl, just generic chinese 425ozin 3amp Nema23 steppers, connected to a 1/2" 4 start leadscrew from Roton, moves 1/2" per rotation.

First paying job done with it - couple 1/8" brass nameplates to go behind some brass handles for another woodworker making a tack trunk.

 

[Alan in Little Washington]

What kind of linear bearings are you using and where did you get them?

I'm looking for some that are economical, have no slop, and slide with no or absolute minimal drag. I have pretty much ruled out anything with plastic (Delrin, Nylon, etc.) slides. The ball bearing linear guides I've looked at (McMaster-Carr, etc.) tend to be too expensive for my application. For my application the bearings can ride on precision ground rod or extrusion slots. Any thoughts?

 

[Chemeleon]

I used 15mm ball bearing linear guides and profiled rails, ordered through aliexpress from china. These ones in particular:

http://www.aliexpress.com/item/1pcs...ndard-Blocks-CNC-Router-CN455/1239923327.html
http://www.aliexpress.com/item/2pcs...r-Guide-CN457its-Linear-Guide/1240137476.html

The guide rails say "Free shipping" in their title, then charge it, so I'd emailed the seller to ask about it. He offered to ship free via fedex, and knocked a couple extra bucks off as well. Ended up running $278 shipped for the 3 sticks of 1m rail, and 10 bearing blocks, and they arrived in about a week (which is unusually fast from china, presumably because he used fedex instead of the cheaper alternatives). So far, they seem to work very well; no noticable flex anywhere in my machine. They don't glide like some other options do, but I've yet to see any downside to that. Would have been nice to have adjustable preload, but doesn't seem necessary with them so far.

The other sort I particularly liked, were Redi-Rail from PBC Linear, like these: http://www.pbclinear.com/RRS-Redi-Rail-Linear-Guides---Metric-Slider . They sent me a sample, which seemed to move very securely, easily attached, adjustable preload, etc. But they'd never reply to my request for a quote, and the prices I found on ebay and the like were significantly higher than I wanted to spend.

A friend with a cnc used the typical supported round rails and split bearing blocks for his. He's said it works pretty well, but he thinks its the cause of some small amount of flex in his machine.

Overall, if/when I build a third, bigger machine, I'll most likely go with those profiled rails again, as they just seem to have the best balance of stiffness and price.

 

[SubGuy]

This thread has gotten me looking into this CNC world. I really never gave thought to building one before, but I am now. TY.

 

[batk30msu]

I was wondering if you could give us a price range for the project. I know it cannot be cheap but do you save a ton when you DIY? Thanks.

 

[Chemeleon]

Its really not that expensive as far as tools go. This one cost me right around $1000 to build. To be fair, I did pretty much everything by hand, where a lot could have been ordered premade - made my own anti-backlash nuts with a homemade acme tap from some extra leadscrew, did all the 8020 cutting/drilling/tapping when I could have paid them to do it, etc. If you bought things premade as much as possible, it'd be even easier to build, but probably run closer to $1500.

If you wanted to build a larger version, it'd be probably around $50-100 more per foot per axis. I'd say you could probably scale up to around 4'x4' without any real modification to the overall design, though at that point, due to the moving table design, you're looking at a 4'x8' footprint, so I'd probably switch to a moving gantry design at that point to save floor space.

 

[bguil]

I'm just wondering if there are enough users of cnc out there to do a little review with pros/cons of available cnc machines or kits such as Shark, ShopBot, etc.?

 

[Chemeleon]

Keeping in mind I've not seen these in person, here's my general thoughts on things:

The Shark's main pro seems to be that Rockler sells it, so its easily discovered and bought. I think its overall design is too weak given its price though, in particular, the unsupported rods for the linear movement. Obviously its a functional machine, plenty of people are using one, but those're limiting its stiffness, and thus speed, quite significantly I'd say. It does include Vectric's VCarve and Cut3D software though, which is about $900 or so worth of CAM software. From what I've read, people tend to run it around 50-70 inches per minute, which is ok for cutting, but if you want to do 2.5d relief carvings, it'll take quite awhile. At that speed, you're looking at something like 16 square inches per hour give or take.

ShopBot's are made in NC, so they've got the local advantage I suppose. I've never really heard a bad thing about them; they seem to be solidly built, well liked machines, they're just more than I can justify spending for now. They also have a new little "handheld" cnc that's kind of neat, the Handibot. Idea with it seems to be you take the machine to the material, rather than the material to the machine. They've also said they want it to develop a whole library of special purpose apps, so it'll be interesting to see how it grows over time.

CNC Router Parts sells several kits that are quite popular. They tend to fall in between the DIY and Commercial units for price, but if you don't want to design things yourself, they're probably one of the better cost/performance options. Like ShopBot, I've never heard anything bad about them. They use their own design of linear guides and rack and pinion rather than leadscrews for the motion. I like leadscrews, but once you get about 4' or so, rack and pinion tends to be more practical for the DIY crowd, since you don't have to contend with screw whip (basically, screw vibrating/wobbling as it spins due to its flexibility over a larger distance - thicker screws are needed to combat it). They can be a fairly good source of some parts too even if you aren't following their plans/kits.

My first machine was built based on the Grunblau MDF cnc plans. Being MDF definitely limits its stiffness and speed a fair bit, though I recently realized I can up cut quality quite a bit by taking shallower cuts at faster speeds, but its pretty easily built - theres no metal machining to deal with at all, other than a few holes drilled in some aluminum angle iron for the linear rails. Knowing what I do now, I'd probably just spend the extra money and go straight to an all aluminum/8020 machine, but at the time, I wouldn't have done any CNC machine if I'd had to spend the extra grand or two metal would have cost for that size machine.

Really, I suspect pretty much anyone here is capable of building a cnc router; if you can build furniture and wire up light switches, you can probably handle building one. Aluminum can be cut with table saw (though you want a non-ferrous material blade) and a router. Its fairly easy to drill and tap (though a little gummy at times, so baby your taps). 8020 in particular makes it easy to frame things up, and they have a handy little calculator app on their site so you can get a feel for how the various profiles might flex. If you buy 8020 though, check ebay first - they sell all their seconds and whatnot (often as simple as a single screw hole or some tong marks) for pretty big discount - 30-60% off the retail typically. Like most things, the more of a machine you build yourself, the more machine you'll get for your money.

 

[Chemeleon]

Finished up the metal mounts today. Tried the first one yesterday with a 1/2" carbide endmill, that turned out to be too much for the wooden mounts, as they broke and jammed the router. Luckily, the surge protector strip tripped right away, so nothing was harmed. After that, I cut one by hand using a circle cutter on my drill press - I wouldn't really recommend it, especially with the included bit as they're brittle (*snap* *twang* *ping*), but with some homemade ones of soft steel it went ok. The homemade bits bent rather than snap, so was relatively safe, but still took 4 cutters and some sanding on the spindle sander to finish it.

After that one was mounted though, switched to a 1/4" bit, and let it cut out the second one, went without any hitches that time. So the machine is now officially done. There's still some stuff I want to make like a magnetic attaching dust shroud, need to fine tune the step count for the motors, etc, but everything that needs made for it to perform 100% is now made

I also did some brief testing at higher speeds, looks like I shouldn't have any trouble hitting my target goal of 300ipm with a 1/8" bit. Once I've played around with those speeds a bit more, I'll make a new video of it. Suffice to say, it flies