Building a cnc router

[lukedude43]

So my team has decided that building a cnc router would be a great investment of time/money/learning. I basically want any and all information that any of you care to share be it small tips, what did and didnt work for you, etc.

[sanddrag]

I'm re-doing the electronics on an old Shop-Bot Alpha. I'll be using a DynoMotion K-FLOP board and a Hitachi VFD with a 2.2KW chinese water cooled spindle. The K-FLOP comes very highly recommended, but I haven't tried it out yet. In theory I'll have it done this summer, but I said that last summer.

www.cnczone.com is a great resource for this sort of project.

[mrwright]

We built a 4 x 8 router from cncrouterparts.com. Their kits are really good if you are looking at simply assembling a quality base. They have options of turn key kits or you can build the electronics from scratch as well.

They also have excellent customer service which was really helpful when we needed help troubleshooting. They are also located in Washington.

What size are you looking to build? How much money are you looking to spend? Our router was around $9,000.

[ChuckDickerson]

You might also what to poke around here: http://www.kronosrobotics.com/krmx02

[lukedude43]

We're probably just looking at around a 2x4, we're just moving into a new space that's roughly 400 sq feet so space is a commodity

As far as budget goes we have about 5k on hand we can throw at this plus we can fundraise specifically towards this project if we need to

[Steven Smith]

I've done a bit of research, and ended up heading down the route of a KRMx02 in either the 4'x4' or 4'x8' version. We have some applications like building display cases for the school where not having to step the material through the router would be nice. Regardless, it's a nice and stiff machine capable of doing aluminum (shallow passes) with acceptable tolerances for FRC. Price tag is ~$4K-$7K for everything depending on size and sourcing of material.

An other option would be to just purchase something in the size you're looking for. I've heard good things about the Probotix and there are at least a few FRC teams running them (2468 is one I believe). Just weigh between the challenge of building one yourself, what an acceptable final product will be in terms of accuracy, and how much you will really save. Do you want the router to be a project or a tool for other projects?

Regardless, I think its a great step in to CNC with great value ($$/benefit to a team), and I think it will end up being a work-horse for us... which is why its up next on our upgrade path.

Edit: Adding link to older thread.

[DonRotolo]

Originally Posted by ChuckDickerson

You might also what to poke around here: http://www.kronosrobotics.com/krmx02

Originally Posted by Steven Smith

I've done a bit of research, and ended up heading down the route of a KRMx02 in either the 4'x4' or 4'x8' version.

Seconded (thirded?). I built their 50 x 27 machine and am very pleased with it. With FRC discounts on 80/20 and building many of the aluminum bits myself it came to just over $3200. (I would be happy to advise you on what to buy and what to make, depending on your fabrication abilities and capabilities. Also, don't waste much $ on the "upgrades", I can help you design your own).

Definitely go with the 50" X-carriage. Your table can be relatively small, and you can index and move the stock to get a larger piece through.

I recommend you proceed carefully for a 4 x 8 machine, it's hard to manage clamping stock and such; you'll find you use the front 2 feet 98% of the time. Also don't limit yourself to the suggested lengths; the Y axis can be absolutely any length you wish.

4 x 4 (50" x 50") seems like a good size.

[sanddrag]

To me a good size would be something capable of cutting 5 ft by 5 ft since that's the size of sheets of baltic birch plywood

[cgmv123]

Originally Posted by sanddrag

To me a good size would be something capable of cutting 5 ft by 5 ft since that's the size of sheets of baltic birch plywood

5 by 8 ft would let you do 4 by 8 ft sheets of conventional plywood as well, but that might be pushing it a bit.

[notmattlythgoe]

Originally Posted by DonRotolo

Seconded (thirded?). I built their 50 x 27 machine and am very pleased with it. With FRC discounts on 80/20 and building many of the aluminum bits myself it came to just over $3200. (I would be happy to advise you on what to buy and what to make, depending on your fabrication abilities and capabilities. Also, don't waste much $ on the "upgrades", I can help you design your own).

Definitely go with the 50" X-carriage. Your table can be relatively small, and you can index and move the stock to get a larger piece through.

I recommend you proceed carefully for a 4 x 8 machine, it's hard to manage clamping stock and such; you'll find you use the front 2 feet 98% of the time. Also don't limit yourself to the suggested lengths; the Y axis can be absolutely any length you wish.

4 x 4 (50" x 50") seems like a good size.

What kinds of materials have you been cutting with it? Our team is looking into giving our current mill CNC capabilities, but depending on the cost and space building another one might also be beneficial.

[Steven Smith]

Something like the Kronos can cut aluminum successfully, however it may require a little tuning. Plastics/wood are generally much more forgiving in terms of feed rates, and you should have enough power and stiffness to make short work of them.

Essentially though, a gantry mill analysis is no different than any machine when you're looking at what you can cut successfully. In order to cut something, you need a combination of sufficient rigidity in your machine and sufficient power to make the cut. If your machine isn't rigid, the bit will bend off the cutting surface (rub on it), or chatter in and out of it. If you lack the power, well, you'll find out that out quickly

Gantry mills are pretty awesome in that they can be reasonably cheap for such a large work surface. It comes at a compromise though. It's harder to make a large span as rigid as a small one. Also, the larger you make your gantry (the moving part with the router attached) the more power you need to accelerate and decelerate it. Routers are a great and reasonably cheap tool head. However, they natively spin at a very high RPM (20K range). When you do the calculations for how fast you need to feed material to constantly take a bit enough bite of material to not chatter, you're moving fairly quickly through material. This can generate a cutting load of 25-50lbs+ resisting the bit, which at the best case generates inaccuracy, and at the worst case, results in chatter/broken bits.

At the end of the day, you can always just choose to take a shallower cut, (0.03" to 0.05" passes aren't uncommon), which allows you to move through the material quickly enough to maintain chip load, but minimizing bit deflection. So many machines can "cut aluminum", but the more rigid ones will do so more accurately and more quickly.

[DonRotolo]

Originally Posted by notmattlythgoe

What kinds of materials have you been cutting with it? Our team is looking into giving our current mill CNC capabilities, but depending on the cost and space building another one might also be beneficial.

I cut wood, polycabonate and aluminum. I am using a Hitachi M12VC router and high-speed carbide bits from Onsrud.

Wood is trivial, just about any router bit will do, but upcut bits are best for slots and outlines of parts.

Believe it or not, but polycarb is the hardest to cut, primarily because I am using a 1/8" bit which cannot handle much deflection.

Contrast that with sheet aluminum (6061), with my 1/4" bit I can cut 1/4" thick material single-pass at 50 IPM...As long as the coolant mist and compressed air are on. Chip management is critical for Aluminum, or bits will break.

Repeatability is a few thousandths, but you need to take time and set up the machine carefully.

[sanddrag]

Originally Posted by DonRotolo

Contrast that with sheet aluminum (6061), with my 1/4" bit I can cut 1/4" thick material single-pass at 50 IPM...As long as the coolant mist and compressed air are on. Chip management is critical for Aluminum, or bits will break.

I take it this is slotting, full width, with an Onsrud single flute bit? Still, that seems like a rather hefty cut. FSWizard tells me to go something like 18,000 RPM at 34 inches per minute for 6061 aluminum, a 60 degree helix angle 1/4" single flute bit, in 1/4" 6061 aluminum, full depth, slotting. It says it will take 0.6 Horsepower for this cut. I wonder if my Porter cable would be able to do this cut.

[DonRotolo]

Yes. Specifically, I did the calculations for an Onsrud 63-620 (1/4" diameter) single-flute upcut bit for Aluminum, with a 1" stickout, running at 16,000 RPM and a straight-line cutting speed of 60 IPM. Under those conditions I can cut as much as 0.300 of material in a full slot. I know it sounds crazy, but it really does it. Less than 0.001" deflection. I didn't calculate the HP, but it is well under 2, maybe even less than 1.

But an Onsrud 63-760 (1/8") for polycarb with 1" stickout at 10,000 RPM and 64 IPM can only cut 0.020" deep slots. It is weird, but the thinner bit really deflects so much more compared to the 1/4" bit.

Both assume a fairly strong stream of air to blow chips out of the slot, and a vacuum system to whisk them away. For the aluminum, add mist coolant to the mix. Allowing the bit to cut chips twice is almost asking for it to shatter.

34 IPM doesn't give you the chipload you need, it is 0.0019 and Onsrud specifies 0.003-0.006. Reduce RPMs or run faster. Never any harm in taking a smaller depth of cut if .250 gets you nervous.

[CENTURION]

Originally Posted by DonRotolo

34 IPM doesn't give you the chipload you need, it is 0.0019 and Onsrud specifies 0.003-0.006. Reduce RPMs or run faster. Never any harm in taking a smaller depth of cut if .250 gets you nervous.

Quoted for truth. Get your chip load correct, then step up your DoC until you get uncomfortable (or it starts making bad noises).

I'm currently designing a router for BadgerBOTS. Still wrangling sponsors, but I've been working on the paper design/CAD since January.

Here's the latest render from a couple nights ago. Still very WIP.

Eventually I'll scan my whole engineering notebook on this design and upload it. If you guys want, I could scan what I have some time this weekend, but as I said, the design isn't exactly complete. There will also be a build log when it comes to that, but it may be a while.