[webgeek]

Hi all; I've got a good sized CNC mill that I love but I find it's a bit of a pain to use for small and thin stuff. It's low-ish RPM range is really limiting on what I can get away with when it comes to things like plastic, small bits in aluminum, garolite or carbon fiber. To that end, it seems I need a router. I want the router to replace some of what I do on my big mill though so I'm trying to come up with a compromise solution. Here are the requirements:

• Working envelope of 12"x18" (the hole grid is this size in the pictures)
• Ability to cut at lowish all the way to high RPMs
• Be reliable
• Be accurate
• Support coolant
• Be usable inside
• Cut aluminum reasonably - I know I won't be hogging, but I'd like some good speeds and shallow depths if possible. Not trying to push it here, but I'd like a reasonable surface finish.

Now I've got a fair bit of CNC experience, so I'm reasonably comfortable in putting together the right steppers, drivers, etc. for this project. I've never built a gantry-type design and I've come up with something a little non-traditional so I'd like to know if this is a workable solution. The specs for the design thus far:

• The base of the entire machine will be 1/2" MIC6.
• X and Y linear rods are 16mm from VXB. Not decided on the Z, but probably 12mm. 4 Linear Bearing 16mm CNC Router Bushing Linear Motion
• Spindle is undecided - I'm thinking the trim router from Rigid combined with the Think and Tinker precision collects and a Super PID to get the low RPM torque. Would love to know people's thoughts on this set up.
• Keiling steppers and drivers - I've had great luck with his stuff and plan to keep on using it.
• Roton 1/2-10 ACME screws with their anti-backlash nuts. These screws will be machined to have a flat for bearings, threads for a lock nut, and a flat again for the coupling to the steppers. I'm very nervous about the 25 lbs rating on these as I've never used anything but ballscrews in the past. Any thoughts here?
• Screws will be supported with two skate bearings on one side and clamped in place with the locknut. A single skate bearing will be on the other side. Not sure if this is good enough for a small machine or not. Would love comments on this too!
• The rods are held in place with shaft collars - not sure I like this in hindsight, but I'd like to avoid the slitting trick everywhere if possible.

Now for pictures...
The first shot is the obligitory render. I ran out of time tonight so it's missing lots of nuts and bolts but the basics are there. The last shot is the same, but just shaded for visibility. The z-axis is still in the air. The plan is to use the Mic6 as the bed combined with dowel pins to help secure things and locate them properly. I show how this works with the next three pictures with the gantry detail.

The bearing blocks have pins in them along with 1/4-20 tapped holes. The gantry base plate gets tied with 8 dowel pins and 6 bolts. In turn, the pins are shared for the stepper mount and the bearing mounts on both sides for the X axis.

I'd love any advice on any of this as it's all kinda different from what I've done in the past. Thanks much!

-Mike

[DonFrambach]

Sorry to be a curmudgeon, but I don't think that unsupported rails will work very well -- especially if you want to cut aluminum.

[webgeek]

Sorry to be a curmudgeon, but I don't think unsupported rails will work well -- especially if you want to cut aluminum.

Do you have more details on this? 16mm shafts tripled up like that (for the Y axis) have a deflection of .00093 with a 15 lb gantry over that length. I'm certainly not disagreeing with you, but I'd like to understand more. Obviously cutting force isn't included in that equation but that's going to be some weird combination of directions so I don't know how to ballpark it. Thanks!

-Mike

[DonFrambach]

Do you have more details on this? 16mm shafts tripled up like that (for the Y axis) have a deflection of .00093 with a 15 lb gantry over that length. I'm certainly not disagreeing with you, but I'd like to understand more. Obviously cutting force isn't included in that equation but that's going to be some weird combination of directions so I don't know how to ballpark it. Thanks!

-Mike

Sorry, I have no personal experience but I've read reports by others that unsupported rails vibrate excessively so that there would be extreme chatter cutting something hard like aluminum. I did make a slightly smaller router for which I used 20mm supported rails for one axis. Using this router, I have successfully cut aluminum and brass. For yucks, here's a link to my build thread: http://www.cnczone.com/forums/diy-cn...ry_router.html

[dbsharp]

You have a few forces to consider, static weight is the most obvious one, but must also consider the force from the cutter, the force from accelerating the mass of the gantry and motor, and also resonance dampening is also something to consider. I would recommend more space between each bearing on a shaft. Supported shafts will be more rigid and worth the extra effort. It looks like your planning on driving the gantry on the far side, you would be better off driving it from the middle or both sides.

[louieatienza]

If you want to cut aluminum, I would tend to agree with the others that at this size it might not be a good idea to used unsupported shafts, regardless of what the deflection specs are. The rails in addition to deflecting from the gantry weight, has to resist racking forces when cutting on the opposite side of the drive side, as well as resist the forces of the spindle/cutter... As to my thoughts on the specs:

1 - Using mic6 is great, but is somewhat negated by the use of smallish unsupported linear rails. Unless you plan to cut extremely slow, which would be impossible since the minimum speed with the SPID is 5000rpm, which for me would be at least 30ipm with a 2-flute bit. I think you'd find even at that speed you need a very ridgid setup to miimize chatter.

2 - Related to 1. Minitech however makes a mill that uses unsupported shaft; however the travels are way smaller and the rails are larger, and it's intended for micromachining...

3 - I personally use a SPID with success, but you'll find that generally aluminum cuts best at higher feeds, which would be difficult to do with unsupported shafts.

4 - Keling is a great company, and many here use them

5 - 1/2"-10 might be fine, but keep in mind the speeds you realistically need to be at to cut; even with a SuperPID, a 1hp router might not have enough "guts" to cut aluminum at 5000rpm, and you'll find you might get better results at about 10000rpm, at 60ipm. Asking your leadnuts to survive 600ipm continuous is asking a lot... That said I (as well as others) use Delrin nuts and machine aluminum fine. It's a moot point to worry about the 25lb rating of the leadnuts if your rails flex with only 15lbs...

6 - There's a mill on the metalworking forum that uses skate bearings to support ground ballscrews, on a machine made of steel, and it works just fine. The man, however is a toolmaker and his work is impeccable...

7 - The shaft collars are slitted, what's the difference? Money saved that can be used to buy supported shafts!

[Thinwater]

I used skate bearings press fit into the aluminum frame and they seem to be good. This is the build thread http://www.cnczone.com/forums/diy-cn..._24_plans.html

I have been engraving aluminum with it and it works very well.

Supported rails are the way to go. No flex in the structure that supports everything else. A gantry can flex enough to mess up aluminum milling, add this on top of a flexing base and you have problems.

I used thrust bearings on each screw end to limit any end play and keep the thrust off of the skate bearings. The motors are connected with lovejoy couplers. I use .003 backlash comp with the couplers and acme thread screws. Small circles engrave true, no steps anywhere, the ends join back perfectly.

My machine is ugly and made from 50% scrap, but it is smooth, very well aligned and uses hwin rails on all axis.

[webgeek]

Thanks for all the great feedback everyone. Back to the drawing board for me. I was hoping to get away with the shafts are they are dirt cheap for the precision they offer. Even small rails are vastly more expensive. I'll save money elsewhere!

5 - 1/2"-10 might be fine, but keep in mind the speeds you realistically need to be at to cut; even with a SuperPID, a 1hp router might not have enough "guts" to cut aluminum at 5000rpm, and you'll find you might get better results at about 10000rpm, at 60ipm. Asking your leadnuts to survive 600ipm continuous is asking a lot... That said I (as well as others) use Delrin nuts and machine aluminum fine. It's a moot point to worry about the 25lb rating of the leadnuts if your rails flex with only 15lbs...

I was pretty soft on the 1/2-10, looks like I'll switch to the Hi-Lead's from Roton and up the size of the motors slightly. That increases the travel speed substantially. It's also far more efficient as well.

7 - The shaft collars are slitted, what's the difference? Money saved that can be used to buy supported shafts!

I wanted to avoid doing the slitting myself, that was it.

Alright, back to the drawing board. Thanks much for all the feedback everyone!

-Mike

[vnlvet]

If I were you, I would mount the bearings that are on the opposite side of the double rail with lead/ball screw, much wider than it already is or put a second lead/ball screw on that side to keep the main beam parallel. You will produce a moment load on those round linear bearings in your design that will either bind them up or break them over time when the cutting head is extended as far from y-axis servo as possible. It would increase the leverage against the bearings the further it runs from the lead/ball screw. Making the pad wider would resist the possible moment.

Consider a design where the points of contact of the bearings on a slide are shaped like an equilateral triangle, or a square. The closer a bearing configuration resembles a long skinny rectangle, the more prone to premature bearing failure can occur. The more it resembles a square the better. Try to follow a rule of thumb of using a minimum 1/3rd of the column length as the width of point of contact of opposing bearings.

V