[cffellows]

After months of planning and designing, I've begun construction of a small, Desktop CNC router. It will be a fixed gantry for the X-Axis and a moving table for the Y-Axis. Work envelope will be about 12" x 9" x 3" for the X, Y, & Z respectively. The X & Y axes will be 5/8" diameter hardened steel shafting with the moving parts attached with bronze bushings. The X axis rods are 16" long and the Y axis rods are 18" long. The main frame construction will probably be MDF although that may change. The inspiration for my design comes from Microcarve's MV2 and BBox routers with a lot of my own changes. Here is a picture of the assembled Y Axis:



The end plates 1/2" thick aluminum, 8" long. Spacing between centers of the shafts is 7". The back piece of aluminim is 1.25" high and the front is 1" high. The back will have some of height cut away to make room for the Y axis stepper which will be inside the frame (maybe). Other parts of the router will be made largely from items and materials already on hand. The X & Y axes will use 16mm ball screws, pretty beefy, I know, but I already have them. The spindle will be a 400 watt, 12,000 RPM DC model with integral ER11 collet chuck. Stepper motors will be older, NEMA 23 which I already have on hand, although I may have to upgrade these to something a little newer and stronger. The Z Axis will be a dovetail setup which I made from aluminum. I have the good fortune of having a small machine shop so am able to custom make a lot of the parts for this router. Don't know how fast this build will go, since I'm still designing as I go, so stay tuned.

Chuck

[datac]

Well, good luck with your little project. I've become quite fond of smaller builds..... After you already have the larger footprints machines, it does become interesting to see how small you can make a competent machine. I feel a little windy today, so if you have nothing to do, I'll critique, not just your design, but what I see too often in general.

For me, I sometimes wonder why so many builders take the unsupported rail route ? I never build anything with unsupported rail... Everything with cast aluminum tooling plate and supported rail. Frankly, I still yearn for MORE rigidity for each machine I build. Yet, understand that I realize that first and foremost, any machine needs to be designed and built surround the actual NEEDS of the builder or requirements of the purpose. If one is building a non-contact machine like plasma or 3d printing, it is a whole different requirement over contact machining.

The second aspect to my wonders is if builders realize just how much more difficult it can be to build a machine, often without the right tools, or at the expense of having someone else make the parts, when unsupported rails are used. At the risk of being critical of your initial design, What exactly serves as your designs "backbone" ?..... the thing that keeps the machine straight and true when it is lifted off of the table top you build it on ? If the only major thing that keeps a machine straight is the set screws that tighten the linear shafts to the end plates, that's simply not enough for any contact machine. When builders realize this inadequacy, they often apply in their design, additional "support" that is really just a patch to fix the initial shortcomings. Too often, this patch involves aluminum extrusion, and frankly, I've never seen any straight enough to make a reliable CNC machine.

Now obviously, the smaller platform does make your current design more tolerable in that it all could be dropped onto a single tooling plate and end up relatively straight and true... but you still have the unavoidable "give" in unsupported rail. If the spindle you want to use ends up being not much more than "Dremel" sized with light cuts on wax.... then your design is fine. Putting a spindle capable of throwing an actual chip rather than dust ?, ..... well, I've never seen unsupported rail end up worth a darn.

I guess the point I wanted to share is that for small footprint machines, BUY TOOLING PLATE ! Small sizes are plentiful as cutoffs, and at really great prices. I buy mine out of a place in Milwaukee (howard precision metals). They have everything on computer. Email them with the sizes you need and they will send back what they have close to it. Sometimes I will combine smaller parts into a larger sheet. Remember that they really love to move these small pieces because they can get more than scrap for them, but nothing near the price you would pay for a new cut out of a new sheet.

With tooling plate, you can fasten supported rail and get much, much more rigid operation. Too, a person can accomplish the rail attachment without ANY difficult "machining". You can clamp one rail on, tag all the holes with a hand drill, drill and tap the holes. If you build your sliding axis next, just having it rigidly tied to the one side allows you to properly position the other side so you can drill and tap tho opposite linear rail.

Add to this, is that all the really short supported linear rail components are really affordable as surplus or takeouts. THK RSR 9's, 12' make for an awesome, very rigid slide system, and again, few are building these small machines, and consequently, can be had for little money.

Take some time to reflect on some of the advantages of building with cast tooling plate and supported rail. When you build a "stackup" of components, even without machining parts, your machine can be every bit as accurate as the specs on the tooling plate (which is often very tight). Because you do not have to machine numerous bores to hold unsupported rail, by nature, things are just easily more accurate.

Because COST is always an important part of hobby cnc, think too about how few ENDS or EDGES there are that do need to be machined with perfection in comparison, because many profiles just have borders or non-functioning edges. Non-functioning edges can be band sawed and sanded with ease.

I am attaching a machine design I built some years ago. It was a personal drawing challenge of mine, to see just how FEW aluminum components I could build a machine with. I believe it was somewhere around 14, and the majority of the machines dimensional accuracy was dictated by the thickness of the tooling plate. If you stop to look long enough, you can pick out the parts where EDGES are critical to maintain stack up heights (pretty much any piece standing vertical on edge). This machine used RSR 9 and 12 Rails, and proved to be extremely rigid. Fast too, capable of 400 ipm rapids in a 12" x 15" work area (but not applied for safety reasons).

Note too, that I don't think the plate for this machine costed much more than $50. That is with FREE delivery as I just so happen to live on a route they take every Tuesday.

Finally, I also believe in "less is more" when it comes to control of a small machine. My all time go to is Flashcut. It is not "less" in function by any means, but as far as the control mechanical components for a machine this size, it all fits in a box thats only some 3" tall, 10" wide and 8" deep. Plug it in, plug your laptop in via USB and your living the dream. I bring this up because lately, I see people building "Control boxes" and "Control Panels" that end up being 4 times the size of their machine ! For what ? I have no clue. It doesn't need to be that difficult. Those who say they need a physical button for any and every function are just "newbies", or, just have not ever had the chance to use a well laid out control.

I am looking forward to using one of the many GRBL variants for a small machine down the road. Seems motion algorithms have fallen into place.. just waiting for more of the necessary functions to be added and one of the many GUI's to accel to stardom.

Well, there is my 2 cents. Keep us up to date on your build.. we all like PICTURES, no matter what your building !

[cffellows]

Thanks, Chris, for your comments. I appreciate anyone who takes the time to share their knowledge and experience. I do have a small machine shop with an 11" Logan lathe and a 7" Harbor Freight mini lathe. I also have an Enco Mill/Drill which I converted to CNC several years ago. In addition I have a floor model drill press, 2" x 48" belt grinder, Sears 10" table saw, and 25 years accumulation of various tooling. I've built everything from model gas engines to astronomical telescopes, mostly from scratch. However, other than converting my mill to CNC, I have no experience with building a CNC machine from scratch. I appreciate the fact that good results are most likely attainable from experience rather than any book knowledge or intuition. So, here are the reasons I had decided to use the design I have come up with.

First, I decided I wanted a fixed gantry. Intuitively, it seems like good rigidity would be harder and more costly to attain with a moving gantry design. I understand and am willing to live with the larger machine footprint needed to attain the same work size envelope on a fixed gantry design.

Second, abandoned the idea of using ball bearing races, including v-groove bearings, like the Momus and Solsylva. It seems like the small area of contact between the bearings and rails would hurt rigidity and accelerate wear, particularly if the rails are aluminum like some folks have used. I'm also kind of put off by the large number of parts needed for ball bearing race carriages. Seems like accuracy might suffer and adjustment would be needed on a regular basis.

Linear bearing systems are certainly attractive on many levels, most of which you've pointed out. Cost is certainly an issue for me, but not a show stopper. However, add to that the number of different sizes and manufacturers available make selection a tedious and time consuming process. I'm also concerned about wood dust and/or metal shavings getting into the ball races, necessitating frequent and thorough cleaning, not my strong points.

So, so far, I've decided bronze bushings on hardened steel shafts might be the best compromise for me. The bronze bushings give me a relative larger contact service area between carriage and rails, wood dust and metal shavings shouldn't be much of a problem, and the design is cheap and simple. On the downside, I suspect that chatter and other rigidity related issues might surface if I try to take too heavy of cuts, particularly in aluminum or brass. In retrospect I should probably have opted for 3/4" shafting instead of 5/8". But this brings us to my particular application for a CNC router. I'm mostly interested in building one, one that works well and isn't too expensive. I really have no particular application for it other than to play with it and make the odd part here and there. Aluminum and/or brass would be desireable, but not really necessary since I do have a CNC mill I can use.

Since my first post, I have decided to buy 270 oz in, 3 amp NEMA 23 steppers and SFU1204 Ball Screws. I decided that the 16mm ball screws I already had are probably way more than I needed and I would prefer the smaller mass of the 12mm screws. The NEMA 23 steppers I had on hand are the older, round body type and probably have less than 150 oz in of torque.

Let me add, that I really do like the design in the drawings you supplied. A very nice, simple looking machine. If I might ask, what size and model number are the linear rails you used? Are they double row bearings or quadruple row?

I do have aluminum tool plate on hand and could use that. What are your thoughts on using 3/4" MDF instead?

Chuck

[datac]

Sounds like your plenty well equipped to make enough of the matching parts you require accurately.

Regards moving vse fixed gantry... thats an age old argument you'll never settle amongst all... sort of like stepper verses servo ! I tend to prefer the moving gantry for routers and engravers. Obviously a steel cutting mill is different, but this is a completely different machine with a different intent. I've both used and built both. Over the years, I've used a lot of vacuum hold down, and it kind of stinks to have to have the vacuum lines rock back and forth with the material. I also do not like the almost double sized footprint. It usually means more material than necessary too.

V-Groove methods burn up a lot of dimension on a smaller machine. There are plenty of examples of some over sized, over imagined methods. The problem I have is there is no real savings to be had as they are not exactly cheap either. Now, for a larger footprint machine, they certainly have their place and suddenly do become the affordable option. While the V-rails and wheels are a lesser cost option for large footprint, the REAL problem is getting STRAIGHT material long enough to mount them to.

Linear rail applications do have to have some thought about contamination, certainly the THK (Flat) style as they are much harder to seal up. When I use them, I try my best to keep the rails in a less than direct access location, not much unlike UNDER the table in that small machines design.

For a medium sized router that going to see a lot of fine dust, I believe the best option is supported round linear rail as these do have pretty reliable seals, fully engaging the round rails, removing dust along the way.

The trick to shopping used/surplus/takeouts is to buy rails with only one bearing slide.... All the incoming CNC newbies do not want them ! They always want two bearings. So, the single bearing units often sit stagnant. I've contacted sellers who have what I want sitting there not selling and tell them to change the auction to make an offer... And I usually get the rail for a good deal because the seller must also not be smart enough to add another bearing.

Then, what you do, is search for bearings... on really short, short rails. Hardly ANYONE wants that for their CNC project, and as such, you eventually get a deal on that second bearing ! Granted, this takes a little patience, but I've been successful with it.

Your reason for building is understandable. I too have made my share of things.... "just because". Its a great way to prove things out or learn.

270 oz motors seem plenty enough for that machine, the small machine I showed you uses 240oz and runs like a rocket. Run the highest voltage you can to get the best performance from them you can get.

Ballscrews..... Well, it seems everyone wants ballscrews. For a wood cutting machine, they too can fill up with dirt and dust. They have nothing for a wiper at all. When it comes to dirty applications, I'm in the rolled lead screw camp.... lead screws like KERK makes. They will work in a filthy environment for practically ever. Some immediately question the accuracy, but frankly, all the ball screws people are purchasing affordably are imported.... and we all know how wacky some of that stuff can get, not to mention that out of the box, they have a lot more backlash than a Kerk system.

For travel distanced up to around 30"... perhaps 36", I will run 1/2" diameter screws, and something with a travel speed of 1/2" per turn on a router/engraver. Your 16mm screws should turn just fine, but correct, you could have got by with a smaller diameter.

The THK rails on that little machine are RSR 12's for the X and Y... RSR 9's for the Z axis. You can find the drawings for them easily with a google search. The one quick thing to remember is if they say rsr "12 W", it will mean the rails are nearly double wide than non-W rails. I've used them too as they are every bit as good..

MDF ? While many people have built their dreams with it... its not for me. I like to build a machine that I enjoy fully the first time.... MDF is not exactly stable. One spill of your coffee, any little crack it gets into will swell up like a balloon.... Tooling plate for a SMALL machine is just too cheap to bother !

[cffellows]

So, are the KERK lead screws the same thread form as an ACME thread? I see that KERK offers a lot of different pitches whereas ACME comes in a few standard pitches. Where does one buy the KERK lead screws and how do they compare in price to standard ACME. I looked on the KERK website but couldn't find any pricing. Darn, so much information to process...

[datac]

>>> So, are the KERK lead screws the same thread form as an ACME thread?

Nope.. far more rounded, and I am sure a whole lot more accurate.

If I am working on a machine that will require a share of documentation, I will get what I need directly from Kerk. They come packed similar to a quality ball screw does, with all sorts of linear test data and results.

For a small desktop/hobby leaning machine, I tend to use the Kerk screws that Mcmaster sells. They have changed the name of these in their catalog over the years, but they, as I have been told, are Kerk screws. For reference, look up item number 6350k23 at Mcmaster.com.

If you order 6 feet from mcmaster though, it will arrive flopping around loose in a cardboard tube, not properly packed. It may or may not also require some gentle straightening because it is clear that they dont regard a precision screw to mean anything to them. But they do work well.

NOTE that there was a time when these were stunningly affordable compared to ANY available ballscrew, but then we got China. I wouldn't doubt that there are an awful lot of China made "ballscrews" that couldn't hold a candle to a genuine Kerk screw. Just the word "ballscrew" doesn't really mean anything special to me lately.

Unfortunately, Mcmaster only sells the lightest duty nuts available when it comes to an antibacklash nut. They are fine for a small machine like the one in this thread... you can see it in the photos. For larger machines though, I use Kerks VHD style nut. Crazy reliable, insane longevity, always zero backlash because it is two nuts spring together/apart with continuously maintained tension by design. It is also adjustable for tension.

Too bad the economy went haywire to shoot the costs of fine american made components to the point where they are almost out of reach. Still, $200 worth of screw usually is plenty for a desktop machine.

[louieatienza]

I believe McMaster differentiates standard ACME from precision ACME. And yes. the Kerk stuff has a different thread form. You can check out Roton as well. I've bought AB nuts from DumpsterCNC and still have the same ones on my 6 year old machine.

I don't think you'll see anywhere near the 25lbs force these lighter AB nuts are capable of, and considering the next tier of AB nuts cost 2-3x the DumpsterCNC nuts I think you'll be fine with them. The Chinese ballscrews on eBay are rated C7 accuracy, which is about .002"/ft, and are called anti-backlash, which does NOT mean zero backlash. It does mean an axial clearance of .06mm or less. I've heard of the nuts being re-balled with success. However this will cause variable torque needed to drive the screw due to variations in pitch, mainly due to the rolling process.

However, ballscrews are at least 90% efficient, whereas an ACME screw about 30-60% depending on pitch. So a ballscrew requires less torque or smaller motor to drive an equivalent pitch ACME. But, that extra resistance can actually help a stepper as far as damping resonance, and it's less susceptible to back-driving. I like using ACME as Chris mentioned for the more worry-free and maintenance -free operation.

[cffellows]

Ahh, I hate it when well-informed and knowledgeable folks cause me to second guess my decided course of action... :^). Wish I had not purchased the 12mm ballscrews. But, since I have more time than money and won't be making parts for the defense department, I'm going to proceed with my current design. That is, bronze bushings on hardened steel shafting and 12mm ball screws. I'm still considering using aluminum tool plate for now. Again, I really do thank you folks for sharing your knowledge and taking the time to provide a lot of good information. I know you probably repeat this information a million times since a lot of people, myself included sometimes, would rather just ask questions than do a little research. So, thanks for your patience, it really does help.

Chuck

[datac]

Ahh, I hate it when well-informed and knowledgeable folks cause me to second guess my decided course of action... :^). Wish I had not purchased the 12mm ballscrews. But, since I have more time than money and won't be making parts for the defense department, I'm going to proceed with my current design...........

By all means use what you have.... You can always swap the screws out, but I would imagine you will just get hooked on "building" and "designing" anyhow..... So another machine is probably in your future !

I did not want to rain on anyones parade,.. I just saw this thread un-responded to and thought I would pass along my experiences. I do shudder sometimes when I see the MDF rollerskate world, but people have to build what they want to build.

My main point I guess is simply that building a small machine today can be not only really affordable because most people are focusing on larger builds, but also very, very rigid and capable. Where some hardware costs have gone up, like that of screws, nuts and bearings, small motors, small quantities of raw stock and certainly motor drivers and CONTROL aspects have dropped significantly since I got into this stuff. I wish I would stop running out of money (or just limit my interests better), so I could build continuously. Sometimes I just draw machines to draw machines.

It's all great fun.... plenty to learn and at least these days, PLENTY of people and forums to lean on. Back in the early 1990's, the DIY aspect of home brew CNC machines was a pretty well kept secret !

Keep us all up to date on your progress.

[louieatienza]

I think you'll be fine with your parts. I would consider something other than MDF. Since you already have machinery to mill metal, it doesn't make sense to use MDF. It's used because it's accessible, cheap, and easy to cut. It takes a bit of imagination to use it right, and most who do substitute parts out for plastic, phenolic, or metal. Some will even use it to build temp parts to get the machine running, then use the machine to make better parts for itself.

[datac]

I don't think you'll see anywhere near the 25lbs force these lighter AB nuts are capable of...........

Boy, a long time ago, I ran thru the math calculations for side loading of sharp tooling running at the ideal rpm and travel. I was looking at it from the wood and plastics aspect and indeed, you would have to have a really dull bit making an absolute mess of things by the time you were pushing 25 lbs.

On a small Router/Engraver machine, you dont usually end up with motors and gearing capable of 25 lbs.... unless you built something that has crazy fine step positioning and practically no travel speeds to brag about.

[cffellows]

I am using a dovetail slide for my Z axis. I made this out of two pieces of 3/4" thick aluminum with the dovetail overlap being about 5/16". Overall dimensions are 7" x 2.5" for the bigger piece and 3.75" x 2.5" for the smaller piece. I've included gibs to adjust for wear and overall travel maxes out at at about 6" although I will only use about 3.5" of travel when installed.

Attachment 280692

Attachment 280694

[victorofga]

aluminum itself will mar each other..
at least some Teflon strips need on the surfaces..

[datac]

........ I've bought AB nuts from DumpsterCNC and still have the same ones on my 6 year old machine........

Nice. It's been a while since I've visited that site. Those nuts do look to be very nice, especially for the cost.

I think I will be trying his ACME 1/2-8, 4 Start nuts on a future machine. I bet it works out well !

[louieatienza]

Nice. It's been a while since I've visited that site. Those nuts do look to be very nice, especially for the cost.

I think I will be trying his ACME 1/2-8, 4 Start nuts on a future machine. I bet it works out well !

I also used 1/2"-8, 2 start nuts on a smaller machine (which I sold) and was able to achieve 400ipm rapids and 50in/s/s accel as my normal setup, and tested to 600ipm rapids and 60in/s/s accel. This was using 180in-oz Kollmorgen CTP steppers, 48V Meanwell PSU, Leadshine DM542 drives, and EdingCNC controller.

[louieatienza]

BTW Ahren at CNCRP has a really nice aluminum housing/mount that incorporates DumpsterCNC nuts, worth a look for a build.

[DSpeck]

If you are still interested in doing a fixed- gantry machine, check out microcarve's various machines. He uses MDF, unsupported round rail and the bronze bushings, as you do, and his machines are capable of very fine work. Do be sure to check out all his threads, as he gives a lot of his though/design ideas as well as construction techniques. He has a forum at microcarve.com/forum, which also has very useful information.

Don't be so sure unsupported rails won't work at this size of machine. Not for large machines, obviously, but in these smaller ones, can work fine, apparently.

BTW, I would not do an aluminum dovetail without at least some anodizing, as the aluminum galls (sticks to itself). You might be better to use the round rails on the Z-axis as well as the other ones.

[cffellows]

I actually got most of my inspiration from Microcarve's MV2, then made some of my changes, including a modest increase in size. He's removed a lot of his pictures and descriptions from the his website since he is now going commercial with his kits. He's using laminated bamboo for the frame, which is probably pretty strong and stable, but it sure looks weird. First thing I'd do is slap a coat of paint on it...

I'm not too concerned about the aluminum on aluminum dovetails. I think with some lubrication and relatively slow speeds, galling won't be a problem. I've used aluminum dovetail slides on metal working machines before and it worked fine.


I think I'm stretching it with the 5/8" shafting I used. In retrospect, 3/4" would have been a better choice for an 18" rod. This might cause rigidity to suffer, resulting in chatter if I try to make too heavy cuts in hard material.


Chuck

[rocketflier]

Thread: Making Acetal leadscrew nuts the easy way

https://www.youtube.com/watch?v=kO8gejSGYuo

https://www.youtube.com/watch?v=Cm450aXeaMc

https://www.youtube.com/watch?v=xN7QgBMuEDg

https://www.youtube.com/watch?v=M0D3_hRUvTo

[cffellows]

Thanks for the videos. I had thought long and hard about the scaled-down BBox. It's straight lines and assembled side panels use less material and make it easier to keep everything square. But, at the end of the day, silly as it sounds, I just didn't like the looks of it. I had decided to build something more closely resembling the MV2, yeah, mostly because I like the looks of it better. One of the things I didn't like about the Microcarve designs are the sockets used to secure the rods to the frame. Those are at least 3/4" long and add 1 1/2" to overall width and length of the frame for a given maximum cutting size.



Funny that you included a video of the Mantis. I actually started building a scaled up version of the Mantis late last year. Mine had a foot print of 12" x 12" and 1/2" shafting compared to 10" x 10" with 3/8" shafting on the original mantis. I accumulated a number of parts, including the steppers, and got the frame built and the rods installed. Then I decided I might as make something bigger and more practical, so I sold the partially-built Mantis with all the parts and actually recovered my investment. Much as I hate to admit it, I had purchased the rods and bushings for a larger machine before I ever started on the Mantis.

UPS is scheduled to delivery my steppers today so I'm pumped about that. The ball-screws probably won't come in for another couple of weeks. The amount of time it takes to get things from China drives me to distraction. Guess that's the price I pay for "Keeping My Options Open" until the last minute. I hate to plan too far ahead for fear I'll lose interest in the project or change designs before the parts get in. As you might guess, I spend a lot of time being indecisive.

All that aside, I'm resolved to finish this project as close to the original design as I can and using the parts I have ordered or have on hand. I've attached a 2 dimensional drawing showing what I have in mind. All the internal parts and dimensions are pretty well fixed. Only the frame design & shape remain to be finallized.



Chuck