DISCLAIMER: This website is for entertainment purposes only. Not a source of professional advice.You can view my disclaimer HERE.

DIY CNC Router – Homemade Linear Bearings

After a couple weeks of testing and trying different methods, I finally got my linear bearing situation taken care of.  Thought I would write up the process I went through.

update (10-16-2012) – if you go the drill press and template route, make sure your press table and any vises are square.  i don’t believe mine was so the batch i didn’t do 100% on the cnc is off by 1mm, making them useless.  They are going straight into the foundry… when i build one…  I’ll try to post an update when I create the next batch.

They post the models for these small linear bearings from cncrouterparts.com.  I had to increase the height of their design to full inch so I could use common 1″x2″ aluminum bars.  I also changed some sizes of set screws and little things.

I’ve learned with CNC things can go wrong quick.  Best to prototype out of cheap stuff first.  Especially when you are creating a batch of parts.  So instead of jumping straight to using the aluminum stock, I decided to create some mock ones out of 1/2″ mdf to test everything first to test the changes I made.  So I cut some 2″x4.5″ pieces

Milling up the test piece stock

I just had to glue them up in pairs to get them into full 1″ depth.

Gluing the 2 half inch mdf pieces to create a 1 inch thick piece

Gluing the 2 half inch mdf pieces to create a 1 inch thick piece

Here’s the finally test product.

The final test stock piece for the linear bearing

The final test stock piece for the linear bearing

Next thing I had to do was to create my jig for the mill so I could do multi-sided milling.  I use 3/4″ mdf.  My machine doesn’t have any homing or limit switches.  To get things zero indexed I used a 30 degree v groove bit to find the top north west corner between milling.  I really need to get some limit switches and learn to use offsets.

Showing how the pieces fit into the jig

Showing how the pieces fit into the jig

Here’s a shot of the test piece’s top milling.

First test piece top milled

First test piece top milled

I flipped it over to do the other side and re-zeroed.

Test piece side milled

Test piece side milled

I flipped it over to do the next side.  Unfortunately I didn’t flip it the right way and ended up milling this piece the wrong way.

Test piece other side milled (the wrong way)

Test piece other side milled (the wrong way)

So that ruined that piece.  Had to create a whole new one.  And when I did I noticed that the nuts I had didn’t quite fit and broke off the stand offs created by the top milling.  Not critical but it made me dig deeper into my cam stuff being generated.  Turns out there wasn’t a hight enough tolerence in those tool paths and too much stock was being left for finishing operations.  Increased the tolerances and when it came to the milling aluminum everything went of great.  Another great reason to do prototyping in cheap stuff first.

Test linear bearing mocked up

Test linear bearing mocked up

Here is the underside. I’ll eventually replace these capped hex bolts with rounded over ones.

Underside of test bearing

Underside of test bearing

The whole system works with a pair of these ganged together.  Here I wanted to show how it fits on the rail.  The real rails will be much wider and made of 1/4″ cold rolled steel.

Demonstrating how the test bearing fits on the rail

Demonstrating how the test bearing fits on the rail

After I got things to reasonably fit I went ahead and created the first one out of aluminum.  It came out great!  All the test runs really paid off.  Here’s the top side.

First bearing's top before adding all the parts

First bearing’s top before adding all the parts

And here is the bottom

First bearing's bottom before adding all the parts

First bearing’s bottom before adding all the parts

I went ahead and created 3 more making it a total of 4.  But I noticed that this was really taking a beating on my carbide bits.  After a while I was having to stop the machine and clean out aluminum from the bits that had turned to clay.  Also, I didn’t have any bits with enough clearance for the deep hole drilling so that was causing things to jam.  After a while I decided that I could keep the pocketing and profiling stuff on the mill but all the holes probably needed to be done on a press.  So I created some jigs.  Here’s the top one

Top drilling jig

Top drilling jig

Here’s the right one.

Right drill jig

Right drill jig

And finally the left one.  This one might not have been necesarry if I had a drill press that had a greater than 2″ travel and I could do pilot holes all the way through.

Left drill jig

Left drill jig

Here they all are together to give you a better idea how they all fit together.  All the holes are drilled out to 1/4″ to fit my 1/4″ transfer punch.  This lets me mark up the stock and then take it to the press and use the appropriately sized bits to make the actual holes.  More about this in the video above.

All the drill jigs together

All the drill jigs together

Here is a demonstration of how the top one fits.  Notice I still drilled holes for the more oval shaped holes used for the adjustable post bearings.  Since its oddly shaped, it couldn’t really be pulled off on a press.

Demonstration of the top drilling jig

Demonstration of the top drilling jig

Using the press sped up the time in creating all 16 of my bearings but required more time hands on creating them.  If I could do it over again I would have only made one side jig for what I have labeled “R” for right.  I would have used a smaller bit to drill a pilot hole as much of the way as I could through the 2″ of aluminum, then used the same bit to finish the pilot hole with a hand drill (since my press has slightly less that 2″ of travel).  Then used the appropriately sized bits to finish off the holes.  I believe this is called concentric drilling.  If you notice in the picture below I had to actually increase the size of some of the clearance holes since they didn’t line up completely with the other side.  Kind of heart breaking putting in all this work and having to do that.

All the bearings created. The 4 original CNC'd and the 12 made with the mill and drill press. 16 total.

All the bearings created. The 4 original CNC’d and the 12 made with the mill and drill press. 16 total.

All in all I’m actually impressed with the cncrouterparts.com design.  Also with their pricing.  With all the work and hours I had to put into this, I think its pretty much worth it to buy them from cncrouterparts.com.  They come with abec7 bearings and all the nuts, bolts and split washers I believe.  But since I’m cash poor and have entirely too much time on my hands, it makes more sense for me to make them.  Plus, now I have the whole linear bearing situation taken care of for any other project I might need.

Beauty shot of the first bearing all put together

Beauty shot of the first bearing all put together

For the first 4 bearings I used screws and stuff at Lowes and got my bearings from Academy sports, which was the cheapest I could find abec 5 bearings in a physical store. But even then it was probably gonna break the bank.  So I ended up buying my parts in bulk from amazon for the rest.  Here’s roughly how it breaks down

$0.80 per linear inch of 1″x2″ aluminum bar x 4.5 = $3.60

$0.16 per set screw x 2 = $0.32

$0.31 per 50mm screw x 2 = $0.62

$0.22 per 30mm screw X 4 – $0.88

$0.06 per 8mm locking washer x 6 = $0.36

$1.00 per ABEC 5 radial skate bearing x 6 = $6.00

%?.?? cost in man hours machining, drilling and tapping (roughly 2 hours per bearing) = How ever much you value your time and your aptitude for machining.

So for about $11.78 I can turn out a bearing.  But considering how cheap they have theirs priced and that it comes with better bearings, I don’t know how they do it.  This also doesn’t factor in all the time it too and the fact that I had to buy a new set of end mills, a new set of drill bits and my first set of transfer punches.  So if you are not a hard core DIY or super cheap person like myself, I recommend you just buy them.

But for me this was more of a write of passage.  I figured if I couldn’t pull this off with a smaller cnc mill then I had no right building the bigger one.  I setup a code to follow that if I could build it, I had too.  Even if it didn’t cost less.

I did figure out a number of things doing this.  How to create jigs for work pieces, especially for multisided milling.  How to use transfer punches and screws (or what tranfer screws even really are).  How to use a sharp pointy bit to zero my work without any kind of homing.  How deep I can mill with certain bits.  Oh, and to only ever use 2 flute bits with aluminum.

So for me it was a well worth it exercise :)

5 comments

  1. Rodolpho says:

    you plan to sell this?

    • ThunderDork says:

      Certainly not! Its not really my design. You can get the models or buy them pre-made from cncrouterparts.com. At his price point, you can’t really beat it. Makes sense for me to make my own because I’m gonna need 16 in total. But if you need a limited run I’d go with him.

  2. Miroslav says:

    Great idea with creating sliders!
    I looked on the site cncrouterparts.com, and cheaper to make my sliders. If it is not a problem, would you send me a G-code for this? The problem is the postage and customs and taxation with us, and went four times more expensive. I apologize for the bad English!

    • ThunderDork says:

      Gcode is specific to individual CNC machines, tool configuration and interpreters (Mach vs emc2 etc). What i’d do (and did) is download their STEP or IGES files and setup your own gcode. Otherwise, best case, you break an endmill. Worst case, you damage your machine.

      Case in point, I learned my zenbot has a weak z axis and can’t really do routing with an endmill over 1/4 inch in aluminum with a hitachi router. So I had to dial back the plunges but could boost the cut feeds.

      On a personal note, it was a great learning experience to figure out how to replicate it. I figured if I couldn’t replicate it then I had no business building the machine I was designing.

      Hope this kinda helps!

  3. Miroslav says:

    Ok!
    Thanks for the description!
    I’m just a beginner in all this, and still looking for information about this. My desire is to make CNC least 60 x 100 cm, and these linear sliders seem to me the most practical and cheapest solution.

Leave a Reply

Your email address will not be published. Required fields are marked *

DISCLAIMER: This website is for entertainment purposes only. Not a source of professional advice.You can view my disclaimer HERE.