BenchMaster CNC Build

Since nobody would give me a CNC machine, buying a turn-key machine was out of the budget and I needed one, I decided my best option was to build my own.  I'm a fan of classic American iron.  Most people see old equipment as pocket money when scrap prices are up.  I see it as recycling.  Why landfill something that can be repurposed and put back into useful service?

After a bit of research I narrowed it down to what I wanted:  A small but sturdy mill on a blue collar budget.  That ruled out the Chinese imports (crap) and newer USA made ($$$) stuff.  As a rare stroke of good luck this little Benchmaster mill turned up locally.  Being from the early 1950's she's pretty wore out but exactly what I wanted.  One day when time allows (ha ha) I'll teach myself to scrape and get the slop out of her ways.  The middle of the X axis is pretty deep compared to the ends. 

Most of these pictures and videos are quick convenient shots from my cell phone.  So, don't expect any Pulitzer worthy documentation shots here.

Scroll to the bottom for videos.

To the right is what I started with:  A well worn and rusted but perfectly usable piece of equipment on a wobbly stand plucked from the corner of an old man's garage.

The first thing I had to do to start the CNC conversion was to get rid of the ACME thread lead screws.  Even new ACME threads have a certain amount of backlash (slack) in them.  Any backlash is a fail for CNC, much less 70 years worth of wear. 

Look closely at how much sharper the threads are near the center vs the ends.  The table had about .060" of backlash.  They needed replacing period.

I didn't want to cut up the original lead screws.  I'm not too sure why now seeing that they are worn slam out.  The part where the handles mount had a good bit of runout from being slightly bent anyway and probably would have made belt adjustment next to impossible.

First order of business was to make these two little beauties.  They're close to being duplicates of the original ends except the od is slightly different to give the pulleys a tight fit and a bit longer so the original handles can still be mounted.  The rear is 3/8-24 threaded.  I put a corresponding thread on the ball screws.  A dab of red Loctite and they shouldn't ever come loose.

I went through the trouble of making them so the handles could be mounted.  I may remake these from 1018 steel later and forego the handle capability.  Unneeded.

One of my best shop investments was a set of quality YG-1 spiral flute taps.  You can hand hold the tap in this fixture threading 60% overlap threads in steel and go slam to the bottom of the hole in one shot without stopping.

Now to mount the ball nuts.  PITA.  These little mounts were time consuming.  Threaded from round stock on the lathe with a boring bar then milled, on this little Benchmaster no less, to final dimensions. 

Yes, the table had to be pulled apart, mounts measured, reassembled, whittle on them, disassemble the table again, test fit, reassemble, rinse and repeat until done.  Rough finished in this picture while test fitting.

After several test fits.....

Clear RTV was put on each end of the ball nut to (hopefully) keep out debris.  Oil the ball screw, apply RTV liberally, let dry 24 hours, turn the screw and hope it didn't stick.  So far so good.

The single tiny #8 set screw was later replaced with two 3/16" set screws on each flat after the ball nut worked loose.

The Y axis is already installed.  Unfortunately, these 5/8" ball nuts weren't a direct drop in.  Oh no.  The X axis one was too tall to clear the underside of the table.  No pictures, but I had to have a slot machined for clearance since I had no way to do it.

I didn't think to take pictures before installing, but I did include thrust bearings on each side between the lash nut and table.  There's a pair on each axis.  One inside and one outside.  Without them there was .005ish" of backlash.  Tightening the lash nuts any further made it too difficult to turn the pulleys.

I still have about .001" with them tightened to 10ish lb-ft.  I attribute that to the belts.  Either way, good enough for gubment work. 

There was nothing to attach the motor mount for the Z axis to.  I made this from 6061 aluminum round bar.  Bored out on the lathe to get a tight fit, machined the sides and drilled the bolt holes on this mill, sliced in half on the band saw, then faced on this mill to get them back even. 

Bolt holes were later drilled/tapped in the front for the motor mount to attach to.

I did not convert the Z axis ACME lead screw to a ball screw.  It has 40ish pound of Grade A American cast iron on top of it.  I figure if it has any backlash with that much weight holding down on it then I have bigger problems.

Now the fun part (why is there no universal sarcasm font?!?)

The wiring.  Kudos to my better half, Sommer, for the soldering job.  I can machine intricate parts to within .0005" tolerance press fits, but thanks to an accident, lack the fine motor control to hold delicate objects for operations like this.  These are the motor plugs.  I won't go into detail on the actual connections.  Every setup is different, and quite frankly, I don't care to think about it again.  I'm no electrimagician.

I'm using a Gecko G540 Motion Controller.  Plenty enough power for the three 600 oz NEMA 23 motors.  The power supply is a generic 48v 12.5 amp unit.  Nothing special but far more than needed.  I may add a little rotary 4th axis eventually.  It should power it just fine too.

I have not and don't plan on rigging up limit switches any time soon.  I'm using LinuxCNC for control software.  Set its limits up and home correctly and you'll never exceed the machine's boundaries and break anything... So far anyway.

I made this panel on my 3d printer to mount everything under my desk.  Crap picture but it came out looking and functioning like a mass production factory piece.

Completed assembly.  I didn't take pictures before using it some.  The motor mounting plates were cut from 1/4" scrap.  Nothing special.  They did come out looking like they belong on it.  Table travel is 5.25" on the Y, 12" on the X, and 6" on the Z.  I only use about 9" on the X.  If I need to go farther I have to loosen the gib due to the table wear, which puts slack in the middle of the travel and causes chatter.


I got tired of the belts.  They worked admirably but were not a permanent solution.  The set screws had to be tightened every few hours and would slip if pushed too fast or hard.  Mostly my fault for making the shafts from aluminum.  Oh well.

The motors have enough power to not need the reduction and the direct drive works better all around.  It's made from 3/16" scrap plate whacked out on a CNC plasma.

Yes, it's filthy.  She's being (ab)used as she models for these photos.

I printed this plate out for the X axis on my 3d printer.  Works like a charm.  Don't pretend like you're not diggin' the yeller color.  I planned for this to be the prototype and print another from black.  Didn't need any modifications and laziness has discouraged replacement.

Some intricate scribing done on her.  These are gage plates I made to QC parts by at work.  It did good with a diamond drag tip.

CNCing a timing pulley for another project.  It can go faster than this.  I'm still a bit of a wuss at pushing it.  I didn't record the finishing pass.

Not CNC.  Manually jogging the controls to fly cut this flange.  It takes .050" passes easily with this hss fly cutter I made and leaves a near mirror finish.  I did the bolt pattern on this by CNC. 

Note the 3d printed collet wrench I made.  Works better for me than the normal cheap stamped steel ER20 collet wrenches.  I made the steel wrench to fit on the collet chuck.  Swing around until it hits the mill and then use the plastic collet wrench to break it loose for one handed tool changes.

A simple bolt pattern.  I have it programmed to drill one round 1/8" deep, go to Z home, await input command, manually touch off z height, then redrill the same pattern 1/2" deep.  It works pretty good for doing manual tool changes since an ER collet system doesn't have a way to get really accurate and repeatable tool heights.

I didn't record the drilling because I was manning the kill switch in case something went wrong.  It took me much of a day to manually machine that part.  It wouldn't take my little electronic creation 3 seconds to turn it into scrap.  Fortunately, she did good and it fit perfect.

CNCing a drive socket to go on my Bridgeport mill

The completed part being used. Spiffy.

  • No Comments
Powered by SmugMug Owner Log In