Hello Guest it is March 19, 2024, 10:02:01 AM

Author Topic: CNC CONVERSION - 6" ATLAS LATHE  (Read 71455 times)

0 Members and 1 Guest are viewing this topic.

Offline RICH

*
  • *
  •  7,427 7,427
    • View Profile
Re: CNC CONVERSION - 6" ATLAS LATHE
« Reply #20 on: September 04, 2009, 08:32:09 AM »
Russ,
 A friend may have "Telescoping Spring Steel Covers" which will work great with the lathe.
I found them in McMaster -Carr and they are pricy , but, mine will be free.
I will post here when the time comes and also may try to make my own that are rmovable for kicks.
RICH

Offline RICH

*
  • *
  •  7,427 7,427
    • View Profile
Re: CNC CONVERSION - 6" ATLAS LATHE
« Reply #21 on: September 06, 2009, 12:38:58 PM »
X AXIS BEARINGS AND BALL SCREW NUT

Two bearings will be provided for the X axis taking axial thrust loads and eliminating play.
Figure 27 shows one side mounted after a drawing layout was made. The ball screw and nut was centered and made parallel to the cross slide as shown in figure 28. Dimensions are taken and used to design the ball screw holding nut. The internal threads were done manually on the lathe as a tight fitting metric thread was deisired. The ball body nut is actually used to check the thread and also used to form the thread.  Figure 29 and 30 show the fitting of the ball screw nut. Three set screws will keep the holding nut from turning. Figure 31 and 32 shows the finished holding nut relative to the carriage and it’s slide. Figure 33 is just a tapping holder which I made and use for small taps. The tap is held in the holder, internal bearings provide for nice finger “feel” of the tapping and allow the holder to be held by the chuck. Works like a charm and worth making as I have never broke a small tap yet ( even 0-80 or more ). Next steps will be to make up the other bearing block and mounting of the stepper motor to the carriage.
Oh well, more layout to be done.

RICH

Offline RICH

*
  • *
  •  7,427 7,427
    • View Profile
Re: CNC CONVERSION - 6" ATLAS LATHE
« Reply #22 on: September 09, 2009, 04:45:44 PM »
A little redesign of what was in the last post for the bearing. Just didn't like the bearing or assembly.
RICH

Offline RICH

*
  • *
  •  7,427 7,427
    • View Profile
Re: CNC CONVERSION - 6" ATLAS LATHE
« Reply #23 on: September 12, 2009, 11:33:55 AM »
JUST AN UPDATE ON THE CONVERSION
The ball screw was turned down on both sides. I used a chuck which uses 5C collets to hold the ball screw. Only went through 2 triple sided carbide inserts to machine down the screw. The screw hardening was at some 0.070” deep ( just a real PITA! ) on one side.

The motor mounting bracket was mounted to the carriage to confirm fit. The linear travel was tested to see if there was any binding post adjustment of the cross slide gibb and bearing preload. This was done by feel and manual turning the screw. The motor and ball screw were also aligned. I have dimensions on how to position everything thus very easy to do.

I figured I would do a preliminary check of the torque required to just move the axis. Only 11 oz-in is required. Figure #35 shows the torque gauge used for the check. It is very accurate gauge and comes in handy. Just one of those things that you get a bargain on over the years ( they are rather expensive! ).

Some trimming of the motor bracket was done along with providing access holes for the carriage gibb screws. I am waiting for bearings ordered.  A single bearing will be added on the motor side. So you don’t see one in Figure #36. Additionally the coupling will be replaced with a solid one since axial loading will
be very minor, if any at all, back into motor. The adjustable carriage and motor mount along with the screw bearing arrangement provide for precise coupling alignment.
 
I just had to know what the backlash would be. Steps per unit were calculated and slightly adjusted based on testing. Calculated was12690 and was adjusted to 12700. The adjusted steps per unit provided repeatable actual movements, both full and incrementally, when going in one direction over the full length of screw travel. Same for the opposite direction. The screw is a precision preloaded ground screw.
That done, backlash was checked ………..life will be good ……..only 0.0003”. So it seems that the x axis positioning can be used without any backlash compensation. I use a SmoothStepper and it doesn’t provide for backlash compensation in Mach. Backlash is a PITA when trying to cut threads  it doesn’t take much to screw up a small fine thread.  All starting to look promising, but,…….never know until you actually use the
lathe.

RICH

Offline RICH

*
  • *
  •  7,427 7,427
    • View Profile
Re: CNC CONVERSION - 6" ATLAS LATHE
« Reply #24 on: September 14, 2009, 02:33:54 PM »
Time to get back to working on the Z axis. Here are some picks.

Offline RICH

*
  • *
  •  7,427 7,427
    • View Profile
Re: CNC CONVERSION - 6" ATLAS LATHE
« Reply #25 on: September 14, 2009, 02:36:56 PM »
A long way to go but the hard stuff is somewhat out of the way.
It's starting to look like a CNC's lathe.
The next psoting will have some interesting info you may want to review for undestanding.

RICH

Offline Dan13

*
  •  1,208 1,208
    • View Profile
    • DY Engineering
Re: CNC CONVERSION - 6" ATLAS LATHE
« Reply #26 on: September 14, 2009, 03:10:42 PM »
Hi Rich,

A nice looking machine is starting to show up there :)

Thanks for all the updates and photos throughout the process - almost like a DIY manual ;)

Looks like the mechanical part is almost done... we'll see it coming alive soon :)

Daniel

Offline RICH

*
  • *
  •  7,427 7,427
    • View Profile
Re: CNC CONVERSION - 6" ATLAS LATHE
« Reply #27 on: September 14, 2009, 04:30:31 PM »
Z AXIS ADJUSTMENTS – BACKLASH  REMOVAL –BEARING PRELOAD

The Z axis components were adjusted by feel. The axis was turned using a makeshift handle attached to the screw. Now  we get into testing for refining of  the axis components. This may get a little long, but a lot of how to and description is included which may be of value to someone.

The Z axis calculated steps per unit and a basic velocity of 40 was used as rough motor tuning units in Mach. I just needed something so I could jog around using the controller. What I wanted to see is if the carriage is trying to lift or bind in any way. The Z axis ballscrew is not a precision ground screw such that the overall straightness meets a tolerance, but, that tolerance is not like a ground screw. The screw is anchored at the spindle end and the bearing at the tail end allows axial float ( just used for radial loading ) Slowly moving the carriage a slight adjustment of the ballscrew nut to saddle had to made. This was repeated until no movement ( of the carriage off the ways was happening – initially it was around .002” ).

Then the velocity was increased until it skipped at 120 IPM. The velocity was reduced and set at 80 IPM and no skipping happened. ( Just for info the X will do 200 IPM and is set at 100 and it only took 11 oz in to just move the axis ( 31 in oz with a new / different stepper installed ). The difference is the internal pole attract / detent. So we have an initial velocity that works, but, it is governed by how well things are aligned and associated torque required.

I will elaborate more in the refinement comments.

BACKLASH CHECK and REMOVAL
The steps per unit were confirmed. Now backlash was checked by moving one direction and then the other at different distances. The backlash was 0.012”…..

The bearing at the spindle end provides axial anchoring so any play between the outer rings must be removed. The ball nut is adjustable so that is another backlash point. Belt, motor quality, pulleys can play into this but the main two were noted first.

The bearings in the anchor end were re-shimmed. The bearings sit in a housing and the outer rings of he bearings are fixed by compressing them together via the mounting plate and the bearing housing. The inner ring of the bearings receives a preload by tightening a bolt into the screw.

Another back lash check was done, didn’t touch the rough ball nut adjustment, backlash now reduced to 0.008” and I know that the outer bearing rings are making contact. I then tighten the screw which adds the preload on the inner bearing rings. Another backlash check is made and it improves, but the steppers are skipping which indicates the bearing preload is too much. So I reduce it, the steppers don’t skip, and backlash is now at .004”.

Now the ball screw nut is progressively tightened until back lash comes down to around 0.001” and the motors are not skipping. Remember the carriage movement check done in the beginning. When all this stuff is tight there is not much give and alignment is very important. Not bad for a rough testing.

You will never have “0” backlash ( the X axis has 0.0003”) . Remember that lubrication needs to be between the moving components. So absolute “0” backlash is rather difficult to achieve if not impossible.
You can however do your best at each component of the axis train to eliminate it.

There is another part to these adjustments, since as you increase the loading on the bearings and ball screw nut, the torque required to move everything also can increase. So you need to play some to find a good combination.

TURNING TORQUE
Torque will be measured using the gauge shown in one of my other posts.
The torque to rotate the axis was measured and it took 80 in oz ( that’s a lot ). The belt was removed and the motor static detent torque was measured at 20 in oz. Can’t do anything about the motor so your left with 60 in oz. to try to reduce. A stated before the X axis has 21 in oz and one less bearing when measured.
The belt is replaced so I can jog the axis.
The ball nut preload was reduced / played with such that the backlash didn’t change. Then I further played with the anchor assembly preload again not causing an increase in backlash. That done the belt was removed and the torque measured. It’s now 35 in oz to just start to move the carriage. So the torque was refined rather easily and reduced almost in half. ( 60 down to 35 in oz). I would guess that at most the best I can reduce it is around 10 in oz.

The ball screw is not perfectly straight so when it turns you are putting a radial load into the bearings. I believe I can reduce required torque by adjusting the bearing mounted at the tail end of the lathe. That requires shimming the mounting plate and just takes some time fooling with it.

But for now, I know that I can almost eliminate  the backlash  (0.001” current on the Z ). Everything needs to be removed, but I just needed to confirm that all is within what I would like to accomplish before continuing on.

You say what’s the big thing about the torque readings?
Well if you can minimize the required torque for the axis without undue strain on the components keeping backlash within a desired tolerance, then you have a refined machine which will improve velocity and acceleration.

I will note that if you go back to the speeds, feeds, and torque that I was after, they have already been exceeded.
LIFE IS GOOD!

RICH

Offline RICH

*
  • *
  •  7,427 7,427
    • View Profile
Re: CNC CONVERSION - 6" ATLAS LATHE
« Reply #28 on: September 14, 2009, 04:37:30 PM »
Daniel,
Thanks.
Still a ways to go, guards, covers, re-machining the slide for whatever will be attached to it, box for the two encoders, etc.
Movement always seems to light up the fire in oneself.  :D
I forgot why folks go out and buy one!  ;)
RICH

Offline Dan13

*
  •  1,208 1,208
    • View Profile
    • DY Engineering
Re: CNC CONVERSION - 6" ATLAS LATHE
« Reply #29 on: September 15, 2009, 12:48:52 AM »
The ball screw is not perfectly straight so when it turns you are putting a radial load into the bearings. I believe I can reduce required torque by adjusting the bearing mounted at the tail end of the lathe. That requires shimming the mounting plate and just takes some time fooling with it.

Rich,

That is not good. If the screw is not perfectly straight it also means that the ball nut experiences radial loads, and that is really bad. Ball nuts are NOT designed to take radial forces. Subjecting them to this kind of load dramatically increases the wear and rapidly shortens their life.

Daniel