|
Posted by Eric R Snow on April 20, 2006, 8:56 pm
Please log in for more thread options
wrote:
>
>>>
>> Robin,
>> Any screw with any clearance will make the table jump in a sort of
>> stick/slip scenario. Unless the table has drag on it. Using a
>> ballscrew will cause all kinds of problems unless it too is run with
>> drag. On a CNC machine the servomotor provides this drag. It's not
>> generally looked upon as drag, but it does prevent the pressure from
>> making the screw turn enough to lose position within the parameters of
>> the machine control. So if a climbing cutter pulls the table into the
>> cut then the ballscrew will be rotated to move the work out of the
>> cut. Then the direction reverses so the cutter is again climbing. Then
>> the screw reverses, etc. If effect dragging.
>
>Are you sure the servo actively acounts for backlash? As I understand it,
>the ballscrew nuts are preloaded to virtually eliminate backlash. I know
>that in older worn machines, there are control paremeters that can account
>for backlash, but because there is give in the (worn) screw, the fix is not
>absolute. Interpolating arcs which cross 0º, 90º, 180º and 270º is
>impossible and results in a slight flat is left on the work piece as the
>axis reverses direction and the table is left floating.
>
>Again, I'm no expert but I cannot imagine 3D machining would be possible if
>not for a mechanically rigid machine which would include preloaded nuts.
>
>> On the manual machine
>> using acme screws the table locks are set to provide enough drag to
>> prevent the cutter from taking up the backlash in the screw.
>
>I am again interested in whether you have issues with the cutter grabbing
>you work. I suppose I'm beating this subject to death, but you're basically
>the only person I've ever heard specifying climbing on a manual machine.
>
>> As
>> regards your method of milling it has been my experience that when
>> hogging material it works better, and gives longer cutter life, if the
>> cutter is fed into the work at almost the full cutter diameter. So
>> instead of using a shallow radial, deep axial cut, I use a deep
>> radial, shallow axial cut. This keeps more of the teeth cutting at one
>> time and is easier on the machine. BTW, I believe the rotating cutter
>> that removes the most material in the shortest time is a twist drill.
>> That is if all the material in converted to chips. Obviously an
>> annular cutter will remove more material, especially if the diameter
>> is large.
>
>I've always shared your observation (and have read articles which support it
>as well). I remember there were quite a few articles written about "plunge
>roughing" a couple of years ago. I don't do work which requires that type of
>roughing but I can imagine it would be very efficient if applied correctly.
>
>Regards,
>
>Robin
>
Greetings Robin,
I guess I wasn't clear. When climb milling I use the table locks to
put enough drag on the table to keep the cutter from pulling into the
work. And I'm not alone. In all the shops I've worked in this is a
common method of machining. Especially because of the increased cutter
life. Now, let's talk about the active backlash compensation going on
in a machine tool. Ballscrews can be set up with virtually zero
backlash. Close but not quite zero. Thrust bearings are set up when
the machine is new to have zero backlash. But everything wears in the
machine and backlash will eventaully occur. So CNC controls have a
parameter for backlash compensation. Even my 28 year old Fanuc 5T
controls have this parameter. So whenever the the rotation of the
ballscrew reverses the backlash amount is added to the move. But even
in a machine with zero backlash the servo motor has to reverse
direction when the cutter grabs. It then must go forward to advance
the cut. This all happens very fast and so it looks like the motor is
always going forward. If the machine has a brushed DC servo motor, and
only two brushes, then the motor can only spin one way or the other.
It can't just stay locked in one place. If a load is placed on the
motor, and the current is adjusted to exactly match the load, then the
motor can be locked in place. But if the load varies enough then the
motor direction will need to reverse. First the load opposes the
rotation, then it helps the rotation. If you don't want the motor
spinning faster than the command the voltage must be reversed and the
motor will reverse direction if there is any backlash. If you can,
notice what the motor is doing on a cnc machine when it is stationary.
It will be vibrating back and forth just a little. The machine has
some kind of position feedback and usually the position sensor is not
mounted directly to the motor shaft. So if the encoder is on the end
of the ball screw then the very tiny amount of belt stretch will
result in backlash that does not need to be programmed for, but will
cause the motor to vibrate because it will spin one direction, stretch
the belt a little, and then the ballscrew will rotate and the encoder
will sense this. When the ballscrew rotation overshoots then the motor
reverses. When it first reverses the tension on the belt is released
and then it is stretched in the opposite direction. Finally the belt
stops stretching and the ballscrew begins to turn. Eventually it will
turn enough for the encoder to sense the rotation and the motor
reverses again. This is where the vibration comes from. Now, my hands
hurt from too much typing so I'm gonna quit.
Cheers,
Eric
| 
Yahoo! 
Google 
Windows Live 
del.icio.us 
digg 
Netscape 
XML Sitemap