Backlash - by John Robbins (continued from above)

“It’s a 16 x 20” vertical mill.  We’re making a fly cut .030 deep across a 6 x 6” block of 4140 steel.  We use a 1/2” end mill to interpolate a 4” pocket .150 deep.  Then we drill four 3/8” holes in a square pattern through the block.

You say this part has been running for weeks.  Has anyone altered the program? 

“No, the only thing we change are the tool offsets.” 

How about the fixture.  Could it be moving? 

“We’re using a Kurt vice with a hard stop to position the part.  I’ve checked it to be sure it’s tight.”

Has anyone been working on the machine?  Is there a chance that any parts or parameters have been changed?

“No one has serviced the machine for months.  It’s been running fine up until now.”  We’re having two problems.  The holes are supposed to be in a square pattern.  Their moving anywhere from 2 to 30 thousands out of position.  The other problem is with the circular pocket.  It’s leaving step overs at the 3:00 and 9:00 positions.                  

Are you pushing the machine too hard?  What’s the spindle load?

 “The spindle is pushing 80%.  It’s been like this since we started.  The operator on third shift tried turning the feed rate override up last week.  He broke off two of the1/2” end mills so he set it back to 100%.

(In the cartoons this is the part where the light bulb goes on over your head)

Oh, by any chance were there any servo alarms when this happened?

“I don’t know.  He powered the machine down and replaced the tool.  Now that you mention it, this is about the time the problems started.  But the machine kept running after that happened.”

I suspect that the thrust bearings or ball screw may have been damaged when the tools snapped.  Do you notice any unusual noise when the axis is moved in rapid?

“The X-axis seems like it’s growling more.”

 Here’s a simple test that I’d like you to make.

1. Set up dial indicator in the spindle and touch off on the solid jaw of the vice.  Make sure the styles is parallel to the axis of motion.

2. Program a feed move of three inches away from the vice then move 3.050 back into the vice.  Insert a dwell into the program at this point and observe the reading.

3. Now move .050 away from the vice.  Dwell again and record the value.

4. Repeat this procedure several times for all axes.

“The Y and Z are on the money but the X axis is off by .030.”

You’ve got a backlash problem in the X-axis.

The difference between the commanded move and the measured value is called backlash.  This is mechanical slop in the drive mechanism.   Something’s slipping.

When a machine has backlash the axis doesn’t keep up with the motor movement.  This is caused by the ballscrew shifting instead of moving the axis.  As long as you’re moving in the same direction you won’t see a problem.  When the motor reverses direction it thinks it’s gone where it was supposed to.  The problem is the axis didn’t move along with it because the ballscrew shifted.

This is most evident in a circular cut.  If the backlash is in the x-axis, such as we have here, a step over will be seen at the 3:00 and 9:00 positions.  This is where the x-axis reverses direction.

It can also be seen when doing a bolt hole pattern.  When moving straight the holes are in the proper location.  When the Y-axis shifts down and the X-axis reverses direction the motor moves but the table hesitates causing a miss-positioning error.

In this situation the most likely culprit is the thrust bearings.  The bearings are supposed to prevent the ballscrew from shifting back and forth while allowing it to rotate.  In a crash they are the weak link in the system.  One of the signs of a bad set of thrust bearings is a growling sound coming from the axis.

The best method to check for machine inaccuracy is a Ball-Bar test.  This gives a report on the machine similar to that given by a car analyzer. 

If you want to be certain that the ballscrew is shifting use a dab of grease to attach a small ball bearing to the end of the ballscrew.  Set up the indicator to touch off the bearing. The bearing gives the indicator a smooth surface to measure off of.  Now move the axis in jog or hand wheel mode back and forth an inch.  When the motor reverses directions watch the indicator to see if the ballscrew is shifting.  It shouldn’t move.

In many situations you can feel backlash with your hand.  With the power off grab the ball screw and rotate it back and forth.  If you feel any free motion before the table starts to move then you have backlash 

Many controls have the ability to compensate for small amounts of backlash. Adjusting for a couple of thousands usually isn’t a problem.   A backlash of .030 is a major mechanical problem that will only continue to get worse.  Replace the bearing and retest.

Keep in mind that this is just one probable cause.  I’ll cover the other possible solutions in future issues.  I’m running out of paper.

TOP

Programming Tips - by Rick Monday (continued from above)

The ability to write work offsets comes in handy when you are running palletized machines.  Most times the fixtures or vises stay mounted to the pallet.  When the job changes the operator simply changes the pallets.  He then calls or downloads the correct program.  The operator would then have to punch in the new work offsets by hand.  Since the fixtures are never removed from the pallets, why not put the work offsets in the program.  This way when the cycle start button is pushed the correct offsets are loaded via the program.  This will eliminate operator errors caused by hitting the wrong button, or by misreading an indicator.

For machining centers with more than 6 work offsets, such as G54.1 P1 thru G54.1 P48, the line remains the same with the exception of the L word.  The line will read as such:

G10 L20 P36 X-9.5000 Y-11.0000 Z-11.8000 ;

The L20 word refers to work offsets above G59.  In the above example G54.1 P36 would be changed to the new values.  Remember, the offsets in the registry are actually changed, the old values will be replaced with the new values.

Horizontal machining center programmers have used this method for a long time.  Many work offsets can be used on one pallet.  When the next pallet comes in, the program will load new work offsets.

Try this the next time you have fixture that will stay on a pallet, I am sure you will find that it saves time.

TOP