I hope you've enjoyed my saga with the BMC20.
Before putting all the chip shields back on, I was doing some final checks to make sure the gibs & such were set nicely. I ran into something funny....
First, the Y axis. It's perfect. Shoving the table around results in 1-2 tenths of motion, so the gibs are adjusted nicely. Checking for lost motion, which I did by putting a tenths indicator on the table, jogging towards the indicator a couple thou, then jogging away and noteing how much the control's position indication changed before the indicator moved resulted in about 4 tenths of "motion" from the control before the indicator moved. Perfect. (Remember that it's got 38,000 hours!)
The X axis is a different story. Again, checking the gibs by moving the table showed 1-2 tenths of motion at one end and 4 tenths at the other. Good enough for a 20 year old machine.
The lost motion check is another story. The same procedure, jogging the axis towards a tenths indicator, then jogging away, revealed that the axis moves smoothly towards the indicator, then when I reverse the jog wheel's rotation (to move the axis away) the axis jumps TOWARDS the indicator about 1.0 to 1.2 thou before moving smoothly away.
When I invert the directions, ie, move away from the indicator then reverse to move towards the indicator, I get the same jump in the wrong direction before movement in the correct direction starts. I slackened the gibs & keeper plates with no change in results. I indicated the end of the ballscrew, and it doesn't move lengthwise enough to deflect the indicator at all.
Then, I touched the timing belt while the axis was stopped. It's got a lot of miniscule motion going on even though the axis is stationary - it feels like vibration rather than actual movement. I presume this is typical, as the Y and Z axes do the same.
While touching the timing belt I repeated the jog one way then jog the other, and I can feel the servo turning the wrong way for an instant on axis direction reversal before turning the correct way. It does this consistently. Thus, I don't think I have a mechanical lost motion issue.
It sounds like (1) the servo needs tuning; or (2) the backlash comp is set wildly incorrectly; or (3) the axis acceleration is set wrong; or (4) something else that I am not familiar with.
OK, my question: How do I set the backlash comp? I'd like to check what it is, at least. It's running Ultimax II V8.60. I tried all the secret screens that I know (102, 488, 642) and I did see axis acceleration values on 102 but not backlash comp. It was updated from V7.something (I have the old EPROM set), and the dual-axis cards have four, 8-position dip switches on them, and one 4-position dip switch. I haven't seen cards like this before:
Of note, the machine was run for quite some time with the auto lube system inoperable. Well, there was one metering unit that worked on the Z ways. To compensate for this, the previous owners had upped the volume pumped by the auto lube pump to the highest value possible. This had no effect, of course. Presumably running with relatively dry ways required the backlash comp to be adjusted much higher? You'd have though they'd just fix the auto oiler.