Sunday, January 29, 2012

Today's Machine Shop Learning Experience

I normally use a stubby .3125" twist drill to create a 3/8"-16 tapped hole in the end of a piece of acetal.  This hole is used to (among other actions) hold it in position so that the chop saw can cut a 30 degree slice off the end.  (It fits into a 60 degree angle tripod leg for the Losmandy G11.)

I could not find the stubby .3125" twist drill, so I went to my drill box, and pulled out the full length drill that is in the .3125" index hole, and assumed that because it fit there, it was the right size.  So I used that to drill the hole, then tapped the hole.

The lesson: the 3/8"-16 bolt that I used to hold these workpieces in place on the lathe would not hold very tightly.  It was not flopping about in there, but it also wasn't tight, either.  It was tight enough to hold the workpiece in place while trimming the outer diameter to size, but it did not feel right. 

The consequence of this was that when I put the workpiece on the jig that holds it in position for the chop saw to make the 30 degree cut, it was a little loose--but it did not feel like it was going to go anywhere.  Oh my, was I wrong!  The blade hit the workpiece, and it proceeded to spin the workpiece off the 3/8"-16 bolt with an impressive velocity.  When it hit the far wall of the garage, it made a rather impressive thunk!  I do not know how fast it was going, but the accuracy with which it hit the target was startling, and for the same reason that rifling in a gun barrel makes the bullet accurate: gyroscopic stabilization!

At this point, I concluded that something was terribly wrong.  Had my son done something wrong when tapping the holes?  Or was that .3125" twist drill not what I thought it was?  My son had done everything right; the twist drill was actually more like .3145".  A .020" difference does not sound like much, but it is the difference between a tight thread and one so loose that workpieces become inverse rifled projectiles.

At this point, I should have stopped, and redrilled and tapped these holes for the next size up, such as 1/2"-13.  Instead, because I was in a hurry (always a mistake in the machine shop), I decided to use a different fixture to hold the workpieces in place while doing the 30 degree slice.  Of course, this fixture is for a different purpose, and was not quite as steady.  But it looked like 30 degrees.

So now I compounded the error by drilling and tapping the bottom of the 30 degree cut, and trying to assemble the final product.  And the holes are in the wrong places!  Instead of analyzing the failure, I move on to a second set of miscut parts, and just move the holes.  Now I can assembly everything--but when I put them into a tripod, I discover that they are more like 35 degree cuts, not 30 degree cuts.  It looks bad, it won't work for the customer, and I have now wasted $22 worth of materials and about three hours of labor.

1. If you change tooling for any reason, verify that the new tooling is dimensionally identical.  Do not trust an index hole in a drill box.

2. If any part of your process produces unexpected results, do not keep moving forward with the output.  It may indicate a serious defect.  Moving forward just wastes time and material.


  1. Another thing to look into would be not just the diameter but the straightness of your full length twist drill. Chuck it up and check for any wobble out at the end. I think that would be a bigger problem than just a couple thousandths diameter change.

  2. You could charge Boeing Millions of Dollars to tell them that. They'd ignore you, but they'd pay.

    BTW, personally, I've been bitten by the CNC bug and am planning to build a CNC router. From what I've been reading, it's both simpler and more complex than you'd first think.

  3. If any part of your process produces unexpected results, do not keep moving forward with the output.

    While Americans have a tendency to focus on the inventory savings of Just In Time manufacturing, the above quality guideline is as or more important. Catching a problem as soon as possible results in higher quality and less waste; the counter example I like to use is apropos:

    "Suppose you order widgets for quarterly delivery, and a jig at your supplier gets out of alignment partway through the quarter. Everyone is going to lose when you take delivery of a quarter's production and at some point discover you can't use a large fraction of it."

  4. 0.3150" is 8mm. The closest ANSI drill is "O" at 0.3160".

  5. The good news is that the stubby .3125" drill bit reappeared while cleaning up.