Wednesday, March 9, 2016

Well This is Kind of Embarrassing

For the last several months, more consistently since I retired, I've been making the metal parts to convert my G0704 milling machine to CNC control.  So far, every part has been usable.  (Perhaps I should say "appears usable"!)  Today, I mentioned to Mrs. Graybeard that I hadn't botched any part.  Everything seemed usable.  So you know what happens next, right?  It's like saying, "our pitcher has a no-hitter going" or, "gee, I haven't had a flat tire in years".

Immediately after that, I went out into the shop and made my first piece of scrap. The last piece on the list to make is a spacer for the Y-axis.  It's a block of aluminum, 1 3/8 by 2 1/2 on the face and 1 1/2" thick.  It has three through holes: two that are 1/4" diameter and one 0.406" (13/32).  The big hole is counter bored an odd size: 0.945" diameter by 0.360 deep, on both ends.  After examining all of the options I have for cutting a flat bottomed hole like that, I decided to use an end mill and use an interpolated circle cut.  This can be done directly in G-Code, a circle is a single line command and to cut out a circle like this is a four line subroutine repeated: go to the center to drop the cutter, step to a new depth, go back to the circle, cut a new layer deeper, and repeat.  To make it sound even easier, I just did that a few days ago to complete my Z-axis motor mount.  Only problem is that I punched in a wrong number in my calculator, coded it wrong and made the circle off center.
Actually, it's worse than that.  I barfed it up twice - as you can see.  The first path, which I aborted before it cut the first layer, was way far off.  The second was barfed up differently.  In both cases, the root cause of the problem was that I did it in my head without checking it in my CAD program. 

There's an old saying that there's no such thing as whiteout for metal, so measure twice and cut once.  This is why.  Let my bad experience be your learning experience!


14 comments:

  1. Hi SG , run all your new programs at Z-zero depth. The lower your datum .005-.010 for prove out. You will only kiss cut the top and ave a good witness mark for reference. Then go back and edit Z to preferred depth.

    ReplyDelete
    Replies
    1. You know that saying "pride goes before a fall"? I always do that, except I skipped it this time.

      If possible, I have a home made "end mill holder" in 7/16" diameter - because that's the size of Sharpie marker. I chuck up the marker, tape some paper over the work and draw the curve. On the second try, I did that and didn't even notice the circle was off center. I just checked that the OD was right if I added the diameter of the tool.

      Delete
  2. What do you mean by "there's no such thing as whiteout for metal"
    ???

    Heck, even Harbor Freight sells that stuff:
    http://preview.tinyurl.com/btnwayp
    }:-]

    ReplyDelete
  3. Depending on what is going on with the piece, would boring and sleeving be an option?

    Or maybe a clean up pass then an adapter plate (spacer) either on the top or the bottom. The real trick is to place the adapter plate so it looks as if you planned it that way all along.

    ReplyDelete
    Replies
    1. Oops. 0.945". I can't recall what the smallest hole I can bore with the boring head is, but it might not go small enough to bore this.
      Maybe chucking the piece in the lathe's four jaw and then some work with the boring bar?
      I hate to make mistakes, and I hate to throw an almost finished piece into the scrap bin. (sometimes you just can't fix it)

      Delete
    2. The diameter is within the limits of the Sherline boring head, but it's pretty light duty like all the Sherline stuff (which is the reason I got the bigger machine). I'd have to sneak up on it a thousandth or two at a time.

      The problem is that the hole is off center to the left. Boring and sleeving sounds like a possibility, or boring a large hole and putting a smaller cup in there. Not sure how to bond it - I don't have any welding capability. Epoxy might be good enough.


      Delete
  4. Toss it. Start new. You will waste more time on a repair than making it from scratch.

    ReplyDelete
    Replies
    1. But he will learn FAR MORE from the time and effort spent on the repair.

      Besides, he's retired! He ain't got nuthin' better to do anyway...
      }:-]

      Delete
    2. While it's true I'm in that phase of life where I have more time than money, it was a piece from my "junk box". I'm not even sure what kind of allow it is.

      The learning part is the part that seems worthwhile.

      Delete
  5. Never brag on something until it's done, it's the kiss of death! It's like saying, "This will only take a second, and the safety glasses are on the other side of the shop...", or "I don't think I need a backup jackstand for just changing oil..."

    In the first case, you'll be lucky if the piece merely sticks in your forehead. In the second, you'll knock the car off sideways while underneath yanking on the stuck drainplug.

    As somebody said, you're retired. Take the extra time! Pace yourself!

    ReplyDelete
  6. It would be more challenging to make that exact same thing twice.

    ReplyDelete
    Replies
    1. Ah, but that's the nice thing about CAD/CAM. Unless and until you change the code, you will make the same thing over and over and over and over again...
      }:-]

      Delete
  7. The build new vs repair decision can be tricky- I have watched guys chase screw ups all over the part- mis cut one thing, then trim a little on another part so it matches, then the alignment with a third part is compromised ,that sort of thing. Then there is the concern the repair may fail down the road.
    Material cost plays into it , and also of course, meeting spec, but in this case you have no one to please but yourself.
    This supports a critical component? What are the failure consequences?

    ReplyDelete