Special Pages

Saturday, April 23, 2022

Update on the 1 by 1 - Part 24

I'm not going to call these "Weekly" updates until they are again.  

The best summary of things since update 23 two weeks ago is that old saying, "the best laid plans of mice and men oft go astray."  At the time, I had hoped to complete the test piece of the epoxy-embedding method I was going to try within a week.  I had decided to do two separate operations; a rough pass with 1/4" ball end cutter followed by a fine pass with a 1/4" square end cutter.  Ordinarily, the rough is done with a square end and the fine with a ball nose.  My mistake.  For the rough cut, I told the CAM program to leave .025 margin everywhere.  The tool paths look like this in the visualization program I use (GWizard Editor).

The process of cutting down the rough pass took about 90 minutes but I ran into troubles with something I've had trouble with before - the cutter getting pushed up into its holder as the cut progressed.  

Take a look at the bottom center of the picture and you can see a group of parallel lines running diagonally between the circular stack of lines on the left and the complex shape on the right; that's a perspective view of the layers.  This is where the cutter is taking off layers in 0.025"deep cuts as I told it to do. When it started clearing out the large flat areas, like from the bottom or upper left corners over to that complex shape on the right, the CAM program told the mill to plunge the cutter to full depth all at once.  Once you're used to seeing the part in this way, it's easy to see it doesn't do the large flat areas in 0.025" at a time passes; there's just one layer at full depth.  Instead of lowering the cutter 0.025", it plunges it a full 0.200".  It should have cut down to 0.220" but the program wouldn't do that because 0.220 isn't an integer multiple of 0.025" so it stopped at the closest multiple.  The problem is that during those deep plunges the cutter can get driven up into its holder which totally messes up the depth of cut.  The first pass looked like this (the picture was shot while it was still cutting). 


After a few tries at repositioning the cutter in the collet that holds it, I switched to a purpose made "end mill holder" that has a setscrew.  This meant switching to the square end cutter because the other didn't have a flat spot on its side for a setscrew.  That held the cutter in place, but then I got hit by another problem, this time one I've never seen before.  The CNC controller program, Mach3, stopped moving the mill giving me an indication that the Emergency Stop button had been hit.  I was straightening up in the shop, so maybe it's possible I brushed into the E-Stop button, but if I did, I didn't press it far enough to lock off.  I don't know if that's even possible.  My guess is it was a power glitch, which we've had issues with before.  When I resumed the program, it was cutting too deep and the computer's coordinates for where it was didn't match where it actually was.   It cut off one of the major design features of the part.  At that point, the part became scrap but I kept going because this is a learning experience and I have more material to make the real one. 

It took days to get over this hump and start cutting the fine pass.  The results of the first fine pass cut looked like this:

Those "stair step" artifacts on the edges of the tapered rod between the two round features (and everywhere you look) are apparent on the simulations of the tool path but I was rather disgusted with this result and started trying to figure out how to improve my results.  The fix was to do yet a third pass with a different method than the other two tool paths. I ended up doing a third and fourth pass.

The rough path uses what's called waterline.  It cuts layers around the design, with fixed distances between the layers and between the cutter and the part.  Both X and Y change constantly.  The fine path leaves no clearance around the part, but cuts in in X motion, back and forth, left to right then right to left.  The only time Y moves is to move the work for the next purely X-axis cut.  All that I control is how close the tool paths are to each other and how far along the path the cutter moves at a time.  That was what caused the stair steps.  The cure to this is what's called a "contour" path, which is cutting along the edge of the part at a set distance - a waterline path at full depth of cut.  I ran two different contour paths; the first getting no closer to the final shape than .004" and the second allowing it to touch the outline.  That gave me this dramatic improvement.

The sides of the shaft are much better.  A nice bonus to cleaning it up is that the first pass took 1-1/2 minutes and the second (with finer steps) took 1-3/4 minutes.  The drawback is it took me until Thursday to get this result.  Compare that to the 90 minutes on the rough pass and re-cutting parts of it several times, and about the same 90 minutes for the initial fine pass.

Since the part is scrap, I went ahead and figured out how embed it in epoxy, which is hardening now.  

The second side should take far less time.  The guy I'm copying says he let the epoxy cure "several hours" and it was fine, but it's a different epoxy than I have.  I'll let it cure overnight.  



8 comments:

  1. Any progress is better than no progress. At least you're getting a handle on what needs to be done.

    And it's really starting to look like a Connecting Rod!

    ReplyDelete
    Replies
    1. The end result of all of this is that I'm going to use a different scheme for the other side of the blank. I think it will improve things all the way around.

      Delete
  2. You're getting there!

    I just have a little 3017 CNC that I use to do plates for switches and other instrumentation in .017 aluminum and plastic and wood. I have to pause for two months for cataract surgery and a trip back East.

    I am NOT a machinist, but I do understand "feeds and speeds" enough to not bang the cutter into the workpiece. At least, not yet.

    ReplyDelete
    Replies
    1. I was recommending one of those to someone for the new ATF "everything has to get a serial number" edict. I engraved my 80% lower on my Sherline. Probably the best way to make a decent looking engraved SN. It has to be .003" deep (I assume so people just don't paint it on).

      I do have some dings in some of the work holding fixtures around here, but I haven't had to replace of them because of it - or broken any cutters. I think that makes me ahead of the game.

      Delete
  3. This comment has been removed by the author.

    ReplyDelete
  4. Often we learn more from our mistakes than from our successes. The lessons you learn here will translate to fewer mistake down the line. CNC is not nearly as simple as most people think.

    I have had a little training in a machine shop (that was a small part of the company where I worked) for some basic things on manual machines. I was blessed with being taught by a gentleman who had around 50 years of experience. I really miss having access to a machine shop since leaving that job 9 years ago. It's amazing how much simpler some thing are with the right tools, instead of having to try to make do with a hacksaw, power drill, files, and emery cloth.

    ReplyDelete
    Replies
    1. Often we learn more from our mistakes than from our successes.

      True in all of life, in my experience.

      I've had no formal training in any of this. When I got my first (broken) CNC milling machine, it was the first milling machine I'd ever touched. Got it resurrected, running fine and still use it, for the smaller "fiddly bits" I have to make.

      Delete
  5. for the new ATF "everything has to get a serial number" edict

    People made things without serial numbers because they thought that was conservative. Now Simon Says they must have serial numbers, so they'll put serial numbers on them, in a legal climate more menacing than the earlier legal climate that inspired them to make private things? I'm just waiting for ATF to tell all the older conservatives to report to Carousel, which is next door to the Purina Soylent Green factory. They'll go.

    ReplyDelete