Friday, June 19, 2020

Shop Update

It has been almost three weeks since I posted an update.  I ended up spending more time than usual in the ham shack instead of the metal shop last week.  The parts I'm working on are either small or fiddly or both.  Let me show you some of what I've been up to.


The top piece is the simplest, but still a fiddly bit.  It's a washer made from brass rod that's 1/4" OD with a 1/8" hole down the center.  It's easy to do this; the only touchy operation is cutting off a .015" thick sliver of a brass rod with a 1/8" hole down its center.  The cutoff tool usually leaves a nib that has to be trimmed off and then the whole thing needed to be lapped on sandpaper to bring the thickness down about .003".  I could make washers like this and sell them, but one of these would cost you the same as a box of 1000 stock washers.  It's kind of silly to make things like this.

The piece of bar at lower left is straightforward cut a piece of 1/4" drill rod to 0.938" long and then move it to the mill (or good drill press) to drill a 1/16" cross hole through it 0.844" from one end. 

The piece at the lower right is the blank for the cam pictured just to its left.  I've never made a cam like this and there's a couple of ways to make them.  The plans I'm using recommend using a method developed by a guy named Hamilton Upshur, which requires building a fixture to turn the cam eccentrically - that is, off the centerline of that big 3/8" reamed hole in the center.   I had no idea what that fixture needed to be until I found a video that's one of a series where this guy takes you through building an entire engine that Upshur published.  Late in the video, I got a good look at some prints he kept using, paused the video, screen captured the frame and improved its appearance. 


The fixture I'm making is at the upper right.  In the first picture, the 3/8" "peg" the designer calls out is lying on the top right edge of the print; the big 1" diameter cylinder that holds the peg is waiting to be drilled for it on the mill.  The peg is through-tapped, 8-32 instead of 6-32 because I prefer 8-32 screws.  Tomorrow, I should be able to complete the fixture and maybe make my first attempt at the cam.  A finished cam is in the middle of this screen capture, and it looks very much like the one I'm trying to make.  With this method, the cam is largely machined by hand; the fixture just makes it easier to hold the work for those odd cuts.

There's another method for cutting cams that has been shared by a home engine modeler named Chuck Fellows, and his video details how he got there.  It seems a bit fussier to set up than this, but probably gives a more controlled shape.  Fellow's method fixes the cam blank on a rotary table which then moves for all the cuts while Upshur's method uses a lot of hand work; tighten down the screw that holds the blank in place, make a cut, loosen the screw, change the blank's position, tighten the screw, make a cut, and repeat.  Over and over.  
 
The topic of making cams is of great importance for engine manufacturers, especially generating the curves mathematically.  I've read several things on this, including equations that can be plugged into an Excel-compatible spreadsheet.  Take this exercise:  the green circle is the starting blank; the big white circles cut the left and right flanks; the center points where the center of a boring head should sit while rotating to cut those big white circles are computed and displayed.  The boring head is cutting an inside diameter, not an outside diameter. 


This rabbit hole is as deep as you want to go and as techno-geeky as you want to get. 



14 comments:

  1. The only thing like this I made was a crankshaft for the project I had to make in college. Learned all about offset turning, and getting things fixtured up.

    I wouldn't attempt to make that cam lobe on a bet.....a BIG bet!

    ReplyDelete
  2. Yeah that looks way too fiddly for me to be attempting but I'm wondering why you can't just offset the piece in a 4 jaw chuck?

    ReplyDelete
  3. Cam looks like a job for the cnc mill to me. Bolt the blank down to the table, through the hole, on top of a sacrificial piece of aluminum. Indicate on the outside to figure out where it is. CNC the outside profile a little bit large with the side of an end mill, then improve the surface finish manually with a fine grit sanding drum on a dremel.

    ReplyDelete
  4. Anonymous has it right. You have a CNC mill, right? Make whatever profile you like, and forget the lathe (unless you want to do it the hard way, for nostalgia or something).

    ReplyDelete
    Replies
    1. I've been thinking of CNC for the fall back.

      Everything I do on a milling machine is under CNC control. Even if it's not running a program for a part, but just running a series of "GoTo" commands (G01 X1.234 kind of statement) to either square an edge or drill a hole. Lately it's been bothering me that CNC slows me down for some things. When I needed to turn a 1-1/2" diameter piece of aluminum down to 1", I just stick the piece in the lathe and cut. I converted one of my Sherline lathes to CNC to cut threads; it could just as easily reduce diameters, but I do those things on the manual lathe.

      This fixture looks like it's going to be more work than I first thought but maybe it gets kind of like doing step and repeat often enough and you remember the motions.

      Delete
    2. You may not have this option available to you (I don't know what your background in software is), but I found it convenient to write up a few simple Windows dialogs that make different basic shapes. They output G-code from a collection of parameters to do things like cutting mortises and tenons, or a collection of holes at various locations and depths. If I had to produce cams of various types, I would definitely parameterize the process (with numbers like lift and rotational angle range) to feed g-code into linuxcnc.

      Of course, g-code is a whole language and you can make parametric programs with it, but I find C++ to be so much easier to use.

      Delete
    3. I'd never thought of that, but the calculation spreadsheet I mentioned around that last graphic might parameterize the process. I've written some large pieces of software but over the years went to more elaborate platforms, stuff like IDE programming. For the last 25 years, I've only written in Delphi. I have done things like program tool paths in Excel and port them to text files.

      My attempt at using the fixture resulted in something that looked vaguely cam-like, but any dimension that matches the drawing is purely coincidental. Meanwhile, I do have a CAM program that can do "waterline" tool paths and I have a Gcode file that's ready to try this afternoon.

      Delete
    4. The open source drawing program Inkscape has a plugin that produces gcode. It's very primitive and only suitable for 2D; for example you have to compensate for the width of the cutter by moving your path. But you could trace over an image of the cam and turn that line drawing into gcode.

      Delete
  5. I consider it a criticism of cnc that the most attractive manual semi-interactive use cnc interface I've seen on youtube is...the hydraulic tracer attachment Keith Fenner uses. When he wants a round profile, he just clamps a radius gage to the tracer rail and fiddles with it. Meanwhile the manual knobs provide offset (geometric translation) controls which real cnc makes harder to do. If he wants the full cnc treatment he can cut a full-length profile, which is functionally a cam, out of sheet metal or masonite on his cnc.

    But nobody says you aren't allowed to simulate a tracer attachment user interface in software. There's a joke that goes, every software problem can be fixed with an additional level of indirection.

    ReplyDelete
    Replies
    1. It took me a while to go find a video, but (assuming this is what you meant https://www.youtube.com/watch?v=c6d57HgTUm0 ) that's an "old time" pattern duplicator. There's an example of one making stocks for M1 Garand rifles in WWII here: https://thesilicongraybeard.blogspot.com/2018/07/interesting-page-on-m1-garand.html

      There is software for duplicating shapes on a CNC lathe, but what I'm familiar with is for making something on center - you program the profile the entire part gets, like a baseball bat, say. For something like this that's turned off center, the cross slide has to be exactly in sync with the headstock. I don't know if mine could do it because my headstock motor is a free running DC motor and sync is carried out once per revolution. I've been meaning to upgrade that combination but haven't.

      Delete
  6. This comment has been removed by a blog administrator.

    ReplyDelete
  7. I don't know if you even have a 3D printer but I thought you might like the info anyway, if you can manage to follow Ol' Uncle Bumblefuck's rambling instructions.
    How to snag 3D parts files from the McMaster Carr website to use for making 3D printed parts.

    https://youtu.be/kLL-ydbr9DQ

    ReplyDelete
    Replies
    1. I don't have a 3D printer. Yet. But I've been using McMaster's models for screws and generic hardware for a year, maybe more. There is not enough time in life for drawing a screw every time you need one. It's only something you do when you need to make sure something will fit.

      Since I don't have a printer, I never thought of saving a pulley or something to print later.


      Delete