Sunday, January 30, 2022

Weekly Update on the 1 by 1 - Part 22

The work on the engine this past week has all been here in the house on this computer with zero out in the shop.  I'm preparing the connecting rod for the piston.  The piston isn't worth putting into CAD because all the critical things I'll do will be by hand on my big lathe.  

Connecting rods are fiddly little pieces with lots of design features that tend to make us need to use several setups for machining.  I always look at them and wonder why there's so much complexity to them.  I think some of the small features are necessary for clearance and others are just because "that's the way they should look".  This is a lift of the pdf drawing for this engine's rod with a lot of the dimensions removed because they get in the way of seeing what the part is supposed to look like. 

I took out the various radii at the small and large ends.  The small end will hold the pin (a wrist pin) that goes in the piston and transfers the force pushing or pulling the rod onto the walls of the piston.  The big end goes on the crankshaft I made back in September/October (and that tried to to take off the end of my right index finger).   See that taper to the rod from the big end to the small end?  Why is that there?  Other than to make it look better, probably to ensure the rod clears the piston adequately.

The whole reason for importing the drawing into CAD is to make a solid model of it so that I can cut the intricate features with my CNC mill.  The solid model, without the holes seen in the bottom (big) end looks like this:

If you look in the big end you'll see a line that's some sort of line (?) artifact.  I didn't really notice that until I was getting the render ready to post.  The line shows up in this view, but none of the ways of selecting that in the CAD software shows that there's actually something there.  Experience says delete the whole thing and start over.  

There's a big complication here.  I've been using a 3D CAD program called Rhino 3D since I got started with it at home, back around 2005, and while it's full-featured, it has been expensive for a retired hobbyist.  I bought a copy of version 5 in '13 and that stayed current until a few years ago.  The new version (6) received a lot of hoopla for all the the new stuff it does, but the things they emphasized - both prettier rendered drawings and a programming language to create programmable shapes - really didn't mean anything to me.  Last year, they updated to version 7 but I essentially "slept through" the period where I could get it at a big discount.  

I've been trying to learn what the options to replace Rhino are for much of '21 and one that I had a favorable impression of, called Alibre Atom, had their lowest cost option on sale for Black Friday and I took the bait.  There's one big hitch.  Alibre's interface is completely different from Rhino's, which means I'm starting over learning CAD.  Rhino's interface is much like AutoCAD, one of the original CAD programs and probably the best known.  Alibre is said to be more like Solidworks, which I've never really seen.  Whatever the interface is like, it's nothing like Rhino.

Long story abridged, I have some lessons in a book and I'm slogging through it.  This is probably going to initially be a translation effort ("how do I say cylinder?" - or "box?" or any of the commands I use frequently).  After some period of slogging my way through it, things will suddenly become much more clear.  

A saying from school, all those decades ago, comes back to me.  Some of you will grok this in fullness.  "A good Fortran programmer can program Fortran in any language."  That's what I'm waiting for.


  1. Not an engineer, but I suspect the critical features of a connecting rod are tensile strength, mass reduction and stress relief -

  2. I once wrote a (partial) Fortran program and just stopped - I had 100% on all my projects and the instructor dropped the lowest grade, plus it was Friday night and my buddies wanted to go out for beer.

    They pulled it from my account (it was automated) and tried to run it. It ran an infinite loop and auto dumped the memory to green-line paper when it exceeded the memory allocation.

    A lot of green line paper, at least 4" thick, all ribbon-folded up.

    They gave me a zero. I argued, "the grading standard says I get at least 60 points for it compiling . . . "

  3. The taper is in the rod to save material and keep the reciprocating weight down. The loading is different on the small end, and it doesn't need to be as strong, so they remove the material. It also gives plenty of clearance for the piston as the rod moves around in the engine.

  4. Like learning a new programming language, or a spoken language, each new CAD program's interface gets easier to pick up. There are only about three different basic ways to do things, and the rest are all variations.

    The key things are how easy it is to build and modify solid objects using a mouse input. The systems that require you to specify objects with essentially a text command (like "box( 1,2,3,8,8,5)") are a major pain in the ass.

    It should be simple to measure things by clicking on two points in sequence, or edges, or faces, or tangent points on compound surfaces. "Snap to" is critical for want to work in "infinite resolution", the theoretical mathematical point and not where it is in pixels. You need to be able to snap objects together at well-defined joining surfaces. Etc, etc, etc.

    Every CAD program I've used does better or worse at all of these. Stay away from ANY program that works in pixels!

    If you want an example of a brilliant program that uses a purely-text-driven way to describe its world, play with POV-Ray (free). It does beautiful renderings, but the method of describing objects is hell on wheels.

    JMHO. Did it for a living for most of a decade.

  5. Forgot to add: I nearly lost my licensed copy of Autodesk Inventor 2012 (before they went to the leasing model, which I hate) and did a significant search for open-source CAD programs. It seems that there are STILL no decent things to be had out there, though many other fields have free s/w that meets or exceeds expensive dinosaurs (compare Photoshop to Krita for painting, for example).

    This is a big, big lack, people! Get out your compilers and get busy!

    1. Since I did no CAD when employed doing engineering work, I can only state my experience with recent FreeCAD releases were not bad. It is far from a 1.0 release though.