• CAD - offer my comment on AutoCAD and Solidworks

    From Richard Smith@21:1/5 to All on Sun Sep 11 20:55:09 2022
    Hi there

    I had someone I know contact me for advice on CAD programs.
    We share a science specialism, by the way.

    I visualised he was making a career-significant decision, so gave him
    a huge answer.

    Turns out he needed to draw in 2-D plan the position of a garage for a neighbour...(!)

    So - I offer what I wrote to my colleague.

    For what it's worth...




    Issues - what you will have to choose:
    * pointy-clicky (eg. "Solidworks") or interpreter / command (eg. "AutoCAD")
    * 2-D drawings or 3-D visualisations
    * how many computer architures do you want / need to be able to work
    on? (see further on - some leave you "trapped" on just one
    computer-family)

    I self-taught.

    Was 20 years ago.
    Results were like as in http://www.weldsmith.co.uk/tech/casting/casting_fdry_credentials.pdf
    see page 3

    That is using AutoCAD in 3-D.

    The tutorial I used was:

    May 2000
    "AutoCAD 2000 in 3D - A Monkish Shot Tower"
    University of New South Wales
    Faculty of the Built Environment
    Tutorial Introductions to CAD
    AutoCAD 2000 in 3D - A Monkish Shot Tower
    Jim Plume

    This is a PDF, obviously.

    When you can get through that, you can probably run from there.


    Now is your big choice - "be assimilated by The Borg, or run free as a maverick".

    "AutoCAD" has been around for a long time - as have others with that
    "computer architecture".

    I am writing in "emacs" (plausibly it's an abbreviation for "editor
    macros").
    It has been around since the late 1970's - 46 years now.
    (maybe see https://en.wikipedia.org/wiki/Emacs)

    I'm putting a steel beam across a 14m span - it's 200by20 flange and
    400mm tall with 10mm web (that's a fictitious size), from S275 steel
    (275MPa yield). What force will it bear at the midspan before
    bending?

    (/
    (beam-fmax-ibeam-simple-cload 200e-3 400e-3 10e-3 20e-3 14 275e6) ;; 128836.19047619044
    9.81 ;; g, N.m^-1
    1e3 ;; kg per Tonne
    ) ;; 13.13314887626814

    I did not type those answers.
    The text editor inserted the answers into the buffer because I told it
    to do so.

    So, that's 13 Tonnes...

    How did that happen?
    I accessed the interpreter and got it to run my program - actually a
    suite of functions where I started off with little ones down at
    calculating the Second Moment of Area, and built from the bottom up
    until the top invocation is like a language for expressing the
    challenge.
    I've even written a simple computer-numerical solution for heat flow.
    That runs for ever, until an inspect-the-current-situation function
    it calls every time it goes around the solution says "stop" (you've
    got there).

    AutoCAD has that architecture. At the core it has an interpreter,
    written in a compiled language and compiled to a machine-code binary.
    Just about all the functionality you see - it's layers and layers of
    functions "riding" on the interpreter.
    My little "beam load-bear" function is a simple version of that.
    You read from the inside-middle outwards to the top-and-bottom.

    "
    ...
    (simple-support-dblbeam-loadcap
    (beam-moment-capacity
    ib-stl-sigmamax
    (z-plt-ibeam ib-width ib-depth ib-web-thk ib-flange-thk))
    ib-length)))
    "

    "z" hides inside it 2nd moment of area "I"

    Programs with this architecture have gone on for decades, only
    terminating if something major changes (equivalent of steam-engines to diesel-engines).
    All the commands stay the same.
    You do get additions in time - eg. when computers became powerful
    enough and had enough storage to handle images, you got image handling
    and inserting commands... But in the totally familiar style of all
    other commands.
    And - they the run on any computer... The interpreter is the only
    thing which needs to be "compiled" into the machine code of another
    physical computer type. All the "scripts" / "functions" - they are
    unchanged. The "ported" interpreter swallows them unchanged. It
    presents the same interface swallowing "functions". That's the point.

    I think "Solidworks" only works on "MSWindows" computers, and is
    "trapped" there.
    And such programs get superceded. So all your work gets "orphaned" occasionally.

    If there were a computer the "McKludge ZYX321" (eponymous other
    less-well-known computer), all interpreters of the interpreter/scripts architecture would have been "ported" onto it - including "emacs",
    including "AutoCAD", etc.

    AutoCAD exposes its command-line interface for drawing, and I used it
    as an educational tool for developing the first principles of
    geometrical thinking in my students. I could see who'd "got it"
    because I could stand at the back of the class speaking instructions
    and watching what was coming up as the shape expressed on their
    screens.

    You should search the Web.
    Find your own way.

    Best wishes,

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Richard Smith@21:1/5 to Jim Wilkins on Sun Sep 11 23:44:57 2022
    "Jim Wilkins" <muratlanne@gmail.com> writes:

    ...
    Your work is impressive. I could immediately see why the initial
    casting was flawed and the revised one sound.
    ...

    Not really.
    This is "Foundry 101"
    Commercial foundries use chills all the time in sand-casting, I
    understand. Blocks of steel embedded in the sand where they need to
    force the casting to solidify first. In order to get sound castings.

    If the comment were relating to clear style of explaining, I will have
    that...


    I learned sand casting at age four at the small foundry next
    door. They expected me to play in the sand pile but I was more
    interested in what sprues and risers did, which I could see when the
    mold cooled enough to approach. They used simplified patterns and
    sculpted the details by hand.

    :-)

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Jim Wilkins@21:1/5 to All on Sun Sep 11 18:27:51 2022
    "Richard Smith" wrote in message news:ly35cxptya.fsf@void.com...

    For what it's worth...

    -------------

    The engineers kept the juicy parts for themselves, Solidworks and Spice, and
    I had to content myself with the leftover crumbs which were CAD design of printed circuit boards and simulation of digital circuits. I use a 2007 demo version of PADS board layout software for mechanical drawing but it can't do any mechanical simulation and doesn't export .dxf properly, and anyway I
    didn't retain enough of Statics and need to practice manual beam strength calculation. Then I proof test my assemblies with load cells.

    Your work is impressive. I could immediately see why the initial casting was flawed and the revised one sound.

    I learned sand casting at age four at the small foundry next door. They expected me to play in the sand pile but I was more interested in what
    sprues and risers did, which I could see when the mold cooled enough to approach. They used simplified patterns and sculpted the details by hand.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Jim Wilkins@21:1/5 to Jim Wilkins on Sun Sep 11 21:38:44 2022
    "Richard Smith" wrote in message news:lyy1upzg2e.fsf@void.com...

    "Jim Wilkins" <muratlanne@gmail.com> writes:

    ...
    Your work is impressive. I could immediately see why the initial
    casting was flawed and the revised one sound.
    ...

    Not really.
    This is "Foundry 101"
    Commercial foundries use chills all the time in sand-casting, I
    understand. Blocks of steel embedded in the sand where they need to
    force the casting to solidify first. In order to get sound castings.

    If the comment were relating to clear style of explaining, I will have
    that...

    ----------------

    It was.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)