re: HPGL vs. Gcode was Re: re:Compucut ?

Rab wrote:

>Could anyone explain the advantages of G-Code over HPGL ?
>My current machine, a Roland uses HPGL to cut in 3D very
>successfully, although I have nothing to compare it with, it
>contours well enough for me. It seems that compucut would work
>similarly ?

>I don't understand the unipolar v. bipolar thing yet, is there a big
>difference ?

Rab,

HPGL (Hewlett Packard Graphics Language) was designed by HP as a
language to drive vector plotters to output computer drafted images.
It is a 2D language BY DESIGN. The SP (select Pen) is often used
to 'select' height in CNC programs which input HPGL. This usage makes
it a 2-1/2D Language. The third axis 'height' can be 'selected', much
as a pen "would have been" selected, so the drawing would be made in
differing colors.

There have been many attempts over the years by many manufacturers,
to expand HPGL to 3D. This is/was largely due to the fact that nearly
EVERY drafting program had a standard HPGL output implemented(usually
the HP7475 'flavor'). This was at a time when many other drafting
exchange standards (IGES, DXF,etc.) were either poorly implemented,
poorly defined, or evolving.

The usual way this is done is to "ADD" commands to the basic HPGL
language. Large companies like Galil and Roland do this. So do small
companies like Digital Design Technologies. (funny thought; Big
company=short name, small co.=long name...)

Another way is to "re-interpret" the existing commands, as shown
above with the SP command.

A third approach is to "add functionality" to EXISTING HPGL commands.
This is what Compucut does by allowing a third coordinate in the PA
(Plot Absolute) command. Galil, Roland and others also make use of
this technique.

The problem with the first technique is that SOMETHING has to "write"
the new commands! For Roland and Galil, the answer is easy. They
create and sell a complete system (proprietary) which takes you all
the way through the process. The second way is the most compatible
and seamless, but may not provide "enough" functionality. For
example, the SP 'trick' works really well for a LOT of parts. It
WON'T work very well for a 3d mold! (but it can be done!)The third
approach has the same problem as the first; SOMEONE has to write the
code! But you MAY be able to 'massage'(manipulate) the output file
easily enough to be worth the bother.

Roland has,IMO, probably done the BEST job of implementing HPGL to
3D. Also of supporting the largest number of HPGL commands. Last time
I looked, the 'translator'(drawing to 3d HPGL code) portion of their
system was proprietary and unavailable separately. I've long thought
that if someone could make machines like Roland, at a better price
point, and with better support, they would be rich... The machines
are VERY easy to use, in general, and the multiple processes of 'art
to part' are pretty seamless... Definitely 'plug and play'. However,
they then seem to shoot themselves in the foot by not realising what
they have, and how it "could" be applied....

HPGL is also "resolution limited". Both by the 'size' of the expected
plotter and the expected units (.025mm). The 7475 plotter 'driver' of
most CAD pkgs is a B size (roughly 11x17 inches maximum). The same
sorts of work-arounds as described above are also used here, but be
aware that HPGL can be limiting in resolution and overall part size.

G code was DESIGNED for MACHINE CONTROL in MULTIPLE AXES. Okay, so it
was also used in plotters originally, but it was a three dimensional
language from the start. It is NOT resolution or size limited in any
way (except that the N word originally only had 4 digits, which meant
the longest program would be 9999 'blocks'(lines). This has since
been updated, as well as similar limits imposed on other numerical
parameters of the language(the computer power, or lack of it was the
reason for these limitations).

The only real "down sides" to Gcode, IMO, are 'perception of its'
difficulty', and the HUGE amount of variation, which I tend to
call 'flavors'. I think these two feed each other: "Since its' hard,
I'll just do this my own way", adding a new flavor and making it seem
more difficult. Another thing that feeds the "g code is difficult"
belief is the generally VERY poor texts available. But the 'flavors'
definitely confuse LOTS of people...

Now to your second question:

Unipolar means the current going through the coil in the motor
ALWAYS goes the SAME direction. Uni=one

Bipolar means the current going through the coil in the motor goes
EITHER direction. One direction at a time, but ABLE to go BOTH ways.
BI=two

Unipolar drives are easy to make, as they require only
4 'switches'(transistors). Bipolar require 8 'switches', so they cost
more, and the possibility of a direct electrical short circuit, makes
additional considerations necessary.

There are lots of resources detailing further differences, so I won't
waste bandwidth here, except to state that it is the DESIGN of the
UNI- or BI- polar drive which makes IT better or worse, NOT the fact
that it IS UNI or BI!

Having said that, IN GENERAL, using commercially available drives,
BIPOLAR is better!

Hope this helps.

Ballendo

P.S. Rab, Compucut may be 'just the ticket' for you, but there are
other choices which may result in your being able to get more help
more easily than with Compucut. Realise this has nothing to do with
whether Compucut itself is good or bad... It can obviously (from the
website pics) make some NICE, TECHNICAL parts...

MOST of the people on this list 'speak' G code (and many are
learning). MANY people on this list 'speak' HPGL. SOME of the people
on this list 'speak' Roland. So far, only ONE person on this list has
replied that he USED TO 'speak' Compucut...