Cool, thanks a lot! Good to know. Over a short enough distance, any
curved line looks straight. :-)
Simon
On 10/26/2016 12:23 PM, Andrew C. wrote:
Simon,
G02 and G03, the codes for clockwise and counterclockwise respectively,
have no way to program splines within them. Furthermore, I know of no
style of controller that allows for programming a spline in a single
line of code.
Of course, splines in 2D and 3D, even above that using more axes, are
used all the time in making parts. So how's it done?
As many discrete lines of small linear movements or small connecting
arcs. This kind of shape though is usually programmed by 3rd party
software, things called CAM software, things like MasterCAM. 3d parts
files are created and the CAM software generates the code for the spline
in many discrete lines.
The thing to know about G03 and G02 is that the curves made with them
can be defined using i and j variables, which is more precise and
universal across machines, or you can use r variable too for radii.
There's a couple ways to do it. i and j allow a bit more finesse.
Does that help any?
-Andrew
On Wednesday, October 26, 2016, Simon Heath <icefoxen@xxxxxxxxx
<mailto:icefoxen@xxxxxxxxx>> wrote:
Thanks for this article, it's very cool! Just one question...
With the "clockwise motion" and "counterclockwise motion" codes, how is
the curve defined? Does it have to be a circular curve, or can you make
it do an ellipse or a spline or some other kind of curve?
Interesting to read this, 'cause from a computer science point of view,
this is a state machine... naturally, since you're setting the state of
the machine. The main difference I see in G code is it has to say how
to get from one state to another... computers are discrete, so when you
switch states a computer just does it. The real world is continuous, so
when you tell a machine to go from one state to another it has to
actually move through all the space between them. :-P
State machines are often used as controllers for mechanical systems,
from vending machines to robot arms, so I'm not super surprised, but
they also get used in scanning text and other such things. Regular
expressions are a state machine for matching patterns in text, for
instance.
Simon
On 10/25/2016 06:07 PM, Andrew C. wrote:
> Ryan-
>
> Depends on how much complexity you want. I havent done much of
this type
> of gcode- as this is often Mcodes (condition codes). they're basically
> the same types of things as codes, but it's a letter m instead of g.
> they change the way parts of the machine operate, and some are
> conditional triggers as you mentioned though these are often not
used in
> general machining.
>
> There is a 3rd type of machine code that hasnt been mentioned that is
> used for what you want to do: macro code.
>
> I can read some of it, but it's not normally ever used in
machining. it
> is the C code behind subroutines, basic if then nor not type
> programming. all conditional statements. I see it scroll by in the
> hundreds of lines running my tool probe subroutines. I cannot code
with
> it, as it is totally linked to variable output and input parameters of
> the machines themselves in binary.
>
> My guess is that many of you could fluently read and write this
type of
> code as I call it something different, but it looks like simple C
to me.
> I just cant code in C yet, but I get the logic of how the
statements are
> read by the machine, I just cant follow what its doing as there are
> hundreds of statements all interrelated calling upon each other.
>
> example code I can get you, or look on Haas website for omp-40 spindle
> probe routine code.
>
> This is the kind of code that looks for binary state of switches
in the
> machine, if they are 0 or 1, on or off, and does things with that info
> based on referencing those parameters. a lot of square brackets!
>
> if you want to do something complex like these conditional
triggers with
> machine readable code in a normal cnc machine, youd probably do them
> with a lot of lines of these codes, or interlinked these with the
few M
> codes that do some of this too.
>
> Don't know if that helps or confuses.
>
> -Andrew
>
>
> On Tuesday, October 25, 2016, Andrew C. <soshumasamune@xxxxxxxxx
<javascript:;>
> <mailto:soshumasamune@xxxxxxxxx ;<javascript:;>>> wrote:
>
>
>
> On Tuesday, October 25, 2016, Andrew C.
<soshumasamune@xxxxxxxxx <javascript:;>
> <javascript:_e(%7B%7D,'cvml','soshumasamune@xxxxxxxxx
<javascript:;>');>> wrote:
>
>
> Its honestly not difficult guys, what's difficult is selecting
> safe tools to cut with, and spinning them at correct speed and
> moving them at correct rate (feed) at a safe depth through
> material. That's the knowledge part behind machining, and that
> is what takes more time to learn.
>
>
> Wow. Ok, sorry. rereading that made me finally realize how I feel
> when the coders of C and pc code make me feel when they say
the same
> thing to me- "it's not hard". A bit stupid, as I don't have an
> understanding of the framework yet.
>
> What I did mean was genuinely, machine code for basic parts really
> doesn't take long to learn. Even with no experience at all, I
could
> teach someone to write a working simple program in an
afternoon. not
> long to show codes, a little time to show how to combine them to
> make a working program you could make a simple part with.
>
> For simple parts made of just linear moves at different
depths, you
> could learn to program it in an afternoon.
>
> After a week, if you're paying attention, programming to do
> threadmilling even isn't difficult to pick up.
>
> I'm happy to show people how to do it if anyone's ever interested.
> Maybe this is another class though?
>
> -Andrew
>
