¿ªÔÆÌåÓý

ctrl + shift + ? for shortcuts
© 2025 Groups.io

Re: #pcbgcode #pcbgcode

 

I mean here:?


Re: #pcbgcode #pcbgcode

 

Oh, and almost forgot; I did solve this with a modified version of PCB-Gcode, here:


On Sunday, April 25, 2021, 11:39:01 PM EDT, mnoyes1@... <mnoyes1@...> wrote:


Thanks, I was beginning to think I was more stupid than I am. Why wouldn't this obvious issue show up in even the simplest testing before release? I even tried to put my screen into Portrait mode and then I could see the form looking sideways(HA!) but trying to use it is impossible. Seems to me there has to be a simple explanation.


PCB-GCODE Version 3.6.2.4 Vector Text Issues #mill

 

Hi All,

I have been using PCB-GCODE 3.6.2.4 with vector text on the milling layer I get some very strange effects.

Delving into the code I noticed that this version of PCB-GCODE the way arcs are drawn has changed. I assume this is to do with arcs on the tracking layers.

Unfortunately this has effected the way text is drawn.

The fix I was able to make was to revert the arcs drawing for vector text back to the old method but with small changes to fix an issue with the letter U

Change I made was to change the code in pcb-gcode.ulp from the directory ulps/pcb-code

The from the line 'case ST_ARC_END:' to the line break change the code to the following:
user_arc_end(rx1, ry1, rx2, ry2);
real cx = rx2;
real cy = ry2;
real end_x = rx1;
real end_y = ry1;
?
if (1 /* USE_IJ_RELATIVE */) {
cx = rx2 - rx1;
cy = ry2 - ry1;
}
?
rz(DEFAULT_Z_UP);
rxy(end_x, end_y);
fzr(z_down_or_radius, FEED_RATE_MILL_Z);
if (g_side == TOP || MIRROR_BOTTOM == YES) {
out(frrrr(CIRCLE_TOP, m_arc_begin_x, m_arc_begin_y, cx, cy));
}
else {
out(frrrr(CIRCLE_BOTTOM, m_arc_begin_x, m_arc_begin_y, cx, cy));
}
rz(DEFAULT_Z_UP);
rxy(m_arc_begin_x, m_arc_begin_y);
break;


Add the following to the end of your .pp file which is held in the directory profiles.
string CIRCLE_TOP? ? ?= ARC_CW + MOVE_XY + IJ_FORMAT + EOL;
string CIRCLE_BOTTOM? = ARC_CCW + MOVE_XY + IJ_FORMAT + EOL;

These changes fix the text engraving on the milling layer.

MPC
?


Re: Milling Circles and Milling Components milling data #pcbgcode

 

On Apr 26, 2021, at 5:42 AM, support@... wrote:

Is there a way to mill circles without building them with arcs?
Unfortunately, no. There are lots of places you can add your own code, if you¡¯d like to implement this yourself.

is it possible to mill shapes of different sizes with a smaller tool, (same tool for all holes )?
Unfortunately, no. But as above, you might implement this yourself.

is is possible somehow to mill the data placed in the milling layer of a component, e.g. a square hole for a reverse placed SMD LED ?
Lines/wires on the milling layer (46) should be output in the milling files. Turn on Generate milling to create these files.

Regards,
JJ


Re: #pcbgcode #pcbgcode

 

¿ªÔÆÌåÓý

This might help:


Regards,
JJ



On Apr 25, 2021, at 10:29 PM, mnoyes1@... wrote:

Thanks, I was beginning to think I was more stupid than I am. Why wouldn't this obvious issue show up in even the simplest testing before release? I even tried to put my screen into Portrait mode and then I could see the form looking sideways(HA!) but trying to use it is impossible. Seems to me there has to be a simple explanation.


Milling Circles and Milling Components milling data #pcbgcode

 

Is there a way to mill circles without building them with arcs?
is it possible to mill shapes of different sizes with a smaller tool, (same tool for all holes )?
is is possible somehow to mill the data placed in the milling layer of a component, e.g. a square hole for a reverse placed SMD LED ?


Re: #pcbgcode #pcbgcode

 

¿ªÔÆÌåÓý

Try putting the cursor in the window, press and hold the ALT key, left click and hold and position the window with the mouse, release both keys and the window should stay in that position.


On Apr 25, 2021, at 21:39, mnoyes1@... wrote:

?Thanks, I was beginning to think I was more stupid than I am. Why wouldn't this obvious issue show up in even the simplest testing before release? I even tried to put my screen into Portrait mode and then I could see the form looking sideways(HA!) but trying to use it is impossible. Seems to me there has to be a simple explanation.


Re: #pcbgcode #pcbgcode

 

Thanks, I was beginning to think I was more stupid than I am. Why wouldn't this obvious issue show up in even the simplest testing before release? I even tried to put my screen into Portrait mode and then I could see the form looking sideways(HA!) but trying to use it is impossible. Seems to me there has to be a simple explanation.


Re: #pcbgcode #pcbgcode

 

Happens in Windows 10 too. By hiding the dock I was able to just get to the area that designates whether metric or imperial. Still isn't great though.

On Sunday, April 25, 2021, 08:06:40 PM EDT, mnoyes1@... <mnoyes1@...> wrote:


Help, just downloaded pcb-gcode for Eagle and got the directory set up ok but I'm running a Windows8.1 PC and the pcb-gcode Setup is totally off scale for my 1260x768 screen. Can't get to most of the settings and I don't see any way to scale down the form to make it wholly visible. I tried clicking and dragging with the arrows and can only change the horizontal scale. Vertical scale is off screen and I can't budge it. Tried that little icon in the upper left corner and picked "size" but it does nothing. Is there a configuration file hidden somewhere for screen size?


#pcbgcode #pcbgcode

 

Help, just downloaded pcb-gcode for Eagle and got the directory set up ok but I'm running a Windows8.1 PC and the pcb-gcode Setup is totally off scale for my 1260x768 screen. Can't get to most of the settings and I don't see any way to scale down the form to make it wholly visible. I tried clicking and dragging with the arrows and can only change the horizontal scale. Vertical scale is off screen and I can't budge it. Tried that little icon in the upper left corner and picked "size" but it does nothing. Is there a configuration file hidden somewhere for screen size?


Re: Five paths to cut a square? #eagle #gcode

 

¿ªÔÆÌåÓý

Hi Martin,

I believe that is a modified version, but haven't been able to figure out who did it. I've search my drives, and don't see any version of the code or pertinent files that mention 'stencil'.

It looks like in source/pcb-gcode.h the variable PH_TOP_STENCIL is defined as layer 5, and the bottom as layer 10, if that's any help to you.

There are quite a few messages on the ?about 'stencil', which might be of some help to you.

Regards,
JJ

On Apr 1, 2021, at 1:20 PM, Martin Rowe <mprowe@...> wrote:

Hi JJ,

Ever more questions - sorry.

On 01/04/2021 15:36, John Johnson wrote:
Hi M,

Pcb-gcode only uses the top, bottom, mill, and text layers.?
Does that imply that the stencil output is deduced from the copper layers?
This might be why you're seeing a bounding square - it's actually on one of the other layers.
Sorry. What is "on one of the other layers"?
Is it possible you're using someone's fork of pcb-gcode that processes other layers?
It looks like yours? From?/g/pcbgcode/files/2Software.
<pjehlnllhdlhpebp.png>

Regards,
JJ

On Mar 23, 2021, at 11:52 AM,?mprowe@...?wrote:

Thank you JJ,

Can you also clarify another puzzle for me? Again in the Solder Stencil Making area.
From limited testing, it seems that whatever shape the Eagle PCB pad is (Rectangle, Octagonal, Round or Oval), the pcb-gcode "top stencil" output results in a bounding square? Is this correct?

Additionally, which Eagle layer is pcb-gcode reading to generate its stencil output? As an experiment, I placed a polygon on layer 21 - Top Cream and the pcb-gcode output is effectively empty??

Regards, M



--?
mprowe


Re: Five paths to cut a square? #eagle #gcode

 

¿ªÔÆÌåÓý

Hi JJ,

Ever more questions - sorry.

On 01/04/2021 15:36, John Johnson wrote:
Hi M,

Pcb-gcode only uses the top, bottom, mill, and text layers.
Does that imply that the stencil output is deduced from the copper layers?
This might be why you're seeing a bounding square - it's actually on one of the other layers.
Sorry. What is "on one of the other layers"?
Is it possible you're using someone's fork of pcb-gcode that processes other layers?
It looks like yours? From /g/pcbgcode/files/2Software.

Regards,
JJ

On Mar 23, 2021, at 11:52 AM, mprowe@... wrote:

Thank you JJ,

Can you also clarify another puzzle for me? Again in the Solder Stencil Making area.
From limited testing, it seems that whatever shape the Eagle PCB pad is (Rectangle, Octagonal, Round or Oval), the pcb-gcode "top stencil" output results in a bounding square? Is this correct?

Additionally, which Eagle layer is pcb-gcode reading to generate its stencil output? As an experiment, I placed a polygon on layer 21 - Top Cream and the pcb-gcode output is effectively empty?

Regards, M



--
mprowe


Re: Five paths to cut a square? #eagle #gcode

 

¿ªÔÆÌåÓý

Hi M,

Pcb-gcode only uses the top, bottom, mill, and text layers. This might be why you're seeing a bounding square - it's actually on one of the other layers.
Is it possible you're using someone's fork of pcb-gcode that processes other layers?

Regards,
JJ

On Mar 23, 2021, at 11:52 AM, mprowe@... wrote:

Thank you JJ,

Can you also clarify another puzzle for me? Again in the Solder Stencil Making area.
From limited testing, it seems that whatever shape the Eagle PCB pad is (Rectangle, Octagonal, Round or Oval), the pcb-gcode "top stencil" output results in a bounding square? Is this correct?

Additionally, which Eagle layer is pcb-gcode reading to generate its stencil output? As an experiment, I placed a polygon on layer 21 - Top Cream and the pcb-gcode output is effectively empty?

Regards, M


Re: Five paths to cut a square? #eagle #gcode

 

Thank you JJ,

Can you also clarify another puzzle for me? Again in the Solder Stencil Making area.
From limited testing, it seems that whatever shape the Eagle PCB pad is (Rectangle, Octagonal, Round or Oval), the pcb-gcode "top stencil" output results in a bounding square? Is this correct?

Additionally, which Eagle layer is pcb-gcode reading to generate its stencil output? As an experiment, I placed a polygon on layer 21 - Top Cream and the pcb-gcode output is effectively empty?

Regards, M


Re: Five paths to cut a square? #eagle #gcode

 

¿ªÔÆÌåÓý

Ah, I see.?

I¡¯m not sure why that happens. Pcb-gcode just processes the line segments that Eagle gives to it, so it looks like Eagle saw the need to pass that extra line segment.?

Regards,
JJ


On Mar 23, 2021, at 7:50 AM, mprowe@... wrote:

?... or maybe I'm using the wrong word with "path"? Let me try again...

Line 32 "G00 X19.0000 Y10.5000 " gets the tool to the North-West corner of the square.
Line 34 "G01 X21.0000 Y10.5000 F360.00 " moves the tool to the North_East corner
Line 35 "G01 X21.0000 Y9.5000 " moves the the South-East corner
Line 36 "X19.0000" moves to the South-West corner (modal - G01 still in effect)
and Line 37 "Y10.5000" completes the 4th side of the square by moving back to the North-West corner.

So why is Line 38 "X21.0000" moving the tool to the North-East again?

Regards, M


Re: Five paths to cut a square? #eagle #gcode

 

... or maybe I'm using the wrong word with "path"? Let me try again...

Line 32 "G00 X19.0000 Y10.5000 " gets the tool to the North-West corner of the square.
Line 34 "G01 X21.0000 Y10.5000 F360.00 " moves the tool to the North_East corner
Line 35 "G01 X21.0000 Y9.5000 " moves the the South-East corner
Line 36 "X19.0000" moves to the South-West corner (modal - G01 still in effect)
and Line 37 "Y10.5000" completes the 4th side of the square by moving back to the North-West corner.

So why is Line 38 "X21.0000" moving the tool to the North-East again?

Regards, M


Re: Five paths to cut a square? #eagle #gcode

 

¿ªÔÆÌåÓý

Maybe I¡¯m misunderstanding.
In the file you attached, I only see that it is cutting out the square with a single pass.


On Mar 23, 2021, at 4:09 AM, mprowe@... wrote:

?Setup: OS OpenSUSE 15.2, Eagle 9.6.2, PCB-GCODE 3.6.2.4

Workflow: New Eagle Project with a single component with a single 2.00x1.00mm SMD pad centred at 20x10mm on board.
Run the pcb-gcode ULP and remove all the "generate" options except Stencil (I'm working towards a laser cut solder stencii - hopefully!l).
I have attached the resultant gcode file.

My question is this: Why is the tool path making 5 passes to cut a 4-sided square? Should it be doing this?

Regards, M
<dimensions.top.stencil.gcode>


Five paths to cut a square? #eagle #gcode

 

Setup: OS OpenSUSE 15.2, Eagle 9.6.2, PCB-GCODE 3.6.2.4

Workflow: New Eagle Project with a single component with a single 2.00x1.00mm SMD pad centred at 20x10mm on board.
Run the pcb-gcode ULP and remove all the "generate" options except Stencil (I'm working towards a laser cut solder stencii - hopefully!l).
I have attached the resultant gcode file.

My question is this: Why is the tool path making 5 passes to cut a 4-sided square? Should it be doing this?

Regards, M


Re: Levelling before machining - OR not #mill

 

OK great. Yeah the area is not too big. I'll give it a bash. I have a piece of acrylic I'll use as a mount surface and clamp that to the table?


Re: Levelling before machining - OR not #mill

 

Yeah. Basically Smoothstepper has the same type of port pin. Thanks

On Sunday, January 3, 2021, 06:01:14 PM EST, Art Eckstein <art.eckstein@...> wrote:


I am not sure of your question.? Not having a smoothstepper, I have no idea of how it would be connected there.
I use TurboCnc and there is a port pin dedicated to the probe. It has a pullup resistor on it and when shorted to ground, completes the circuit.

I move to the XY coordinate that I want to probe and issue a G31 Z-.5 and it will lower the probe until it contacts the surface or reaches -.5 whichever occurs first. The probe coordiates (X,Y,Z) are then added to the file for later use.

Has this helped any???

Country

?
At 1/3/2021 05:14 PM, keith3125 via groups.io wrote:

I have a smoothstepper with a probe i/o so I'll try something like that but where is your circuit attached?

On Sunday, January 3, 2021, 05:07:24 PM EST, Art Eckstein <art.eckstein@...> wrote:


My "Probe" for this purpose is my 'engraving' tool.? I chuck it up in the spindle and alligator clip one side of the probe cable to it and the other side to the pcb stock *which is insulated from the frame as it is held in a piece of derlin.
When the 'probe' touches the board, it completes the electrical circuit.

When originally testing this method, I kinda crumpled up a piece of aluminum foil and proceded to probe the surface to see what would happen.? To my surprise, the foil never moved except in one place that it had a spot of epoxy glue on it.

Hope this helps.

Country


At 1/3/2021 04:46 PM, keith3125 via groups.io wrote:
Art, what do you use for a probe?

On Sunday, January 3, 2021, 11:40:52 AM EST, Art Eckstein <art.eckstein@...> wrote:


When you say "because the software and controller don't cater for
that" do you mean you can't do probing or what??

I find leveling software such as <> in
valuable because even though I use a vacuum chuck that has been
milled flat, the cheap boards that I get from the usual suspect are
still quite 'wavy' and again as you say it makes a difference between
success and failure.

Country



At 1/3/2021 11:12 AM, falconeti6969 wrote:
>Hey guys,
>
>So I have basically shyed away from milling boards and just buy perf
>board and this is why;
>I can't do a levelling pass with my machine because the software and
>controller don't cater for that. Its an industrial 1200x1200 machine
>I bought in 2012. I have toyed with moving to mach3 BUT no real
>reason to for any other reason and it just works.
>So I am getting more and more to the point of milling my boards and
>wonder what do you guys, those with my dilemma, what do you do? At
>the margins the boards are machined at .2 deviation can mean the
>difference between success and failure