GBR files

[beikeland]

If I posted the complete projected, could you convert them to .pdf?

Sure, shouldn't take too long. But you can download the free software and try it yourself if you want to.

As I said, they are not to scale, somehow despite the files saying absolute inches and they are imported as such, and the on screen measurements work okay, the prints come out much too large, and using a print ratio smaller than 1:1 crashed the software. Maybe you can find a way to figure it out, or just turn the .pdfs into raster at a DPI that gives you a good fit with known component sizes.

[beikeland]

I'll try to get some screenshots and document the conversion - could be useful for others who find the thread later.

that would be most excellent

So start off by importing the aperture file. You need to manually select the type for this when importing.

If you have reason to change the scale to mm/inches/mills/whatnot do so in the next dialogue. The defaults were okay here.

Then import the layers. Make sure the aperture file is selected as tool table, it should be by default if its imported first.

Set colours and types of the layer in the next dialogue to make things easier.

After all files are loaded you should have something like this


Then just print to cute pdf writer or other pdf output device. Take care to check the box for  "Print all data in black" and for the bottom layer check the box "print as bottom view" otherwise you'll get a mirrored image.


I'll just attach the screenshots here in case dropbox links fail at some point, but external images makes for better formatting.

[Kunaphil]

Thank you so very much!  I do have standard components that I can get an approximate size.  I have been reading the gc tutorial, but haven't too much time to play with it.

[beikeland]

If you measure the components on the screen in gc prevue and they match the real world, then you could just get the board dimensions from the screen as well.

The resistor is about 10mm across sounds reasonable. There are probably many ways to do measurements, to move the cursor to the starting point, press Z to zero relative measurement, and then move the cursor to the far end and read out the relative distance on the screen (top left-ish)

[Zax]

I added the basic import (still need to understand how the apertures are used) and it looks OK to me, any idea if the sizes are correct?

[beikeland]

still need to understand how the apertures are used

In this case the file describe different "pen" sizes and shapes. Sizes in mills.
So

Code: [Select]

D10   ROUNDED            10.000      10.000 Means you should stroke D10 with a 10 mills round "pen". Number is repeated for X and Y direction to allow elliptic shapes and rectangles in addition to circles and squares.

[Zax]

Thanks, that's what I assumed from looking at the file.

What I am not clear on until I put it in the code and see the result compared to other viewers is whether you rotate the rectangle to the angle of the line segment or not. I think you do but may need some other examples to make sure I have it correct.

I'm only importing a single layer, does that make sense? You wouldn't want to overlay when lasering right?

[beikeland]

Hmm, about the rotation, not sure - it would make sense, but it won't make a big difference. I recon most design using the obsolete format will have ample space between relatively wide tracks - so chances for any short circuits or high impedance tracks due to wider or thinner tracks should be near zero.

But I guess the answer could be fond in history books, did plotters and photoplotters have the capability to rotate the tool? As far as I understand a photoplotter has a wheel of apertures that rotate to change apertures, but unsure if each apertures in turn can rotate.

[Zax]

I guess no rotation, this may be more interesting to code.

[beikeland]

Another reason to think it makes little impact is that it would be very unusual to use a oblong or rectangular aperture for PCBs. In fact I was surprised to see they had used square ones.

Edit: thinking a little bit more about it, square makes sense to mark pin 1 on chips etc. but not really for tracks.

[Kunaphil]

I  found my original email that was sent to me with the Gerber files.  The size of the board should be 6.5" X 3.25".

[Zax]

I think it's looking about right...

[beikeland]

Thats the bottom layer, it needs to mirrored - and the whole thing needs to be inverted (as the laser removes the paint, the etchant then removes the copper, so the unlasered bits will be kept in the final result).

[Zax]

It imports as an image so you can select mirror X or Y and invert as needed. You can also use Sketch Plus to add vector around the design as you mentioned before to provide a nice clean edge. The G-code generation is working great too.

I have matched what is displayed in the commercial programs, but can easily change the defaults.

[beikeland]

Looks good! Think the outline will really help with clean edges indeed!