Kicad PLT non fatal errors

[Stonemull]

Hey Zax, import PLT not working with kicad plots.

I only tried one and get a continuous pop up of 'non fatal error' till I use task manager to kill the T2Laser process.
File i tried attached, it looks fairly basic at first glance so not sure where the error lies.


While I have your attention, you mention in the gerber import that you can load drill files, I was hoping I could use them to add pad holes to the copper pour as there is no option to do that with kicad as holes are another layer, but it didn't appear to image anything, is it just for boring holes perhaps ?

[Zax]

I looked at the PLT file from kicad, it's slightly different to what I support from Corel so I will need to make some changes to deal with their format.

I'm not familiar with kicad, does it export DXF files?

T2Laser will load aperture (.APT) and/or drill files (.DRL) in the same directory with either the same name or prefix as the Gerber file. These are layered so the drill file will be subtracted from the layout.

[Stonemull]

I looked at the PLT file from kicad, it's slightly different to what I support from Corel so I will need to make some changes to deal with their format.
I'm not familiar with kicad, does it export DXF files?
T2Laser will load aperture (.APT) and/or drill files (.DRL) in the same directory with either the same name or prefix as the Gerber file. These are layered so the drill file will be subtracted from the layout.

Kicad output options are ..

Gerber
Postscript
SVG
DXF
HPGL
PDF

It is free and open source and could be used to produce artwork in a pinch. The only one I can use currently is the Gerber in raster format but I get no pad holes.

Ok had some success after renaming the .drl file to the same as the copper (is there a way to load a differently named drill file ?) , I now have holes in the vias but not the through hole pads, the holes in the vias are a little offset from the true centres, added a screenshot and also one from Gerbview within kicad for comparison for the front layer and drill file here..

https://www.dropbox.com/s/x9yxszobaj4kksy/Screenshot%202017-04-06%2013.36.26.png?dl=0
https://www.dropbox.com/s/vspz7bbgshylqd5/Screenshot%202017-04-06%2013.38.53.png?dl=0




I have attached the useable outputs of the same board and the drill file so you have a reference to the files I am currently using.

Ok installed 1.4, I downloaded from cnet originally and put up with the stupidly loud videos.
(I am happy to trial beta's for you if you ever need another tester...)

I found the hatch options but I cannot seem to get a hatch on either a DXF or an autotraced GBR yet.
Is there a button to press or is it automatic once enabled ?

The DXF import looks damn sexy on screen, I want it..

Also, there is an asterisk next to the 'Gerber Options*'   .. I am not sure why and I can't shake the feeling I am missing something... any reason ?

[Zax]

The aperture and drill files must be the same name (or prefix) and in the same folder.

I did notice the format of the drill sizes is slightly different, but I can easily fix that. See attached, it appears identical to the PDF (or at least to me).

The DXF looks good too but I don't think you can use that without modifying.

I'm also not sure hatch is going to work on such small detail, you really need to raster at 0.1 or 0.15 with a slightly de-focused beam - that would give you be best removal.

I will fix the PLT and drill size format in the next release. In the meantime I added a space between the drill designation and size.

[Stonemull]

Roger, will try it with the space.
Yeh tried the hatch and not really a success, a few hatchlines cross the outline boundarys, its just not going to work.
Not fussed about the HPGL, PLT so don't bother fixing if it is just for me, as long as I can get the drill marks with gbr's then I will use that method and raster till a good vector solution comes up.

I mentioned inthe updates thread you have some sort of underflow bug with the diagonal hatch sending the laser to 9999,9999 at times.

ok, tried the space and it works, awesome, it does have an issue where the via drills (T1 0.5) are displaying as a larger T2 or T3 but I can work aroubd that, I just want drill marks really so I might just drill everything at 0.5mm for the moment.

I just scanned a film under glass at 0.1mm feed 1000 and laser 30 and the quality is amazing despite being a little over exposed, so good that I am attempting something new, sticking an exposed film onto the copper. Drying it at atm, can't see it working, I think the mylar top film will rip half the resist off the pcb as it is not a full sheet of film.

[Zax]

I will "fix" both the GBR and PLT, as they were developed from just a few examples I downloaded and people sent me. It works perfectly on those but of course your kicad version is slightly different (no big deal to correct it).

The hatch thing will develop over time, your small images / details are a challenge especially when tracing as it doesn't always connect everything up as you think it should and then the hatch lines "escape" and run for the hills!

[Stonemull]

I can send you as many gerbers and plt's as you want
Those few examples will likely get you all the functionality you need though.
The raster is producing good quality and I will likely stick with barring some sort of pure vector fill routine, which I can't see happening
I think when you combine vector movement with a raster there will be misaligment due to belt tensions on these machines leading to a fill being offset from the outline.

I design electronics and write embedded code btw (mostly PIC), I have had plenty of things run for the hills over the years.

[Koskee]

I also use KiCad, and have been working towards being able to burn PCBs with my machine, so that nobody feels like any time is being wasted on this area..

@Stonemull - if you wouldn't mind explaining your burn in a bit of detail, I'd be interested to know how you're going aboit it.. Are you using presensitized boards or something else? And how have you been dealing with the beam divergence lighting up what it's not supposed to (I'm guessing that's what you're referring to when you talk about using the mylar tape over top?)

[Stonemull]

I have a thread in the projects section with photos and details. Still a work in progress though I am using my first board.
I am using Puretch film at the moment which has 3 layers, the top one being mylar which is removed after exposing.
I am using a beam power of between 25-30 and bleed does not seem to be a big issue.

[Koskee]

Sweet, i'll check it out. Thanks

[Stonemull]

Just tried a HPGL file and damn, it looks like really good I think, it is still doing pads and I only put a small piece of film under glass to give it a quick try but from what I can see it is working really well.
I told kicad that it had a 0.05mm pen hence the jiggling around, but even the 0.1mm lines for a border look damn good, maybe vector is the way to go..

It has not filled in the ground plane unfortunately from looking at the initial open, maybe there is a kicad option that can help there somewhere, I will see if I can pin down what is happening later.
Also kicad is not using the auxilliary axis (option is greyed out) so I have a massively oversized offset on the bottom of the board which will make double sided tricky to attempt.

looks like I have a few hours more of mucking around tomorrow.

Well done zaks, damn good start on hpgl for sure.


i did a little video.. https://youtu.be/iRnPdyfp5IM

[Zax]

Maybe you could add hatching to the ground plane, without seeing it I don't know if you can do it in T2Laser or not but it's an option.

[Stonemull]

Or I could just not use a ground plane if I want to plot it for some reason, it is mostly good practice and with my old ferric chloride etching technique saves on etchant costs, less copper to remove.

I have also changed to a hydrochloric acid, hydrogen peroxide method that theoretically lasts forever without depleting the chemistry.

So it may well be the best method to use hpgl for high resolution prototype boards.
I can see the method putting down accurate tracks the width of the laser spot, which no other home laser method can currently can do as far as my googling has shown.
it is also missing vias so I suspect the circular elements are being discarded or coded incorrectly currently.

As I only did a partial ilm ladt night and I was not entirely sure where it was going to plot, I missed the interesting areas on the film, photo shows the plotted film just below the same area on a overexposed raster I did earlier in the week.
(blue smear on the left of the pads is on the plastic card underneath, not part of the plot)

You can see the ease with which it managed the clearances required though.
The ground plane seems to have been plotted with just a very fine outline and no fill, it pretty much fell off during development so not visible on the film in photo.

edit: the missing tracks are interesting, I thought maybe they were plotted before I put the film down as I started it tipo see the motion then decided to put a film under it, paused T2 then hit continue but selected 'no' so it should have started anew, not sure if it did.

[Zax]

The raster still looks good to me, but I guess it's the details that don't show up on the photo. I can definitely see vector being the better quality for the high res. though.

[ta1dr]

BTW :  I tried the dry film but I failed.  I thing  laser leakage problem

https://www.youtube.com/watch?v=ZIrT5eWAw2c&t=390s