flatcam

Open full view…

Different heights in same project file?

pauledd2
Thu, 11 Feb 2021 07:54:56 GMT

Hi there I am investigating a problem with my laser pcb exposing workflow where pcb traces have a different witdh in y and x direction. Currently I suspect some kind of non-uniformity in the focused laser point that seems to form some ellipse-kind formed spot instead of a perfect circle which then causes different track sizes in y/x direction. My idea was to create a test pattern in one project file where the laser is very slightly de-focused from its optimal focus point to form a slightly blurred bit more uniform laser point. For that I would like to have four test patterns with four different laser heights. I usually use 40mm Z height for my laser but I would like to have as series of 40, 42, 44 and 46mm height for each test pattern. Is that possible in one project file? [Screenshot_20210211_084311](//muut.com/u/flatcam/s2/:flatcam:pHFd:screenshot_20210211_084311.png.jpg) I created a simple testfile where I imported a kicad gerber, created isolation routing and created the cnc object (gcode). But flatcam of cause can't take care of my four areas to change the t ool height in each area? Is that possible a t all? If not I have to bite the bullet and create a separate project for each of the four areas and create four separate test pcbs...

pauledd2
Thu, 11 Feb 2021 08:24:37 GMT

Alright, I try to edit the gcode file and insert some G00 Z4x..... commands at the positions where the laser changes to the four different areas. Thats a bit of fiddling but not that effortful as creating four different project files / pcbs... :)

Marius Stanciu
Thu, 11 Feb 2021 12:49:22 GMT

Hi, There are lasers out there that have this problem of having the projection as an oval shape. This is due of the fact that the array of semiconductor is rectangular in shape and not a square. A square will give you a circle after the lens so you need to make sure of that before buying the laser. BTW, what power has the laser? As for your problem, there is a certain flexibility in FlatCAM beta which allows you to add different tools with different properties. Further more, you could create one Gerber pattern, duplicate it using the Copy function. Move them into place (by default they are copied on the same area, will occupy the same area), isolate them, change the Geometry parameters for each (including the Travel Z which affects the focus), use the correct preprocessor and then merge together the Geometry objects into one to be able to generate only one GCode file. That's a general idea on what you need to do and not a tutorial to be followed. Therefore you might need to change some things here and there in the flow, depending on the FlatCAM beta version you are using.

pauledd2
Thu, 11 Feb 2021 18:26:12 GMT

Thanks for your hints. My laser is labeled as 0.5 (China)Watts so I assume it has not more that ~350mW real power. And because I expose pcb at 5/255th PWM I assume a total power of about ~5mW. The strange thing is I already got, for my needs, quite good results down to 8mil. [AnycubicDuetLaserPCB](//muut.com/u/flatcam/s3/:flatcam:YI8c:anycubicduetlaserpcb.jpg.jpg) But currently I already struggle at 11mil with ugly non-uniformity: [Ugly](//muut.com/u/flatcam/s3/:flatcam:PXIi:ugly.jpg.jpg) Vertical tracks are quite thicker than horizontal tracks and I cant really figure out what I did different to my previous pcbs... I'll try to experiment a little bit more with focus point...

Marius Stanciu
Fri, 12 Feb 2021 01:20:39 GMT

I don't think I quite understood what you've said.... Are you saying that you can get good results with *most* trace widths down to 8mil but you can't get good results with some trace widths like for example 11mils?

pauledd2
Fri, 12 Feb 2021 07:19:41 GMT

No sry :D. I got good results down to 8mil some time ago but it seems I changed something over the time that now produces just ugly results, no matter if 8 or 11mil. Maybe it's some printer settings but the only thing I can imagine is something with the proper focus point. Or I somehow introduced some "angle" in the laser's collimating lens because it also doesnt look that good manufactured. Its the cheapest stuff you can get...

Marius Stanciu
Fri, 12 Feb 2021 09:33:12 GMT

In that case, you can make a test by first drawing a couple of test lines (one horizontal and one vertical line) and then, if possible, try to rotate the laser body by 90° in the mount and redo them on a new sensitized PCB. It will help if you describe what exactly is your work flow. How did you mounted the laser on the 3D printer carriage? Your 3d printer is a Cartesian design isn't it?

pauledd2
Fri, 12 Feb 2021 09:41:21 GMT

I made a video some time ago: https://www.youtube.com/watch?v=fE-qiESHCZU Well rotating the laser is a bit difficult because I made special brackets that the laser can slide in. I could dismantle the whole laser module and rotate the diode itself because its a cylindrical shaft. I might consider that but what else than the laser should it be...

Marius Stanciu
Fri, 12 Feb 2021 09:51:12 GMT

Some other people reported a while back that there is a thickness difference between the vertical and horizontal (and maybe oblique) traces when they do engraving with a mill bit. I don't exclude the possibility that something changed in the Shapely module used by FlatCAM to create the trace isolations, maybe a rounding error or who knows what ...

pauledd2
Fri, 12 Feb 2021 09:53:39 GMT

Oh okay, but shouldn't that rounding error be visible in the flatcam plots? Or would the error only go into the gcode file?

Marius Stanciu
Fri, 12 Feb 2021 10:20:21 GMT

Perhaps, right now is just something that could be investigated. I think that is most likely that the problem is on the laser side but there is not much information right now to sway one way or the other. I have one of those laser heads that you have and I am waiting also for 5.5W Chinese laser head to arrive (in customs). Maybe, if I can find the time I will play with them too and see if I have any issues. The 5.5W laser should be able to take cured paint away from the pads therefore allowing the usage of paint as soldermask.

pauledd2
Fri, 12 Feb 2021 11:48:33 GMT

Alright, ill first try to optimize my focusing routine... What kind of paint do you have in mind? 5W is a lot of power :D I can cut mylar with my 0.5W laser in two passes ;)

Marius Stanciu
Fri, 12 Feb 2021 13:08:33 GMT

Some eBay sellers from China sell UV curable paint and even if the laser is UV and will cure it by exposure I thought that perhaps going over the threshold may allow it to be burned in the exposed area by going with a slow feedrate/max power on the pad area and fast/low power elsewhere. This way I can do the UV cure elsewhere and burn it through where I want. It was a thought.

pauledd2
Fri, 12 Feb 2021 16:19:42 GMT

> @Marius Stanciu > isolate them, change the Geometry parameters for each (including the Travel Z which affects the focus), use the correct preprocessor and then How can I change the travelz value on a geometry object? I only see travelz in the preferences as global setting. I tried to right-click on the geometry object in the project-tree and choose properties but those fields are write protected and cant be changed...

Marius Stanciu
Fri, 12 Feb 2021 16:49:06 GMT

In the Geometry Object Property/Selected tab you have a Travel Z parameter that can be changed.

pauledd2
Fri, 12 Feb 2021 16:59:01 GMT

Hmm... [Peek 2021-02-12 17-57](//muut.com/u/flatcam/s1/:flatcam:V8Zn:peek202102121757.gif.jpg) Cant edit the numbers

Marius Stanciu
Fri, 12 Feb 2021 19:00:37 GMT

Sorry Paul, unless you start talking about what you are doing and how you do it, I can't pull words from you. If it's hard to write (different language) maybe you could do a screen capture video.

pauledd2
Fri, 12 Feb 2021 19:41:08 GMT

Sorry, Mutt converted my gif into a static image... Here is a new capture: https://p-bg.de/pics/cantEdit.gif By the way I think I solved my focus problem. I invested much more time and a magnifying glass to get the laser point as small and round as possible and then did a focus test pcb. At 41mm laser height I get good results again and I think I cant get it better. [FokusTest40mm](//muut.com/u/flatcam/s3/:flatcam:2aFB:fokustest40mm.jpg.jpg)

Marius Stanciu
Fri, 12 Feb 2021 20:17:15 GMT

Ok I understand now what you are saying. The idea is that the Properties Tool and also the Properties section inside the.... Properties Tab (I know it's confusing I will try to change some of the wordings in the next version) are just reports, therefore Read Only. They are not meant to be changed. And since you are using a certain Preprocessor file, it will remove some parameters due of the specification of that Preprocessor. I think I've added one preprocessor that allows changing of the Z Travel parameter when using lasers but it may be available in the next version (whenever it will become ready). Most of the other preprocessors disable the Z Travel parameter. Regarding of your results in UV exposure with the laser, I think that better results can be obtained by adjusting (lowering) the feedrate and/or increasing the power of the laser. If the exposure time is greater then the exposed photosensitive material changes the properties faster creating better shape s. You may want to make a few passes too, you can set the overlap in Isolation Tool of 99.9999% which amount to almost complete overlap.

pauledd2
Fri, 12 Feb 2021 20:42:55 GMT

Okay, that explains my desperate search for the Travel Z parameter :D I dont want to lower the feedrate further than 400mm/min because it already takes about one hour for the laser to expose 70x90mm pcb board. And I wont go higher in laser power because at 8/255th PWM I observed some kind of "splashing" of the photosensitive layer. It seems to get burned away rather than just exposed. This leaves some kind of dust besides the exposed tracks which can disturb the developing process with NaOH... I thing I am already in the perfect balanced zone of speed/laserpower and withing the mechanical constrains of my 3D Printer. I also dont want to go much higher in speed than 600mm/s because that introduces another problem: vibration that can be seen on tracks with hard angles. The printer is simply not perfect for that kind of task. I possibly would get much better results with a better laser (better collimation, lens quallity) and with a smaller laser (less mass that has to be moved)... I try my big pcb tomorrow and I will see if I get a better result :) thanks for now.

pauledd2
Mon, 15 Feb 2021 11:22:07 GMT

> @Marius Stanciu > I don't exclude the possibility that something changed in the Shapely module used by FlatCAM to create the trace isolations, maybe a rounding error or who knows what … Just as a side note: I checked the gcode files produced from flatcam in an online gcode viewer and that shows all tracks have correct width no matter which direction X or Y. So I would rule out flatcam itself, at least in my case.