Failed! PCB milling not satisfying yet ...

[okalkan]

Hello everyone,
beside the impressing results of successfull milling results as documented in the forum, here's a not yet satisfying result.

Setup: Shapeoko 2 as sold as complete set from inventables.
Engraving Bit: Aptec 30°, 0.1mm
Milling Bit: 0.8mm
Rotary Tool: the one shipped with the Shapoko 2 (speed set to "6")

Software used:
- Target 3001 (circuits) (cutting depth 0.3mm and line thickness 1.0mm and 1.8mm with an aura or 0.5mm)
- hand written G-Code (test pattern)

The PCBs have been placed in a milled bed to get a plain zero level. They have been fixed with double sided tape.
The bit's Z axis is set to zero by letting it fall down to the center of the board.

Result:
- Using to few table seems to cause the rough edges in attempt 2.
- Tracks are milled with a thickness of about 1mm (regardless of cutting depth) and a precision of 0.1mm)
- 0.2mm seems to be good enough to get rid of all the copper (maybe 0.15mm).
- Milling to fast causes loose of precision ("F x" with x <= 30 seems to be ok, faster does not make sense).

Next steps:
The lines engraved are much to big and the precision is poor. I currently do not really have a clue how to reduce the line thickness and improve the accuracy. Perhaps the bit is wrong (I'v read that some of you use 45° or 60° bits). Regarding the results from attempt 2, I will replace the double sided tape because it might work as a spring that causes the board to wobble. I also think of replacing the rotary tool mount to get a fixture point closer to the lower end.

If you have any hints or did some experiments on your own that can help to identify the parameters to adjust for a better results, please don't hesitate to share them (esspecially if you are using a complete shapoko 2 kit without any major changes - such as ACME z-axis, more expensive rotary tool, etc.).

First and second attempt.

2014-04-17 11.23.58_commented.jpg (462.6 KiB) Viewed 2008 times

Third attempt with a test pattern.

2014-04-17 11.29.18_commented.jpg (305.09 KiB) Viewed 2008 times

PCBTestPattern.nc

Test pattern GCode

(10.49 KiB) Downloaded 55 times

By the way: for wood milling the Shapeoko 2 works fine (I didn't try aluminium nor plastic yet).

Best regards,
Ove

[WillAdams]

Edward recently wrote up a bit which we added to the wiki: http://www.shapeoko.com/wiki/index.php/ ... lling_PCBs

[cvoinescu]

My opinion is that you mill too deeply for that type of endmill. Also, the quality of the supplied rotary tool is, erm, variable, so if you got one with larger runout, it may be that too.

[edwardrford]

I'm with Cat, it looks like your depth pass is too aggressive.

[PsyKo]

Hello,

Shapeoko works quite well for PCB milling :
http://www.shapeoko.com/forum/viewtopic.php?f=28&t=975
http://www.shapeoko.com/forum/viewtopic.php?f=28&t=1780

Here I suggest :
- Reduce the pass depth. If needed, do a second pass.
- Check your used endmill. It looks like the second one does a better job.
- Check the rigidity of your machine. Non straight line can be explained by loose excentric nuts, or loose rotary tool hold).
- Going too slow is usually not a good thing. I see you feedrate is F8... That's incredibly slow if running in mm/minute. Try something higher.

Hope this helps.

[okalkan]

Thanks for your answers.

For the next try, I will do:
- use a reduced depth
- try multi pass
- use a higher speed
- use clamps instead of double sided tape

Best regards,
Ove

[twforeman]

You might also want to tweak your steps/mm settings for more accuracy.

I have some info about doing that here: http://timf.anansi-web.com/wp/ferrules- ... -shapeoko/

[okalkan]

Update:
Thanks to Timothy for the inspiring description. I will also use your ideas on wiring, cause it looks much better than mine yet.

Concerning PCB milling:
I tested some more and figured out two major reasons:

Reason 1: a "not perfect" dremel tool - which wobbles in it's axis a little bit which may be the main reason for the wide cut lines.
Reason 2: the Z-axis lacks of a counter bearing on the lower side in combination with the V-reel. Any kind of gap will cause some less accurate results.

Speed and cutting depth variation did not show major differences because of the other two reasons (I will test this again when I have fixed the major problem).

How to fix:

1. I used spring washer for the z-axis screw (quick fix) to keep it tight (unfortunately it's now a "gliding" z-axis, but vibrations tend to loosen the upper screws very quickly)
2. I will replace the spindle with a more precise one (and use the stock tool for some less accurate woodworking)
3. I will rework the Z-axis completely (the ACME Z-axis goes in the right direction but the problem is the V-reel itself (in my opinion))

By the way, I will use a thinner double sided tape again to fix the board on the zero leveled plane. Seems to work better than clamps.

Best regards,
Ove