Quote Originally Posted by mayday_75 View Post
Is it really that hard to mill 0.65mm pitch SMD's?
I couldn't really tell you to be honest because I never even tried that small. I did try normal SOP (1.27mm) spacing and I could get it to work, with some room left, when everything was going just right - but it was clear I would be pushing things to the limit if I tried. At 0.65mm pitch, you have to be able to mill 0.3mm or so thin grooves _reliably_, and that's not easy considering the typical tool used is a "V"-tipped engraving stylus* with a 0.2 ... 0.1mm width _at the tip_, which grows as soon as you plunge into the copper - the deeper you go, the wider the groove becomes because of the "V" shape. It's possible levelling would solve this for you - as I noted, I never tried it, choosing to pre-mill the sacrificial HDF bed under the PCB flat first instead. It has to be said, this method definitely doesn't guarantee uniform groove depth unfortunately - you might be not going completely through the copper in one corner and going too deep and wide in another. Perhaps probing can fix all that for you though.

Quote Originally Posted by mayday_75 View Post
When I thought about the shady looking wood parts of the mill, I got to thinking that wood might actually absorb some vibrations? Adding precision if done right?
Speaking of vibrations, I've seen "gasket" type things that dampen vibrations from the stepper motors. Could that be something worth looking into?
PCB milling isn't really a high load milling operation (and it should be a high RPM one) so there really shouldn't be any vibration to speak of. What there will be however is lots of rapid movements jumping from track to track (or worse - form some part of track to some part of another track, randomly, depending whether your gerber-to-gcode converter optimizes or not) meaning that the rigidity of your frame does get tested (think of a printer starting to swing the whole table it's on while its carriage goes back and forth). If there's enough give in the frame to not get you back _exactly_ in the groove, you'll have misaligned grooves (and that's assuming you runout and backlash are zero - they never really are - and your motors are doing perfect microstepping). Wood frames are also rather vulnerable to humidity-caused subtle warping (so I hear - mine is aluminium), and in an application like this that can't possibly be good... As for the "gasket" things, I'm not sure what exactly those might be; at these hobby sizes, I haven't seen the motors themselves mounted any other way than rigidly screwed to the frame. However, what I have seen used (and my CNC came with them) are flexible shaft couplings (looking like normal couplings with transversal slots in them), which are supposed to dampen the instant torque change of the steppers against the inertia of the axes. I can't really tell you how indispensable or not they are, as I said my CNC came with them...

Quote Originally Posted by mayday_75 View Post
The probe, should I add an external pull-up resistor as in the limit switch schematic?
I've downloaded a probing SW I found via a YouTube video, autoleveller or similar I think it's called.
Yeah, you should - strictly speaking these things shouldn't be needed because there are internal pull-up resistors in the Arduino itself already doing the same job. However, CNCing is an electrically noisy business (which gets downright horrible if your spindle is VFD driven), so in practice you definitely want those resistors there (at the probe too), and possibly even some small capacitors in parallel with the switches / probe for the same purpose, if you keep encountering false "limit" triggers. Again, good luck!

*There are horribly thin cylindrical mills that don't care about penetration depth for consistent groove width but they're definitely not as widely used by amateurs due to higher cost - and frankly, they look like they're about to break if you just look at them wrong; even the tips of the sharper 0.1mm "V"s tend to break quite easily if you go too fast or spin too slowly (10000 RPM minimum, 20000 or more would be better) - you just notice that the tool suddenly doesn't seem to cut all the way through the copper anymore... There are also some nice "floating" attachments that "ride" the surface of the PCB and guarantee a uniform depth, but those are typically even rarer than the hair-thin cylindrical tools...