Difference between revisions of "PCB Milling on the Roland"

Latest revision as of 00:28, 17 July 2013

This is an on-going log of the adventures of milling PCB's on the Roland.

Cutters

New carbide cutters in 45deg and 60deg, the drill is 1mm for scale. The thickness of the tip is 0.1mm

Turns out the 45deg cutters are a bit fragile and the tips break easily.

This is the 60deg cutter after cutting 5 Ardunio sized PCB (top layer only), you can see the edge is starting to go on the tip (we only use at most 0.2mm of the cutter), but it's still producing good quality cuts.

The 60deg cutters appeared to give a better finish.

Boards

This board was cut with a 0.2mm cut width with a 60deg carbide cutter, the pads shown are for an FDTI chip (28pin SSOP for scale)

Here are some images from the same board, but with a 0.1mm isolation width, you can see the finer cuts by the extra copper left between some of the tracks.

Work Holding

The biggest issue so far has been holding the work flat. As the cut depth has to be 0.2mm or less, you need a good level surface to ensure a constant cut depth. Pro machines have pressure feet and floating heads, some even map the surface of the board prior to cutting and adjust cut depth accordingly. We have none of this so need to improvise.

First attempt, MDF machined flat. This is a no starter, no where near the tolerance required.

Next use a sheet of Model Board, this has good dimensional tolerance and is stable over time.

The next problem was the fact we were using double sided tape to hold down the PCB, that was no good, to much variation.

So tried clamping the board, but this still lead to warps in the board.

Hint: When zeroing the Z-axis, place the sensor in contact with the PCB and zero directly onto the PCB, then move to +15.00mm in Z and re-zero. This will remove any inaccuracy in the thickness of the sensor.

Vacuum Clamping Project

So at this point it seemed a vacuum bed would be a good idea.

This setup gives really good results and allowed a successful test pattern to be milled into an area 150x100mm with no partial cut through as had been seen before.

The problem is now that the system leaks, note the increase in taped edges in the above photos to minimise losses.

While this is to be expected, this causes problems for my 2 Stage Rotary Vane vacuum pump. That pump (which is normally used for degassing resin and resin infusion) does not like a partial vacuum, so the leaks above lead it to over heat, will radically shorten it's useful life.

An attempt using the Henry Hoover, while providing enough vacuum, caused a worrying amount of heat.

Next on the list of things to try will be the pump donated by Rachel, but am unsure if that will provide the required level of vacuum as it's really a blower not a vacuum source. An adapter needs to be manufactured first.

If that fails then I'll look into sourcing a side channel pump, which are more suited to running for long periods in these conditions.