Further experiments milling PCBs

I have spent many hours CNC milling circuit boards and I have documented my process elsewhere on this blog. The workflow is to first design the board in Fusion Electrical (or any similar CAD package) and export the manufacturing package as Gerber files for the copper artwork and an Excellon file for the drilling information. The manufacturing files can be imported into FlatCAM to convert them to CNC files for milling operations. This is all documented in my various ramblings. Double and single sided board are both accomodated by the FlatCAM software (you just need to get your head around manipulating the artworks).

I have been reasonably happy with the results that I achieve but as ever I tend not to rest with a process which has some niggles. As a result I keep searching for a breakthrough lightbulb moment or at least an interative advancement to make the results more repeatable/easier etc.

The first irritation was the cost of the milling tools I was using. I bought these in from Think and Tinker in the US and they cost around USD20 each for a 5 thou end mill. One careless twitch and your bank account takes a hit.

Let’s take a step back and refresh. The FlatCAM process produces a file that moves the cutter around a copper clad board to leave isolated representations of the copper trackwork. The gap between the wanted trackwork and the rest of the unwanted copper is defined by the width of the milling cutter. FlatCAM allows you to widen the isolating gap by taking additional milling paths side by side to each other with a defined overlap. With a 5 thou cutter I used to take two parallel cuts with a 10% overlap so the resulting gap was 9.5thou. This is a pretty narrow gap. The second time round cut does tend to raise burrs/slivers of copper and these can easily cause a bridge between the wanted and unwanted trackwork.

I make this more difficult for myself as I prefer to tin the copper after milling. (Scrub the copper clean, daub with flux and then deftly swish a soldering iron across the copper). The end result is good but the road to it is fraught. Before tinning, the board needs to be checked for milling shorts and all milling debris removed and scrubbed clean. After the tinning process there will be a number of self induced solder bridging shorts which also have to be cleared. A quick post tinning run round the board with SolderWick is a fast track route to remove these. The process works and the tinned end result looks much better than bare copper.

The first area of my current investigation was the milling tools. Could a larger diameter milling tool be sourced at a price advantage to the US parts and how much larger diameter would be practical to give acceptable artworks. The solution came from APT Tooling who are one of my regular suppliers of tooling. They provide a range of micro milling tools starting from 0.3mm (11.8 thou) and upwards. A single pass at this size would achieve the same isolating gap as the two passes with overlap with the 5 thou parts. Potentially a single path would reduce the swarf debris. APT offer two versions of the micromills with different hardnesses.

My PCB artworks are not high density and with my eyesight constraints I tend to stick with through hole components whenever possible and 40thou trackwork. I ordered in 0.3mm and 0.4mm milling bits (55 HRC versions) and ran the FlatCAM simulations of the 0.4mm isolation routing. The 0.4mm (15.7 thou) had the distinct advantage of creating a wider isolation gap while still allowing inter pin gaps to be cut. Adding a second cut with the same 10% overlap created a very acceptable artwork with a generous isolation gap. FlatCAM is nice in that providing the first pass can cut 100% of the isolation artwork, the second cut automatically truncates cutting gaps that would not be compatible with the artwork and damage the geometry. Here are some FlatCAM simulations

Here are the actual results of an artwork cut as a single 0.4mm cut and as a double cut with 10% overlap. (I terminated the single cut as the resulting partial artwork was enough to make a decision on). Notice that the double cut does leave mini floating islands between wanted trackwork as shown in the simulations. These can be easily removed with a hot soldering iron.

Whenever I get a new supply of PCB material I strip peel a small section and measure the thickness of the copper and the overall thickness of the board. Currently I am using RS single sided board (RS#219-2123) and this measures 2 thou copper and 60 thou overall. With this material I am running a 4 thou depth with a feed rate of 10″ per minute. I run all PCB operations on my Tormach PCNC440 with my home made vacuum plate.

Conclusion – for my level of artwork density I now believe I can get acceptable results using a 0.4mm cutter in either a single cut or overlapping double cut process. The APT 0.4mm tools are much lower cost than the US sourced parts and being larger diameter will hopefully be more robust. To date I have only tried their 55RHC versions. Note I always run both the milling cuts and drilling routines with spray coolant from a Fogbuster with my normal spray coolant (QualiChem Xtreme Cut 250C) . This has the advantage that the fluid enters the cuts and shows as a clear black line when the cut depth is correct or at least deep enough to go through the copper.

If anyone wants more information about my process, please get in touch.

Links to similar or related post are listed below : –

Vacuum table update for PCB milling

I have mentioned in the past my home made vacuum plate for holding printed circuit board material flat while milling the copper trackwork.   As a technique this works really well and the results are very consistent using 5 and 10 thou miniature milling cutters from Think and Tinker.

When the artwork design needs holes drilling through the PCB material, care has to be taken not to drill through into the aluminium surface of the vacuum table.  To protect the carbide drills and the surface of the vacuum table I use a sheet of 3mm MDF underneath the PCB material to act as a sacrificial  board .  While the MDF is ‘transparent’ to the vacuum suction it does degrade the downward vacuum grip and this has always been a frustration.   If the board does not need through holes then the MDF is not required.

My wife uses an anti-slip perforated rubber sheet on the kitchen sink draining surface.   This is very flexible and looks like a weird cobweb of holes.   It struck me that this could be used as a sacrificial sheet between the PCB material and the vacuum table surface.   The rubbery material is not perfectly flat but is very easily compressed to become consistently flat over a large area.

I begged a sheet of the material from the kitchen ‘sub stores’ and mounted it on the vacuum table with the PCB on top.   I measured the overall depth of the PCB and the mat under vacuum compression to be circa 80 thou and I usually set the depth of cut for drilling to 68 thou.  This looked like it should leave a safety margin to avoid damage to the drilling tool and the vacuum table surface.

Using anti slip rubber sheet as a sacrificial spacer
The rubber mat in place on the vacuum table

The idea worked a treat.   The PCB was very rigidly clamped  in place and no damage was done to the drill bits or the vacuum plate.  Here is an image of the first board being run using this backing material technique.

Test PCB being milled using the rubber sacrificial mat
Test PCB being milled using the rubber sacrificial mat instead of MDF

This looks like a step forward in my PCB milling process and has the added advantage that the rubber sheet can be hand washed afterwards and re-used.

Anyone needing some 3mm MDF ?

Similar or related subjects : –

Update on CNC milling printed circuit boards on a homemade vacuum table

Update to the use of FlatCAM to mill PCBs

There are quite a few entries on my blog regarding using FlatCAM to convert PCB design software manufacturing files into CNC code.   I also have mentioned my small home made vacuum table and a floating foot compression device both for holding the PCB blank flat while the milling takes place.

I have revised my original write up to focus on FlatCAM version 8.991 and also pulled together notes on these other techniques.   If you like it let me know.  If there are mistakes also let me know.

FlatCAM and milling pcbs updated notes 2021

Similar or related subjects : –

Floating pressure foot for the CNCEST3040T mini milling machine

A new idea for keeping PCB material flat while milling artworks

The vacuum plate mentioned elsewhere on my blog serves me well when milling printed circuit boards on the Tormach PCNC440.   It keeps the PCB material flat and makes the cut widths repeatable when using V cutters.

The plate cannot be easily used on my CNCEST3040 due to the restricted Z height.   We have experimented with various techniques to keep the PCB material clamped flat on the smaller mill with varying degrees of success.

Idle hands and brain during social distancing has produced a possible solution that might be of interest and stimulation to others.   It consists of a circular pressure ring that sits around the spindle chuck and tool.   There is a second ring that sits on the spindle body connected to the lower ring with rods which have coaxial springs pushing down on the lower ring.   The magic is to use mini ball transfer units on the lower ring to press down on the PCB and glide friction free around the PCB as the cutter does its stuff.  The assembly is held in place on the spindle with 3 gripping screws.   The downward pressure is adjusted by 3 screws that press against the spindle mounting frame.

The ball transfer units come in all sizes and are very common in baggage handling systems at airports and in industrial conveyor systems.  The ones I used came from RS and have a 4.8mm ball and a M2 mounting shank

The prototype was made using 3D printed rings.   There is an image below.  Apologies for the yellow PLA but finding any PLA at a decent price is very difficult in the present circumstances.

Bottom view of pressure foot for CNCEST3040
A view of the underside of the lower ring and the four ball transfer units. In the background is the upper ring that sits around the spindle with the pressure adjusting screws and the spindle gripping screws.
Pressure foot for the CNCEST3040 in place on the spindle
View of the pressure foot in place on the spindle showing the tension adjusting screws and spindle grip screws

The idea seems to work and has produced some good consistent quality PCB prints.   It does have disadvantages in that you need to have a larger PCB blank to allow for the larger footprint of the pressure ring.   It is probably only of practical use for PCB milling but then the problem of flatness is less critical in drilling the board and routing the profile.

Similar or related subjects : –

Verified by ExactMetrics