Topics

Rough edges on PCB traces

Scott Goldthwaite
 

I could use some advice on why the PCB trace edges are so rough (see attached pic).  I'm new at milling PCBs.  I have a ShapeOko 2 with a DeWalt DW660.  I'm using 10 degree engraving bit (titanium coated). Cutting depth is about 0.005", 15 IPM feedrate, 30,000 RPM.  I have about 0.010" runout.  Most of the edges are very rough, but some edges are really smooth.   Any thought to what my problem is?  Why are some edges nice and smooth while most are not. 

jeff.birt
 

Not all engraving bits are good for milling PCBs. Some do a really good job, others do a bad job. Whatever bit you use it will need to be carbide or it will not hold up no matter what it is coated with. The PCBs substrates are very abrasive and will quickly ruin a HSS/Cobalt bit. A 10 degree bit would not be my first choice. I mill almost everything with a 60 degree bit s it provides a good compromise between minimum trace resolution and single pass width of cut. The minimum I use is a 45 degree bit, I have not found anything shallower of much benefit.

 

For etching PCB, or anything really, 0.010” of run out is a HUGE amount. With a tiny engraving bit and that much run out you will quickly ruin the bit. The first thing I would is getting a more precise spindle. A Dewalt 611 with a set of PreciseBits collets is a good choice for smaller machines: http://www.soigeneris.com/precisebits_dewalt_611_precision_collets-details.aspx . Then look at the bits you are using. I used PreciseBits PCB bits years before opening up shop and becoming a dealer. This page has the PCB related bits so you can see them all: http://www.soigeneris.com/pcb_making_bits-list.aspx . With these bits a good starting point is to feed 1IPM for thousand spindle RPM.

 

I also found out recently that imperceptible vibrations in your PCB stock during the machining process can cause a poor cut quality. I had someone at the university ask me to machine some ‘Duroid’ PCB stock which is a 0.025” thick dual layer, Teflon core used for HF applications. It is very expensive and a bit like trying to machine a slice of cheese with copper foil on both sides. My first test came out very, very badly. I talked to Ron Reed at PreciseBits and he suggested that the problem might be with the stock vibrating whilst being machined. I took a small piece of the Duroid and used 3M Super 77 spray adhesive to glue it down to a ½” cast acrylic substrate. This not only had the advantage of preventing any vibrations but it held the stock down very, very flat. I was able to use a DOC of only 0.0014” (and 0.001” in one spot) with a 60 degree bit and machine two nice double sided microwave detection antennas (about 1.5”x1” in size). The board shad stapling (vias between the ground planes on both sides) that called for a 0.010” through hole (I had to do 0.011” as that was the closest bit I had on hand.) Even with the 60 degree bit I was able to resolve 0.010” wide traces. To release the glued down board you just need a hair dryer or low power heat gun, the 3M Super 77 releases about 150F and is easily cleaned up with acetone.

 

Jeff Birt

Soigeneirs.com

 

 

 

From: pcb-gcode@... [mailto:pcb-gcode@...] On Behalf Of scott@...
Sent: Monday, May 05, 2014 5:54 PM
To: pcb-gcode@...
Subject: [pcb-gcode] Rough edges on PCB traces

 

 

I could use some advice on why the PCB trace edges are so rough (see attached pic).  I'm new at milling PCBs.  I have a ShapeOko 2 with a DeWalt DW660.  I'm using 10 degree engraving bit (titanium coated). Cutting depth is about 0.005", 15 IPM feedrate, 30,000 RPM.  I have about 0.010" runout.  Most of the edges are very rough, but some edges are really smooth.   Any thought to what my problem is?  Why are some edges nice and smooth while most are not. 

Brian Volken
 

I see a HUGE difference when I give the board a good coat of WD40 before milling.  ALL my 'jaggie' edges are gone.  No need to sand the board.  

See the last two entries in:
WoodWorkerB CNC Router Project: New autoleveller software

 



Scott Goldthwaite
 

Jeff.

Thanks for all the info.  I'm using FR1 boards, not FR4 board so wouldn't that reduce the wear on the bit?  I got really good results on my first attempt, but my PCB wasn't flat, so it didn't mill everything.  But the traces came out pretty good in some spots, you can see a picture here: http://www.shapeoko.com/forum/viewtopic.php?f=28&t=3332#p25444  They didn't have any burs and didn't need any sanding.  Not sure why my 2nd attempt had such different results.  I'm holding the PCB with double stick masking tape on masonite milled flat.

I did some more testing today at different feedrates. I though bringing it way down to 3 IPM might help, but it did not. I did have some success with feedrates between 10 and 20 IPM with WD 40.  The board needed to be sanded a bit, but once it was it looked good.  That test was just with straight lines, so theres still some more testing to do.

I'm gonna pick up some other types of bits to see how they work out.  Probably a small end mill around 0.015" and some tapered bits like you mentioned.  

The runout sucks.  I think it's all in the collet because the DW660 shaft is pretty good.  

I might go with a DW611, but I don't want to put too much weight on my ShapeOko 2, it's not a very rigid design.

Paul Kiedrowski
 

Scott,
From your pictures I don't think runout is the problem, although .010" is pretty bad, its probably a poor quality collet.
High runout just means you'll mill a wider path than expected. Looks to me like you may have a machine vibration problem, which occurs at certain motor speeds.
Check for loose shaft/belt couplings, and run the stepper in microstep mode, not full step (quarter step is the coarsest you should use). In full step mode you always will get a motor resonance at low speed, usually at a few hundred steps per second (i.e. 1-2 rotations per second for a 200 step motor). You can usually hear that type of vibration too.
 
Also check the vertical stiffnes/rigidity of the Z-axis system, in other words it might start to skip across the surface at some speeds.  Spraying WD40 ahead of the tool as it travels definitely helps.
 
Sanding rough edges is fine for me, if its just curled up. There are a couple pics I took at pcbgcode.com showing how its fine after you sand them off.  But I prefer using a straight sided endmill for traces under 15 mils wide. I save my expensive 10 mil endmills for doing fine pitch, or final designs, and use cheap spade bits for initial work.
With end mills you don't have to mess with autoleveller if you can hold the PCB down well and you have a decently designed machine.
 
-Paul
 

From: "scott@..."
To: pcb-gcode@...
Sent: Monday, May 5, 2014 5:53 PM
Subject: [pcb-gcode] Rough edges on PCB traces

 
I could use some advice on why the PCB trace edges are so rough (see attached pic).  I'm new at milling PCBs.  I have a ShapeOko 2 with a DeWalt DW660.  I'm using 10 degree engraving bit (titanium coated). Cutting depth is about 0.005", 15 IPM feedrate, 30,000 RPM.  I have about 0.010" runout.  Most of the edges are very rough, but some edges are really smooth.   Any thought to what my problem is?  Why are some edges nice and smooth while most are not. 


scott.goldthwaite@...
 

Thanks Paul.  I'll look into the microsteps and see how to turn that on on my shapeoko.  I think everything is rigid enough (as much as it can be for a shapeoko).  I've checked to make sure nothing is loose and the v-wheels have a snug fit on the makerslide.  Since I'm using a DeWalt DW660, I can't adjust the speed (unless I buy some external speed control).  I ordered some 0.3mm end mills that should arrive on Tuesday, so I'll give it another try once I got them.

scott.goldthwaite@...
 

I took another shot at milling a PCB on my ShapeOko.  This time I used a 0.0118" end mill at 3 IPM and 30k RPM with plenty of WD40.  Results are pretty good.  Also, I got a new collet for my DeWalt DW660 which has a lot less runout: about 0.003" compared to 0.010" initially.