CNC for printed circuits boards

Sensacell

Joined Jun 19, 2012
2,544
Only for the simplest small boards.

The frustration factor will be huge for anything big (flatness issues) or complex.
Plan on lot's of busted cutters!
 

Thread Starter

be80be

Joined Jul 5, 2008
1,989
I was thinking about getting one to play with I'll make it a little stronger.
I was wondering more about if the motor and steppers be good to mess around with.
 

philba

Joined Aug 17, 2017
960
For milling PCBs, you don't need a lot of strength or super accuracy, for that matter. If you get 25 micron (~1 mils) accuracy you are probably golden. You can probably live with 2 to 4X worse.

Playing devils advocate, a milled board still has all the draw backs of home etched boards - no plated through holes/vias, no solder mask, no silk screen. And, with low cost, fast turn PCB manufacturing it seems like a lot of effort and expense when you can get high quality double sided boards made in less than a week for less than $20 delivered.

That said, a small mill is a great thing to have - making front panels, stencils, test fixtures, all sorts of stuff. Along with a 3D printer, it makes for a pretty decent shop.

On that specific machine you linked to. It doesn't look very rigid. The supports are plastic and rails are only 20x20mm. Take a look at openbuilds part store. Not super cheap but by the time you finish upgrading that Chinese machine, you probably won't be that far off.
 

Thread Starter

be80be

Joined Jul 5, 2008
1,989
I found a 3D printer it looks kind of good. I just don't like waiting so I'm thinking about a little shop to make my toys LOL
Screenshot from 2018-01-25 02-37-32.png
 

Picbuster

Joined Dec 2, 2013
990
it's a nice thing to engrave your front door name plate but useless for real milling.
milling any kind of material need stability. The table mass should be at least 50x your the mass of your milling object ( 1kg material milling table at least 50Kg) and yes, mill size, dept. and feed are also effecting the table mass.

Then the axis drive shaft should be ballscrews with a play of less then 2 micro millimetre.
x,y,z movement should be driven simultaneously this to avoid(minimize) stair case effect.
Again I will not try make you unhappy but try to protect you wasting money. (milling and their problems are part of my existence).

Picbuster
 

AlbertHall

Joined Jun 4, 2014
9,025
I have a CNC3020T mill on which I make very acceptable PCBs. I use a 0.3mm end mill and can get 0.3mm tracks at 0.3mm spacing which makes it possible to put a track between the pins of 0.1" spacing ICs.

I use Eagle with PCBgcode (both free) to produce the Gcode for a board. To get around the flatness issue I used to use the commercial AutoLeveller which adds code to probe the surface of the board and automatically adjust the height while etching the board. As I wanted some other things I wrote a VB6 post-processor which now includes my own version of auto-levelling. This program combines all the files (top and bottom, etch, text, drill and mill) into one gcode file. It also drills the holes for two dowels to accurately locate the board when flipping it for double sided boards. It produces code for LinuxCNC (also free).

If anyone would like a copy of the code I am happy to share but note that it is not very well error-trapped and there are almost certainly bugs left in there.
 

ScottWang

Joined Aug 23, 2012
6,860
I went to a place to see the diy CNC machine for PCB, but it looks like not so good, and a friend bought a CNC machine for PCB from China, the price around US$120, but it didn't including the charge of shipment, and after you buy it then you need do a lots of adjustment for the drill and machine, it seems not that easy to use, a lots of works to learn, although it will meet a little trouble, but I still would like the use that kind of machine to milling the PCB, it seems better than to use the chemical method, but the copper wire of PCB can't layout too small as you can't layout 2 lines between two pins of dip ic, I hope someday could buy one, maybe the quality is more higher, and you will need to use the drill like this -- Drill Bits, it is easy to break, so when you use it must be very careful.

The attached pcb file were made by a friend.
 

Attachments

Thread Starter

be80be

Joined Jul 5, 2008
1,989
AlbertHall If you don't mine i can use the files Im getting the cnc I use strip board to test out ideas i think this would be ok for what im going to do. thanks

I got a cheap lathe it had tubes for the bed i changed them to solid bars it worked great for years.
i think i could fix this into something for printed circuit boards.

ScottWang that looks really good. thanks

I'm just wanting to dabble with this. LOL
 

spinnaker

Joined Oct 29, 2009
7,837
I have a similar cheap machine. I am simply not satisfied with the results of isolation routing. I get inconsistent results. The main issue I understand is that the spindle cannot generate sufficient RPM.

I went with a 'hybrid" process. I first drill with the CNC machine and then use the toner transfer method to etch my board.

If you are interested in this method let me know and I can try to document.

One of the supposed plus sides of isolation routing is eliminating chemicals but when you see the amount of fiberglass dust that gets generated from isolation routing, you will wonder where the environmental benefits are of eliminating chemicals.
 

AlbertHall

Joined Jun 4, 2014
9,025
AlbertHall If you don't mine i can use the files Im getting the cnc I use strip board to test out ideas i think this would be ok for what im going to do. thanks

I got a cheap lathe it had tubes for the bed i changed them to solid bars it worked great for years.
i think i could fix this into something for printed circuit boards.

ScottWang that looks really good. thanks

I'm just wanting to dabble with this. LOL
I hope this is a usable link to the files: https://1drv.ms/u/s!Ah0XLGOdlqiianOkV0eepMUzMJs
 

AlbertHall

Joined Jun 4, 2014
9,025
One of the supposed plus sides of isolation routing is eliminating chemicals but when you see the amount of fiberglass dust that gets generated from isolation routing, you will wonder where the environmental benefits are of eliminating chemicals.
I don't use fibreglass board for that reason, also it wears out tools quickly. I use FR2 boards, SRBF.
 

Janis59

Joined Aug 21, 2017
970
RE:""I want to make my boards at home.""
Then use a Positiv-20 process. Cheap, very accurate, practical. May add the green lack and never will guess its not a commercial product. All process takes just half-hour.
The CNC for PCB - very inaccurate, very expensive, very slow (hours), very bad electrically (parasitic capacitance), bad looking.
My CNC is very accurate - about 5 microns. Yet for pcb the result is so bad that even the iron+paper method gives a better view.
 

Janis59

Joined Aug 21, 2017
970
RE:""I use FR2 boards""
Why??? Them (FR2) are mechanically and thermally so weak into comparison with FR4, and 150% more lossy what is rather big problem for RF circuits? Do You do that for be cheaper?? Its no worth to economy on the quality indeed.
 

AlbertHall

Joined Jun 4, 2014
9,025
I can only say that I have never broken an FR2 board or set one on fire.
I am just about to embark on one of my rare RF circuits on SRBF. Mind you the RF frequency concerned is the extremely high 60kHz.
 

MrSoftware

Joined Oct 29, 2013
1,666
At work we have a desktop CNC by Other Machine Company (OMC). I haven't used it myself, but i have seen it used for proto PCBs and it works. I can't offer any other details unfortunately, but I have seen it used to make multiple PCBs.
 
Top