Success!! milling PCB... First PCB is and ARDUINO!!

[cvoinescu]

As for the vias, could you not just make the hole a bit wider and use the leg of the component and solder to do it? How many are there typically? Could you not use a small wire?

With homemade dual-side boards, I try to avoid using a component solder point as a via, because it would require the component to be soldered on both sides of the board. With resistors and diodes that's okay, with ICs it's iffy, and with electrolytic capacitors, pin headers and IC sockets it's impossible (no access to the top ring, because the part covers it). I try to make the boards single-side, and use jumpers on the top side. However, I still design the boards as a two-sided boards, taking care to do only straight lines on the top layer (and as few of them as I can).

That said, I much prefer to use prototyping boards and Kynar wire for one-offs. Tri-pad is my favorite; I can also work with boards with unconnected pads, but I hate the ones with long strips. For more than about three boards, I use Itead.

[thiagokunz]

I don't think g code cares what extension the file had any more than a text editor - it just assumes it is g code.

Hi Nigel. I made a test. I just changed the .tap extension to .nc. It worked just fine. I milled som soft card board, and guess what... I broke my 0.2mm drill bit. I think I must find a stronger one.
The good news is it worked like a charm. No gcodes unsupported and no warnings.

Milled cardboard - directly from Eagle + PCB-Gcode. 0.2mm Drill bit and the holes with 0.8mm.

1604646_10202501461354506_1013782644_n.jpg (160.06 KiB) Viewed 1650 times

for the DIY side of the world, the chemical direction is not the fastest route, especially if you are trying to do 2 sided boards. There is a lot of blood, sweat, and tears when trying to do 2 sided boards.

MeanderBolt... Yes... my first boards I made with photo sensitive ink and acid to etch. A whole morning to complete this process.

As for vias, I use a very thin wire (single strand wire left over from a phone system installation). For the vias, I usually choose a drill that is just about big enough for this wire. I then strip off the insulation so it is just one long bit of shiny wire and I literally weave all the via holes. I keep a print out with a pattern that I have penned in so I do not miss any, and so that I do not go through the wrong hole. After all the holes are filled with the wire, I solder both sides of the board, then trim with flush snips. This is by no means a fast process.

Do you have any documentation on this process? It seems nice. I saw somewhere some people selling some rivet-like vias. It's interesting for 2 layer pcbs, but it isn't made for multilayer pcbs. Anyway, don't you think the electrolytic processo of making vias could be faster than this thin wire procedure? I'm just throwing some thoughts in the air.

I'm also in a project of making a solder reflow oven. I have all electronics and I'm struggling with the code. There is a good tutorial by Ben Heck. You can do SMD's easily. I have made the board design... And if it works I'll post here somewhere.

[thiagokunz]

Probably I broke my drill because of the rpm and feed rate. I've found this table that could help us to avoid this issue:

http://www.precisebits.com/Reference/drillfeedspeed.asp

There is a stronger material for drill bits for pcb than tungsten carbide?

[MeanderBolt]

Thiagokunz, do some reading in the PCB threads. I have not yet tried, but if I am remembering correctly, I think that you need a 30 degree V bit.

As to the via thing, I read a lot of different ideas and eventually decided after trying several that failed miserably, that I would just thread it and solder both sides. It takes no time at all to solder. The time to thread it is a little more, but that is because I am cheap and try to do it all with one 16 wire. Stupidity, I know.

I try to avoid using a component solder point as a via, because it would require the component to be soldered on both sides of the board. With resistors and diodes that's okay, with ICs it's iffy, and with electrolytic capacitors, pin headers and IC sockets it's impossible (no access to the top ring, because the part covers it). I try to make the boards single-side, and use jumpers on the top side.

I completely agree. This is similar to my approach to the vias. Once you have been bit by going through the whole process then finding out that your board is unusable for one reason or another, you tend to remember it. Home one offs are great because you get them right now, but you have to be very mindful about where things are and how you will be assured of getting signals from point a to point b. It does require soldering on both sides of the board many times. As to the cap thing, myeahhh, I got burned by this too. I had to lift it up and solder then bend the leads over.
I have gone mostly surface mount because 1. there are no holes to deal with and 2. it looks bad ass and 3. people look at you (who do not do this) just befuttled as to how someone could do this.

[miketedeschi]

Another tip for homemade PCb is to adjust the design rules in your PCB software. Inventables did a PCB stress test for the shapeoko 2, and the TQFP was good but not perfect, I think it's because of the V-shaped bits. (based on discussion from another user above this post.)

I would love to skip the Z-axis perfection issue, if possible.

I just ordered some 0.3mm and 0.5mm carbide end mill bits (to avoid the height issue like the other user mentioned) and I am going to attempt TQFP for sure. I also like SMD personally, mainly because leaves the other side of the board flat without risk of shorting anything, but also because i've been getting into reflow soldering with decent paste and it's so much better!

0.5mm ~ 19.7 mil
0.3mm ~ 11.8 mil

In theory, the 0.3mm end mill could be perfect for TQFP spacing, without worrying about the v-bit height variances. The ATMEGA328p and FT232R chips in the TQFP package seem to both have 0.3mm pins, with 0.35 and 0.5mm spacing respectively.

Since 0.3mm is about 12mil, you could mill 12mil traces with 12mil spacing all day, hopefully. I'm going to try it and share the results next week.

Is that eagle to PCB GCODE sender the way to go??? thanks. -Mike

[miketedeschi]

http://www.precisebits.com/Reference/drillfeedspeed.asp this RPM and feed rate chart says 80,000 rpm for a 0.3mm carbite bit for FR4???

can the stock shapeoko 2 spindle do 80,000 rpm??

[WillAdams]

No. 30,000 tops: http://www.shapeoko.com/wiki/index.php/Spindle_Options