PCB Production

[spritesmods.com]

J.

I've only just noticed this thread, I've been doing some experiments with plotting directly onto a copper clad board using a pen, there's a couple of photos of my best results so far here:
[www.flickr.com]
[www.flickr.com]
[www.flickr.com]
The pen I'm using at the moment is about 0.8mm but it just ran out so I'm looking for a smaller one now.

At the moment the program only supports using a pen, but I'm hoping to expand it to be able to use multiple tools at some point (cutter, laser, paintbrush etc). The program can load gerber, dxf & (soon) svg files.

I had a look at that page with the printer converted to make PCBs, but I couldn't see any photos of the actual end results (i.e. working PCBs) anywhere, was I missing something or is it still in progress?

Edited 1 time(s). Last edit at 05/26/2008 09:18AM by greenarrow.

greenarrow~~

Might I suggest a scribe? That way you could etch things quite easily. Just make a channel which has a spring loaded cap, thus always forcing the scribe onto the work surface.

---

Jay

A lot of force will be required for a scribe (both vertical and horizontal), I'm not sure how well my cartesian frame or a darwin frame would cope with it. A motorised spinning bit (something like a dremel) might work a bit better.

Once I get my system up and running, I wanted the most simple way to make PCBs. I wanted to avoid any etching if possible.
What I was thinking of doing was using my dremel to cut a small groove in.... something. The same way that if you lift a trace off a regular board, there is a groove left behind. Then melt some sort of metal into that groove. I don't know what kind of conductivity there is in some of these really low melting point metals, but having a multihead system could end up making the whole board at once (one side). The other side would need some sort of calibration marking or something to get it to line up. The same way a PIC has the upper left pin as pin 1 in the diagrams, put some groove/wall system on the reprap bed to press and lock the board to as a home position. As long as the board is flipped the right way, it should work out.

After six months I found the link!

[www.instructables.com]

Some modifications towards automation and this low-tech approach might be exactly what we need!

The plastic paint could be bio-degradable since it is only used for a few hours..

Then I don't know what the by-products are, copper-chloride? or will the copper
attach itself to one of the poles?

I had a go at saltwater etching a few months ago and just couldn't get it to work. I tried several times, eventually even resorting to covering the board in wax and scratching bit out, still with no success

Quote

What I was thinking of doing was using my dremel to cut a small groove in.... something. The same way that if you lift a trace off a regular board, there is a groove left behind. Then melt some sort of metal into that groove

If you're already milling, why not mill isolation tracks into pcb blanks, or even make your own conductive-metal-over-hear-resistant-substrate sheets? That would be somewhat quicker and easier, I'd say. Melting metal into a groove is significantly more complicated than is sounds.

I'd be more interested in a field's metal extrusion onto an extruded plastic substrate... no need for a precise and rigid cartesian bot that can withstand milling stresses, and the potential for nifty 3d circuits and the like.

extruding the fields style metal was what I meant by melting. I contacted a company last night that deals with low melt alloys to see if they could make a highly conductive low melt alloy. I'm not too sure how this would work with the final product for soldering. We could extrude the base and the metal, but when we go to solder chis and resistors on, it would melt the metal and the plastic. If we use PCB blanks, then I don't think we'd have a problem with the base, and the metal would actually be welded at the temp of the iron. It 'sounds' ok, but we won't know until we try.
Here is a mockup of what I had in mind for the head and how it would animate. If anyone is interested in building one I can provide better details.


If you're wondering how the selected tool will stay in place, there would be an automated locking pin mechanism.
This way it could mill the grooves, extrude the metal into it, use metalic pens to draw the silkscreen.
I doubt this type of head would be good for more than one or two extruder heads though.

Why do you need to make grooves in a board, can the fields metal not just be extruded onto deposited plastic?

Can the metal not also be extruded onto the components once they are placed, eliminating the need for soldering afterwards.

I'm only working on my PCB plotting to be useful in the meantime until there is an effective additive circuit process

Has anyone thought about 3D circuit design yet? I can't see any way we can make use of existing PCB software, we will probably need to design a tool which will import netlists. There possible avenues for designs : placing components arbitrarily in 3D space; or placing components on multiple layers. Layers will probably be easier to design in, but will not be as compact. However it will be easier (but harder to make) for a computer to auto-route efficiently in 3D, in which case layers would not be necessary.

Anyway, its probably best we get the process working in 2D first

The grooves would keep the metal in place. If you try to solder to it with no grooves, once you get two runs together, it would be really hard to get them apart again. It would be nice to put the components on the board first, then put the runs on to save production time too, but you'd need a pretty thin extruder head to not bump into parts, not to mention you'd have to recode the software (gcode I guess) to tell it to not go over the areas that have components on it. You've seen what happens in your tests how it goes to the next component without regard to whats in the way.
We all want the final product, but we probably should start simple first. My vote would be to do something that doesn't require etching, but we would probably need a highly conductive low melt metal.
Does anyone know the resistance in fields metal?

Dylan Wrote:
-------------------------------------------------------
> extruding the fields style metal was what I meant
> by melting. I contacted a company last night that
> deals with low melt alloys to see if they could
> make a highly conductive low melt alloy. I'm not
> too sure how this would work with the final
> product for soldering. We could extrude the base
> and the metal, but when we go to solder chis and
> resistors on, it would melt the metal and the
> plastic.

Good point. It would probably not have too much effect on the plastic since it has a pretty high specific heat. Waving a hot soldering iron around a Field's metal trace, however, could be tricky.

... the problem could be the bad adhesion between metal and plastic ...

First when printing metal on the surface and further when remelting for connecting the pins - it wouldn't stick to the surface and form blobs instead of neat trays ...

I had an idea where you have some metal-powder or metallic salt embedded in the plastic (not so much, so it wouldn't be conducting) - then the aplied metal could stick to the embedded particles.

Or draw the trays with an IR-laser to 'activate' the particles and let grow metal on the activated areas as in the MID-process ( [forums.reprap.org] ) ...

Viktor

Not a bad idea, there is also the conductive glue I've seen on sites like instructables that could be used as a base to adhere to the plastic and would help fuse the metal back to it. This would probably be easier to accomplish than a metal plastic alloy. Especially with a multihead system like I showed above. There is also the head changing unit that I saw on one of the videos on the wiki.
The metal plastic alloy might short some circuits if too close, or with less metal powder to avoid shorts wouldn't bind with the extruded metal runs.

I use the ironing pcb method.
For those who does know, the method is simple:
1. Print the pcb traces on a paper with a laser printer
2. iron iton the raw pcb board
3. if succesfully ironing, then etch it.

The most difficult part in this process is the actual ironing, and if two sided board is necessary its way more difficult to properly align.

So I was thinking about it, and I think it could be automatised with the reprap.

1. We could use the laser printers' powder and directly deposit it on the raw pcb using Adrian excellent idea:
[blog.reprap.org]

2. With a laser diode (is a dvd-writer laser diode is strong enough?) we could melt the powder, and therefore it will stick on the surface.

3. etching as usual.

With this method we can align the two side of the board without trouble, so it can directly made all the reprap's pcb's without modification.

what do you think?

Isn't toner power carcinogenic?

Certainly a method to accurately register double sided boards with a high degree of repeatability would be great.

I have wasted too much photo sensitive board attempting repeatable results.

aka47

---

Necessity hopefully becomes the absentee parent of successfully invented children.

Toner powder isn't too bad, wouldn't want to breath the dust often though. The benzine in the liquid part shouldn't be messed with, I opened a bottle of that for about 10 seconds and it was enough to stink up a 500 ft^3 room for half an hour with the windows open and gave me an instant headache.

If you deposit the powder really close to the surface and you molt it instantanously by the laser diode, I dont think there will get out any additional powder in the air.

What do you think?

I heard too about carcinogenic, this is why I dont fill my toner at home.

What about making a paste of the toner and alcohol and depositing it using a syringe depositor thingy and fine needle.

Toner is usually of a known particle size/class.

Push comes to shove bake it on, in an oven or under a grill.

aka47

---

Necessity hopefully becomes the absentee parent of successfully invented children.

Hi Andy,

... i tried sometimes to make a paste with laser-toner from Sharp - either the toner didn't mix with the solvent, or when, then some of the ingredients thend to stick together, so the paste won't go through the needle

AFAIK the toner consists mainly of latex and graphite with some additives.

So you have to mix with the graphite, but shouldn't solve the latex ...

Viktor

Shame that

Pastes are quite tricky things, I have all but given up on messing around with Epsom salts as a potential support paste for similar reasons. Controlling crystal size and maintaining plasticity is problematic.

aka47

---

Necessity hopefully becomes the absentee parent of successfully invented children.

Hum I had a go at salt-water etching the other day. I didn't try to do anything useful but I managed to get a way a lot of the copper that way.

I also got some brown powder into my salty water, obviously it was some kind of
copper salt.

It could be easily filtered out with a coffefilter.

I used about 0.5 amps with the cathode in the saltwater.

Any one knows enough chemistry to attempt a formula? I think the salt water process is doable, but it would be nice to know if I'm creating a major environmental disaster should this be a success ;-)

Did you manage to etch it all the way through? When I tried it, it would always stop working before it got to the end. I think it must be because there is too much [copper oxide?] covering the copper and stopping the process. It would continue if this were removed, but it would be hard to not remove the pcb masking at the same time.