Homebrew PCBs

I've been getting pretty good results lately with the last several 
boards I did (10 mil traces, 10 mil spaces between lines), so I guess 
I'll share my methods in case anyone wants to try them.

First, cleaning the board is critical.  It has to be evenly clean and 
there can NOT be any scratches in the surface of the copper.  I'm not 
talking about "tooth" to the copper, but actual scratches that you can 
see with your eye.  These larger scratches will not take any toner, so 
there will be a very fine break in your lines if a scratch crosses a 
line.  My cleaning method gets it perfectly clean and makes any large 
scratch very visible before transfer:

1) I clean the board with hot water, Comet (or Ajax) kitchen cleaning 
powder and a gray 3M scrubbing pad.  Use a VERY generous amount of Comet 
(I shoot for almost a slurry in the pad) on the pad and board and do the 
normal scrub up, down and around in circles.  The Comet and Ajax has a 
very fine abrasive in the powder that really helps to scrub that board.

2) GET THE EDGES.  Any time I had issues before, it was almost always 
near an edge or a corner.  It is harder to get them perfectly clean, so 
spend a bit of extra time on those spots to make sure they are clean.

3) Rinse the board in hot water and check for any large scratches.  At 
this point, it will be covered with tiny scratches caused by the scrub 
pad, but very deep scratches will still stand out.  If you have any, 
scrub some more until they are blended in with the rest of the board 
(aka: you can't see them!)

4) (this is where I differ from anyone else I have read about) Sand the 
board with soap and warm water with 2000 grit 3M automotive sandpaper 
and a small sanding block.  This paper almost polishes the copper and 
gets it very flat.  "Almost" is the key.  The surface looks like a 
copper mirror (you can see the fiberglass weave in it), but there is 
still a very fine "haze" of microfine scratches covering the board.  
Frankly, this is part of the same procedure I used to color sand some 
defects out of my car's paint.  That's also why I had the supplies 
laying around and decided to try it the last time I had a crappy 
transfer because the board still wasn't perfectly clean.

5) Rinse the board in hot water and check for large scratches.  If you 
find any, go back over it with the paper a few more times to polish them 
out.  Don't just ignore them or you will be trying to track down an open 
in some small trace that you just can't see once the board is etched.

6) If you are happy with the board, give it a final wipe down with 
denatured alcohol  right before you do your toner transfer.  Acetone 
works as well, but I find the denatured alcohol to be much easier on my 
hands and nose (yes, I'm usually too lazy to go get my heavy chemical 
gloves when I am "only" using a small amount of acetone...)

When I switched to this method of board prep (prior I was using soap, 
water and scrub pad), the quality of the etched board went WAY up.  I 
believe it is because the board has a smoother, flatter surface for the 
toner to transfer to, so the tone transfers more consistently across the 
board.  The 2000 grit paper still leaves enough tooth for the toner to 
grip well, but it is small enough that it doesn't impact even the 
smallest feature of the circuit pattern.  Also, not that I specified 3M 
automotive paper and not just anything that is available.  The higher 
quality papers have a VERY consistent girt size which leaves a very 
consistent scratch pattern.  Cheaper paper may say 2000 grit, but that 
is only an average.   They will have stuff larger and smaller than that 
and those larger scratches can come back and bite you.

The other change I made that improve my boards (but not as dramatic as 
the cleaning procedure) was to start using a GRC laminator.  Mine was 
modified with the temperature controller, so I can run it much hotter 
than normal.  I could do a pretty good job with an iron, but the 
laminator is more consistent and a lot easier (Feed board.  Rotate.  
Repeat 10 times).

Paper wise, I have tried a few different materials.  My normal go to 
paper is the Staples photo gloss paper.  After trying it, I bought a 100 
sheet pack, so I am good for a while.

I have also tried the ink jet transparency (the coating used to allow 
the ink to stick also allows the toner to cleanly transfer from the 
plastic to the board) with good results.  The transparency works very 
well with double sided boards as it is simple to get both sides very 
accurately registered to each other before inserting the board.

I also tried the Avery name tag labels.  This method is a bit different 
as you peal the label off BEFORE to print it.  This puts the toner on 
the waxy backing paper instead of on the label which helps it release 
onto the board.

All in all, before I started my current board cleaning method, I felt 
like I was getting better transfers from the transparency and the Avery 
labels, but I also thought I was getting better details from the Staples 
photo paper.  In the end, I generally ended up using the paper and 
occasionally had to re clean the board and redo it when a piece of toner 
didn't transfer correctly (usually around the edges or corners).  Since 
I changed my cleaning method, I have not had one single failure from the 
Staples paper.

The only "down side" to this method is the paper does NOT want to come 
off the board!  It has to be soaked, pealed, soaked some more, scrubbed, 
soaked, scrubbed some more, etc to get all the paper off the copper.  
Which is actually good because that means the toner is REALLY stuck to 
the copper and won't let go.  I've found a standard bathroom hand towel 
works for the scrubbing as it is abrasive enough to work the paper 
residue off without tearing up the toner.  And it REALLY saves my thumb 
a LOT of abuse.  Just keep the corner of the towel you are using wet and 
the board wet and rinse the paper residue out of the towel often.

Oh, and you don't have to get all the paper residue off the toner.  Only 
the bare metal needs to be completely clear.  That paper residue WILL 
come off when you hit it with acetone after the board is etched.  I 
still try to get most of it off anyway, just because...

Good luck.

Clint

On 11/29/2009 7:49 PM, awakephd wrote:
>
> Howard,
>
> I also got some ink from the catalog paper onto the copper when I used 
> that -- but it didn't seem to interfere with the etching. To answer 
> your questions:
>
> 1) I clean the blank pcb with soap, water, and scotchbrite; then clean 
> again with acetone.
>
> 2) I am still experimenting with the laminator, so I don't have a 
> recommendation on number of passes. My laminator allows setting the 
> temperature and the speed at which it feeds through, so there are a 
> number of variables to play with!
>
> 3) When I used catalog paper, I think I soaked it for 20-30 minutes at 
> the very most. When I used the glossy, I soaked it for 2-3 hours 
> altogether, trying to get it loose!! But as I said before, in the 
> little bit of experimenting I've done with the laminator, I was able 
> to peel the paper off immediately after the final pass through the 
> laminator -- no soaking at all! It remains to be seen whether I can 
> duplicate that result without losing traces ...
>
> 4) I did not download a driver for the HP, or at least not manually -- 
> it apparently is included with my distro (currently ubuntu 9.10, but 
> it has been included at least from 8.04, which is what I started 
> with). My distro also automatically recognized my old HP PSC-750 
> all-in-one inkjet. However, I have had to download a driver for a 
> networked Xerox printer/copier at the office -- as I recall, I just 
> did a google search for ubuntu and the model number and came up with a 
> driver.
>
> As for pcb software, I am running Kicad. It is available both for 
> Windows and for Linux. Here are a couple of links: 
> http://kicad.sourceforge.net/wiki/index.php/Main_Page 
> <http://kicad.sourceforge.net/wiki/index.php/Main_Page> or 
> http://www.lis.inpg.fr/realise_au_lis/kicad/ 
> <http://www.lis.inpg.fr/realise_au_lis/kicad/>
>
> Andy
>


[Non-text portions of this message have been removed]