Homebrew PCBs

Hi,


Bit of a ramble -- sorry about this...

I've recently developed a bit of a problem with my etching process and 
I'm hoping that someone can give me some suggestions for improvement.

Up to now I've been etching FR2 boards at roughly 10/10 (10 mil tracks, 
10 mil spaces) with good yield. I've recently moved to FR4 substrate and 
8/8 rules in order to make a board for a slightly higher-speed design, 
and yield has now dropped to zero. I've uploaded a couple of pictures of 
my latest effort -- they were taken with a camera phone so they may not 
be very useful. I have been experimenting with UV exposure times: the 
top side of the board was exposed for 2 mins 45 secs; the bottom side 
for 2 mins 30 secs. I normally expose for 2 mins 15 seconds but this 
doesn't seem like enough for the FR4 board.


Top layer: http://www.atom.net/stuartw/board-top.jpg

Bottom layer: http://www.atom.net/stuartw/board-bottom.jpg


The details:

Artwork: transparencies printed at 1200dpi on a Canon i560
Board: Farnell FR4 double-sided photosensitised
UV exposure unit: home-built, 2x15W tubes approx. 5cm from the subtrate
Developer: Farnell KOH/NaOH pre-mix
Etchant: Farnell FeCl3 crystals made up to near-saturation

All the chemicals are less than one week old. I've used them to develop 
and etch three Eurocard (160mm x 100mm) boards so far. My process 
involves the following steps: drill some alignment holes in the board, 
remove the protective sticker from one side and expose it, then remove 
the sticker from the other side and expose that. I then develop the 
board in a tray of developer solution at room temp while manually 
agitating the board and turning it over frequently in order to try and 
develop both sides evenly. After about 2 mins 30 secs I wash and etch 
the board.

The problem is that I'm getting regions of board that are clearly 
over-etched (e.g. top right corner of the top side in the picture above) 
and other regions that have barely etched at all (left hand side of the 
same picture). The under-etched regions emerge from the etchant stained 
orange, which seems odd. I suspect that the over-etched regions occur 
because I'm etching for too long as a result of the other areas which 
refuse to etch.

I'm going to put my (home-built) bubble etcher back in to service this 
evening to see if it improves matters -- so far I've been manually 
agitating the boards in a bath of etchant.

Can anyone suggest any likely causes of this problem? I'm willing to 
accept that my process may never be good enough to produce boards at 
8/8, but the fact that I can no longer reliably etch even heavy tracks 
is depressing!

Incidentally, I'm not at all happy with the quality of the Farnell 
(Newark in the US, I believe) FR4 board -- I have found scratches and 
areas of missing photoresist in the latest batch. Has anyone else had 
this problem?

Thanks for your help.


Stuart