Homebrew PCBs

Volkan Sahin wrote:

> 2.5 microsecond value is both calculated and experimentally verified. 
> Here is the calculation
> Dry film requires 10-15mJ/cm2, let's take 15mJ (because of wavelength
> of blue laser is 
> 405nm). 720 dpi resolution corresponds 35micron-meters spot size,
> actually it should be 
> larger than that since on a continuous line there shouldn't  be any
> discontinuity. 
> Let's assume it is 50 micron-meters and  laser output power=150mW. 

So 15mJ sensitive resist and 150mW laser instead 60mJ and 85mW that's
the difference. 

> BTW, this 8000dpi is it equivalent resolution or real hw resolution?
> If it is equivalent 
> resolution it doesn't have so much sense. 

The machine i used i don't remember to see 8000dpi on the control
software, but the actual listed model says 8000dpi. But is just some
spec they trough. Because is an 'analog' machine it pass the film
continually, if can go slower and less power on laser the user can get
the impression of higher resolution.  We had to choose laser power for
each resolution and not the dot size. But these is a stressing
irritating noisy crap rotative machine at prices under cheapest cars. To
print on films that needed to be developed, fixed (bath), washed (in
darkroom), dried and only then used. But nobody uses more than 4000dpi
and who uses it must be for 1mil traces or is crazy. 1000dpi for
soldermask and 2000dpi for copper layers is just fine. End quality is
more about resist application, exposer light, developing and etching
solutions/machines, not film resolution. That machine take 1h to print a
2000dpi 20x30cm artwork, less time if big ground planes and less tracks
(the print traces and not dot thing).
Sub 10minutes machines are really expensive. Try google, there are
rotative mirror ones and raster lcd ones. Right now i am see one that
claims 128.000 dpi (on film)!! The lcd raster method can do things like
what is the antialising fonts in ours lcd displays. Again is a 'analog'
effect that seems to increase resolution by changing power around
corners and so. Also like epson dot size inkjet printers. The artwork
'body', whatever the resolution is, can only be super dark.
Makes me remember stupid ideias i was having 1 year ago after calculated
8hours printing time with the only uv laser led i found (<5mW). I
thought off rast scan a small monochromatic lcd screen like used with
8bit mcus or even a old cell phone lcd, with a uv source behind, but the
light could not heat the screen and uv led wasn't dense enough (or
collimated) , then i thought using a old school transparency projector
with a laptop screen and hundreds off uv led and move the pcb x,y above
the lens. Then found your direct print work, still impossible to do
soldermaskwith inkjet (the only epoxy based thing i found had mixed a
strong solvent used as epoxy remover), and needed metalic resist for PTH
boards. But you addressed this with the laser. Dip coating things like
aq3000 and print with your machine and use acrylic based white ink for
legend in the inkjet is many years ahead most boardhouses.


> Here is a simple demonstration of the problem,

Thanks for the effort but where you edited your email (seems yahoo
webmail) just sucks for ascii art. yahoo just s...  i can't even write
my name here, it's Sim�o with the ~~~~~ but you will get it wrong. 

> > I think it is a good idea, as far as I
> > know Sun Microsystems used similar concept in their laser printers.
> They were using VGA port as far as I know to generate/transfer raster
> image.
> I don't know the implementation details.

Could you say the machine model?