The black ink in my old Epson 660 is extremely UV blocking. I can expose
a PCB over 5 times normal and still not manage to effect the dark areas.
The problem of exposing his length of time is that thin traces become a
too narrow. The Epson 600 black ink (dye based) looks deep red through a
bright white light. I have just recently tried my ($100) Cannon S300 but
with Calidad black ink refill. I have made one PCB so far, but not
measured it's UV blocking ability. Also in the driver settings,
selecting a media for a Photo paper then the printer mixes color inks
with the back ink.
Have you looked at the transparencies under a microscope ?
Does the 1200x2400 dpi printing really show up sharper than 600dpi
printing from the Cannon S300 printer ? I found the line edge jaggedness
from inkjet printers, renders resolutions above 1000dpi meaningless.
mpdickens wrote:
> A member of another mailing list I am a member of
> found the following in a archive. Further, he tested
> and it worked as advertised:
>
> I have been dealing with a development effort for a
> circuit that my company is developing. In order to get
> fast turnaround of boards for testing, I needed a way
> to make high quality circuit boards (multilayer) in
> under 8 hours. Cost from commercial board houses for
> 24 hour turnaround was in the range of $2000-$3000 per
> design. In my case, I had to also develope plating
> systems and through-hole activation, fast etching, and
> a hot 20 ton press which I built by converting a shop
> press from harbor freight and adding a temperature
> controller and heating elements. etc.
>
> For ATM purposes, 2 sided boards can be made for a
> minimal expense.
>
> Because many on this list make their own circuit
> boards on occassion (for stepper circuits and camera
> circuits), I thought I would share my experience
> with the group.
>
> I am currently producing 4, 6 and 8 layer circuit
> boards using equipment now in my basement. Granted my
> basement looks like a chamber of horrors, but I
> suspect this is true for many on this list. Eight mil
> traces and lands are now easily doable and I am
> holding +/- 2 mil registration.
>
> The greatest roadblock to producing good circuit
> boards was getting good artwork on a transparency. In
> that regard, I have made several discoveries which are
> not immediately intuitive.
>
> First, getting really good artwork for the spec above
> is not possible with a laser printer. Phase error
> creeps in and even for printers claiming 1200
> DPI the accuracy just isn't there. I tested this with
> several models of HP printers including the 2000
> series and the 4000 series.
>
> In addition, the toner is just not dark enough. You
> end up having to underexpose the photoresist in order
> to get good removal and then you have a problem with
> undercured photoresist that will not tent over holes
> and whose sides are weak. Further the developing
> process just trashes the underexposed resist.
>
> I finally decided to try an inkjet printer. After
> some research looking for a printer that supported
> high resolution in black, I purchased a Canon.
> Initially, I purchased the S300 but it turned out that
> clever marketing made is sound like it supported high
> res black. In reality, the black was only 600 DPI
> like every other printer... Not enough resolution. I
> then tried the S800, which did support 2400 x 1200 DPI
> in color and in Black - the only printer that
> supported high resolution black printing. Experiments
> with
> this printer unfortunately revealed the problem that
> most people have with bubble jets. The black is
> simply not dark enough in UV. This despite the fact
> that it was a pigment based ink.
>
> I did have moderate success stacking tranparenies.
> This allowed me to increase the exposure time, but
> because only the first transparency was ink down (the
> second had to have a full 5 mil separation for the
> thickness of the first transparency, the edges were
> not very clean.
>
> I then had a brainstorm, I realized that my UV filters
> for my flourescent lighting were amber. I decided to
> try other colors... I quickly discovered that yellow
> was just as dark (in UV) as black. Disappointed that
> it was not darker, I began thinking about ways I could
> change the formulation of the ink to include a
> coreactive UV blocking chemical. I started searching
> the net when I discovered that ink fading as a result
> of UV is a real problem for photography. To my
> surprise, my printer already contained an ink that
> was UV blocking. All I had to do was tell the printer
> that it was printing on high resolution photopaper.
> This automatically switched cartridges to the PC
> (Photo Cyan) and PM (Photomagenta). Yellow remains the
> same because yellow only fades to yellow.
>
> In any case, once I did that, I was able to fully
> expose the Photoresist. In comparing a foil blocked
> section and a photo ink exposed section there
> was little difference. Moreover, in testing artwork
> created by a real photoplotter (costing $200,000).
> There was no difference. The only difference was that
> I settled on "GREEN" as being the color that was best.
> This selected the darkness of yellow in UV and the
> chemical UV blocking in Photo Cyan to produce a very
> dark black in UV and a pretty green in visible... :-)
>
> Perfect exposures! That along with unbelievable
> resolution of these printers make for a killer
> combination for producing your own artwork and
> consequently your own circuit boards.
>
> The bottom line is this. You DON'T want a printer with
> a dark black! Forget whether it is pigment based ink
> or dye based ink. That is all irrelavent, none of them
> are going to be dark enough.
>
> You want a PHOTO printer with PHOTO ink. Further ALL
> photoprinters have high resolution in color! Even the
> cheap ones ($100)! Just make sure a photo ink is
> available either from the manufacturer or for an ink
> refiller. All photo ink is, is ink with UV blocking
> added so the photos you print don't fade.
>
> What will the photoplotter companies do???
>
> Armed with this information, there is no reason
> everyone on this list does not do steves killer mod
> for the Philips Vesta camera or the many circuits
> for telescope motorization and tracking.
>
>
> Best
>
> Marvin Dickens