While searching around I saw the 8 degree LEDs from LEDTronics as
well. I decided to stick with the 20 degree leds.
I had done a quick calculation for the amount of undercut of 10 degree
half angle versus 8 degree half angle. This is not taking into account
the affects that index of refraction would have.
For .0015 resist with 10 degrees gives an under cut of .26 thousands
of an inch. A 4 degree half angle gives .1 thousands of an inch undercut.
If I could find an index of refraction for the photoresist, then I
could take that into account as well. The undercut would be a little
less since the index of refraction(IOR) of resist is greater than the
IOR of air. If we were taking glass, the angle for 10 degrees would
reduce to about 7 degrees. Since I expect that the IOR for resist is
less than glass, the angle would be somewhere between 7 and 10.
After going through all this, my conclusion is that unless I
eliminate other things that cause scattering such as glass that is not
optically flat and better masking materials, I would not reach the .26
thousands undercut of the 10 degree, much less the .1 thousands of the
4 degree.
So when I do build my UV LED box, I personally will stick with 20
degree viewing angle LEDs.
--- In
Homebrew_PCBs@yahoogroups.com, Markus Zingg <homebrew-pcb@...>
wrote:
>
> Just as a note, you don't need a UV-LED based box for the resist. I
used
> it for years with an exposure box using UV tubes. The advantage of UV
> LEDs over tubes IMHO are:
>
> - LEDs have a 20 degree (or less, see below) viewing angle whereas
tubes
> radiate all around them. Alas you get more paralellized light
> - UV tubes are MUCH stronger. Exposure time was here < 10 seconds.
Thats
> a problem if the tubes do not start all at the same time cause then not
> all areas are evenly exposed.
>
> I can say that since I'm using the LED box, I can do finer artwork. I
> already did sucessfully a board with 5mil traces (GigE switch, the
> signal pairs from the PHYs to the RJ45 transformers) Whereas I've not
> been able to go that fine with the previous setup. Exposure time with
> the LEDs is pretty much exactly one minute and obviousely the LEDs
go on
> and off all simultanousely. Exposure time of course varries a great
deal
> depending on the distance and quantity of LEDs used.
>
> The disadvanteage of the UV-LED based aproach is that they are
> relatively costly. I used 80pcs (20x20cm squares, Kingbright L-7113UVC
> in my case). Make sure to get a datasheet and check that the peek
> radiation is around 400nm. Also look at the viewing angle and use some
> trigonometric calculations to find out how many LEDs you need at what
> distance to cover the area your unit is suposed to work with.
>
> It apears like the general distributors (Mouser, Digikey etc.) in the
> meantime no longer carry UV LEDs due to the dangers involved with using
> them. Ledtronics does. However they want you to sign a form where you
> state to have read and understood the dangers involved and that you do
> not hold them responsible. They have a fairly interesting LED (
> http://www.ledtronics.com/ds/L200CUV395/ article L200CUV405-8D
> <dsdc0304.pdf>) whoes viewing angle is only 8!! degrees. However,
such a
> narrow beam means even more LEDs. This one is $1.15/pp, minium oder
> quantity 100. I'm tempted to try them :-)
>
> HTH
>
> Markus
>