Homebrew PCBs

The distance from the board is easily adjusted. If it needs 2 inches
fine. If it needs 1.42, that is fine to.

Thanks for the pointer to the pdf. I will take a look at it.

--- In Homebrew_PCBs@yahoogroups.com, Adam Seychell <a_seychell@...>
wrote:
>
> Damon,
> If you do the maths, you'll find that intensity for an infinite
array of 
> infinitesimal small point sources gives constant light intensity at all 
> distances. In other words, your light array does NOT follow a 1/r^2
law. 
> If I were you I'll move it back to at least 2" so that light blends
from 
> adjacent LEDs.
> 
> These UV LEDs are narrow band, so its unlikely you will see reflected 
> light at other frequencies obfuscating true UV intensity.
> 
> Can you explain why the fluorescence from white paper will produce an 
> inaccurate intensity image projected from the UV light source ?
> 
> For the power supply, I've built one based on a LM3488, but you can 
> probably configure any boost converter IC as constant current source. 
> input is 8 to 20VDC, output 35V @ 400mA constant current, and is
capable 
> of driving 25 strings of 10 LEDs in series for total of 250 LEDs. I'll 
> send you schematics if interested.
> The LED array is based on National's app. note.
> 
>
http://www.national.com/appinfo/power/ledrefdesign/LM26988x4LEDArrayDriver.pdf

> 
> 
> 
> javaguy11111 wrote:
> > I am using the 395-400nm LEDs. Since there is a strong UV component to
> > the LEDs I am not sure if a visual test for uniformity is going to be
> > accurate. 
> > Since I am working in english units obvious I am working around 12.7
> > mm or .5 inch. Not a big difference. 
> > To really measure exposure I need stouffer gauge. I may go ahead and
> > order one from Think Tink at some point.
> > As far as distances, the further away the longer it takes. So
> > initially I will be going for 1.25 to 1.5 inches. Based on my
> > calculations from yesterday in another message it looks like 1.42
> > inches is the distance I want so that the half power points sum up
> > properly. Since these point sources and not infinitely sized, the
> > illumination will never be perfectly uniform.
> > 
> > My design at this point is leaning towards a 4x6 board, which with .5
> > inch spacing will use 96 LEDs. I will also get the board sized so that
> > I can add more if I need a larger size, but my largest board to date
> > is 2.5 by 3.5 inches. 
> > 
> > I am looking into also using LT3591 boost regulator for driving the
> > LEDs. Obviously it will take several of these to drive all 96 LEDs,
> > but I do not like the idea of current limiting resistors or just
> > putting them all in parallel. 
> > 
> > 
> > 
> > 
> > --- In Homebrew_PCBs@yahoogroups.com, Adam Seychell <a_seychell@>
> > wrote:
> >> I also got the 390-395nm BestHongKong LEDs. To test the exposure
time I 
> >> made a 3x3 square array as you did, but with 15mm pitch, not
13mm. To 
> >> look for uniformity, just shine on some white paper and see how much 
> >> variation there is on the reflection. I found an LED pitch of
15mm gave 
> >> fairly uniform projection at 100mm or more. A 30mm distance as
you are 
> >> talking about looked way too non-uniform to my liking.
> >> There is no big penalty going further distance, except the edges
of the 
> >> projected light will tapper off more slowly the further distance
away. 
> >> All this means is your array will need 1 to 4 LEDs in size larger
than 
> >> PCB artwork area to compensate for tapering intensity. Remember an 
> >> infinite LED array has constant intensity at all distances. So, it
> > beats 
> >> me why your trying to get the LEDs close as possible.
> >>
> >> The best way to test exposure times is to expose without any image 
> >> printed on the photomask. Then expose small sections at incrementing 
> >> times. I did this by cutting strips of photoresist film about 100 x
> > 20mm 
> >> , and drawing 4 lines equally spaced across the strip to give 5
> > squares. 
> >> I used some black plastic to mask out all but one square and exposed 
> >> this area for a recorded time. Exposing in small steps, e.g 
> >> 10,20,30,40,50 seconds, you can get very good estimation of time it 
> >> takes to fully expose photoresist. After development you will see
the 
> >> squares which are slightly underexposed will look different (more 
> >> damaged or loss of smoothness) than those 100% exposed.
> >>
> >> Only after doing that you should test artwork, because now you know
> > what 
> >> the minimum exposure time is.
> >>
> >> javaguy11111 wrote:
> >>> I did a little more testing of UV LEDs this weekend. The LEDs are
> >>> plugged into a solderless breadboard and driven with a variable
> >>> external power supply I tried a 3x3 grid with 1 inch and .5 inch
> >>> spacing. The LEDs were run at 20mA and pulled about 10V.
> >>>
> >>> The 1 inch spacing at 1.25 inches above the board showed too much
> >>> variation in exposure. I would get areas of overexposure while other
> >>> areas would not get enough exposure. I tried times of 1,5 and 10
> >>> minutes. 10 minutes was way over exposed, 5 not so bad and 1
left some
> >>> areas okay, but other areas underexposed.
> >>>
> >>> I then did tests with .5 inch spacing with the LEDs at 1.25 inches
> >>> above the board. Uniformity of exposure was much better. I tried an
> >>> exposure time of 1 minute. I think I can drop that down to 45
seconds
> >>> as I still saw a little extra exposure in areas directly below the
> > LEDs.
> >>> So I think for me .5 inch LED spacing is what I am going to use
to get
> >>> uniformity. I will do a few more tests of increasing the
distance to 2
> >>> inches above the board to see how that works.
> >>>
> >>> --- In Homebrew_PCBs@yahoogroups.com, "javaguy11111"
> >>> <javaguy11111@> wrote:
> >>>> I just did a quick test with my UV LEDs using a step pattern going
> >>>> from 10/10 spacing down to 1/1 spacing in thousands of an inch.
This
> >>>> test pattern was printed by an Epson 2400 printer using Pictorico
> >>>> transparency.
> >>>>
> >>>> The leds were 1 1/4 inches above the board with a .5 inch
spacing. I
> >>>> tested with 4 LED's running 2 in series and then 2 of those in
> >>>> parallel running 20ma at 6.6V. 
> >>>> Exposure time was 60 seconds. 
> >>>>
> >>>> After developing in sodium carbonate, I got good lines down to  3/3
> >>>> spacing. The 2/2 looked pretty good as well, but was a bit jagged.
> >>>> Possibly due to the 1440 dpi resolution of the printer. The 1/1
lines
> >>>> were all merged together. 
> >>>>
> >>>> I will try 3/4 inch and 1 inch spacing over the weekend to see how
> >>>> those work, before I commit to a spacing. I will post the
results of
> >>>> those tests as well.
> >>>>
> >>>
> >>>
> > 
> > 
> >
>