design of a UV exposure box / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

design of a UV exposure box

1 2 next

Move Thread

LAN_403

hboy007

Fri Feb 26 2010, 04:37PM

Registered Member #1667 Joined: Sat Aug 30 2008, 09:57PM
Location:
Posts: 374

Hello everyone,

I have started with the design of a new lamp reflector for my homebuilt UV exposure box. I built my UV unit into the housing of a flatbed scanner using two 8W replacement tubes connected to a driver board of a compact fluorescent lamp.
However, I recently became very frustrated when I tried to etch some finer wire widths because the direct light from the two tubes covers a large solid angle at each point of the layout (and exposure uniformity could be better).

Question 1) has anyone come up with a reflector design to alleviate at least some of the drawbacks imposed by a flatbed light source design?

Ok, so here is what I have in mind:
First of all, I have to find a suitable reflector design, the all-time classic would be a parabolically bent reflector, but I think there are better solutions because the area below the tube is not handled and the tube has a finite radius which also broadens the incident solid angle on the surface of the box. The Reflector surface needs to be intersected by absorber plates that segment the reflector and reduce the angular stray in the direction along the tube.
Secondly, I think it is a good idea to add a honeycomb absorber 1cm or a little more away from the glass surface.
Every other optimisations would aim at compensating for the light loss caused by the light former structures. A good part of the light could be recovered by using structured absorber plates that reflect in a preferred direction (I doubt this is worth the effort).

Let me conclude this initial post with a little video of the model I have calculated yesterday

Any replies are much appreciated! Thank you.

Steve Conner

Fri Feb 26 2010, 05:57PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

You did put your artwork with the toner side against the photoresist, right?

If you put it the wrong way round, acetate side against the photoresist, then that's when the diffuse light starts to cause problems.

If you have to put it acetate-side-against to make the PCB come out properly, then you've printed it backwards.

radiotech

Fri Feb 26 2010, 07:27PM

Registered Member #2463 Joined: Wed Nov 11 2009, 03:49AM
Location:
Posts: 1546

I have seen tubes coated with a fine uncoated strip casting a line. Focus is impossible because the arc is not a filament. I dont know how laying an adhesive fine line (or perhaps 2 for fringing) on the tube and airbrushing a highly uv reflective coating in the tube would work. when peeled off it might work. Perhaps a premade gap shroud or strip exists?

hboy007

Sat Feb 27 2010, 01:58AM

Registered Member #1667 Joined: Sat Aug 30 2008, 09:57PM
Location:
Posts: 374

Thank you Steve for your concerns. Luckily it's not the first time I make circuit boards and I have gathered sufficient experience to know what's going on in my box when good common sense fails me. I've seen eveything, starting from blur caused by fine copper burrs that remained at the edges from cutting (which made me sand and clean every photopositive board before and again after removing the protective tape) and extending to abberations such as ghost images originating from lower angle internal reflections of the glass sheet. I can also observe strong anisotropy in the quality of edge resolution because the incident angle of photons emitted lengthwise and across the diameter is quite different.
My approach to investigate an apparently simple topic should not suggest that I lack the practical experience (I know you try hard to forgive me my amplifier simulations in the first place

) but I want to toy around with my programs and push my primitive raytracer code to give me a profound idea of what's a good idea to cut out of cardboard sheets (not that I already knew parabolic reflectors and a few light shaping absorbers will be as good as it gets, anyway) and aluminium foil and I enjoy sharing with you.

hboy007

Steve Conner

Sat Feb 27 2010, 09:34AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

OK, well maybe you've just come up against the resolution limit of DIY PCBs. :P

Remember that there are also chemical limits to the resolution. The developer and etchant "undercut" somewhat, so tracks come out thinner in the copper than they did on the artwork. Maybe you can improve this by playing with temperatures and times, or maybe you can beef up the tracks in your artwork.

But anyway, it means that it's probably not worth putting *too* much effort into the optics. I've seen a few commercial PCB exposure boxes, but never one that had anything like the optical complexity you're proposing. I was under the impression that SMT components killed DIY PCBs, and anyone still making them now used Press'n'peel or the toner transfer method, rather than UV.

I've seen fluorescent lights for offices that had parabolic reflectors, and a kind of segmented grille that fit over the front made of chrome-plated plastic. It seemed similar to what you described, and the grille seemed to do a good job of what you called "reducing the angular stray". Maybe you could salvage one of these.

Also, photographers debate about the differences between diffusion and condenser enlargers, I guess this is the same issue really.

radiotech

Sat Feb 27 2010, 07:22PM

Registered Member #2463 Joined: Wed Nov 11 2009, 03:49AM
Location:
Posts: 1546

Holophane fixtures have a unique diffusion technique. The incand types had a series of wavelike mirror strips. The fluor type are still used.

hboy007

Sat Feb 27 2010, 10:27PM

Registered Member #1667 Joined: Sat Aug 30 2008, 09:57PM
Location:
Posts: 374

I just incresed the number of monte carlo experiments to obtain a scatter plot and made a histogram of the projected count rates

OK, I guess I just focussed the tube onto the top plane. the double peak may be caused by the tube casting a shadow in its own projection. This is so much fun to play around with!

@radiotech: I hope I have figured out what they built in the next few days, thanks for the hint

...

Sat Feb 27 2010, 11:00PM

Registered Member #56 Joined: Thu Feb 09 2006, 05:02AM
Location: Southern Califorina, USA
Posts: 2445

As to the uneven exposure--could you not just print out a transparency with a gradient that is the inverse of the current radiation pattern, and put it between the tubes and the board? It looks like you are quite adept at simulating the pattern, so it shouldn't be too hard to get a decent filter made.

For the rays not being parallel-honeycomb would help, are you sure you aren't just using too thick a layer of resist? I will admit the only time I did a UV board I used the sun as my light source (which didn't have either of these issues

), but it seems like you only need a few microns of resist, so unless you have 10micron wide tracks you should be able to get away fairly easy.

mikeselectricstuff

Sat Feb 27 2010, 11:36PM

Registered Member #311 Joined: Sun Mar 12 2006, 08:28PM
Location:
Posts: 253

The easy answer to get uniformity is to increase the distance to the tubes, and if necessary use bigger tubes or more of them.

Adjust by puttting a sheet of paper on the glass and playing with tube spacing and distance until the illumination through the paper looks reasonably even.

The second thing is to use a silicate based developer, not NaOH, as this gives a very wide margin, so uneven exposure is less of an issue.

Track/gap of 10 mil (0.01") is easily achievable with good board, good artwork (laser print on tracing paper) and a simple non-reflector UV box.
See my page

for more info.

hboy007

Sat Feb 27 2010, 11:51PM

Registered Member #1667 Joined: Sat Aug 30 2008, 09:57PM
Location:
Posts: 374

I am currently running some monte carlo simulations with 500.000 events and play around with different curvatures
(I should jugde the result by the variance among the histogram bins but for now, looking at the scatter plots is fine) . The hot spot can be reduced by repositioning the tube a little higher or lower, even 5mm can make a big difference. I guess a combination of a honeycomb layer with a corrective filter, either near the tube or 1-2 cm below the glass cover will do the trick.

I generally use "Bungard" boards that come with a high quality 5Âµm photoresist coating and are very tolerant to inhomogenous exposure and chemical bath concentrations / residence time. it is almost impossible to ruin the PCBs in the subsequent steps. It's just my UV unit that fails to do a good job.

I have the impression that I lose about 50Âµm around the edges depending on the distrance from the lamp and that is far from being satisfactory.

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob