Electronics > Projects, Designs, and Technical Stuff
Building a UV light box for PCB production
(1/2) > >>
CharlieWorton:
Hey, everyone - I'm building a UV LED based light box for exposing Riston - type films, as one step in the production of PCBs.  I started by choosing what I hoped would be the correct LED; 365 nM LEDs, each one rated at 3 watts.  I wound up purchasing 20.  I arranged them in 4 rows of 5, and came up with what I thought would be a cool idea: use a 200 X 300 mm PCB as the base and etch the circuit, polish the copper until it shone, and then tin plate it.  The idea was that the PCB would serve as a reflective base, as well as being easy to solder.

I laid out the circuit board image using Microsoft Publisher (a desktop publishing program that has been largely abandoned by Microsoft).  I started out by trying to transfer the image using toner transfer; that didn't work very well.  I spent a lot of time trying to rescue the transferred image using felt tip markers, but ultimately gave up.  Then i had a cool idea.  I took some cheap self adhesive shelf liner I had laying around (they used to call this stuff mactac in my youth; vinyl with a pattern printed on one side, adhesive on the other, and a removable sheet to protect the adhesive) and I cut a 220mm X 320mm sheet of it.  I then fed that through the laser printer, printing directly on the light colored vinyl sheet.  Once printed, I removed the backing paper and stuck it to the PC board.  Then I took a sharp knife, a ruler, and all my patience; and i carefully cut away all the vinyl sheeting where i wanted the copper to be etched.  This worked out quite well; I etched the entire board, and got a great result.  Once the etching was complete I washed and dried the board, stripped away the remaining vinyl... and polished the copper until it shone, using a metal polish called autosol metal polish.  This stuff is magic; it will put a mirror finish on virtually any metal.  Here's a manufacturer link: http://www.autosol.com/product_p/1000.htm   You can actually find this on Amazon, and I highly recommend this stuff.  Best metal polish I've ever used.

Once i had it polished, I wanted to tin it.  Tinning that huge board with a soldering gun would be a really slow show.  There's some stuff called Liquid Tin from MG Chemicals (a proud Canadian company) which works well, but they want about $100 CAD for 16 fluid ounces of the stuff.  So i did some digging, and i found a YouTube site called NurdRage where he comes up with chemical approaches to simulate existing products... and he had a formula for liquid tin!  So I tried it, and it worked.  Took me a few tries to get the plating effect I was looking for, and it still isn't gorgeous, but it is utilitarian.  It will work.  Here's a link to the YouTube video:

Then I spent some time soldering all the LEDs in place, checked my work, took a deep breath, hooked it up to my power supply and... success!  Yea!

Here is the board after it was soldered, before i applied power:

And, here all those tiny UV LEDs are all powered up, and singing in unison:

...and the Murphys Law Gods have been testing my patience for the last hour, trying to get this post posted... so i think I will now retire, with as much grace and good will as I can muster.  Enjoy, folks.  I leave in peace.  And pieces.  >Charlie
sleemanj:
Looks good.  That is a LOT of UV.  Your exposure time will probably be measured in single digit seconds if you run that full blast.  Don't look into the light!
Miti:
I wanted to do something similar years ago but then I discovered PCBWay, 10 pcbs for $10, then they reduced it to $5. In my opinion it’s not worth the effort and the chemical danger if I can get 10 professionaly build pcbs for $30 in about a week, or for about $15 if I’m willing to wait for about a month. Or at least I could before the virus.
Zero999:
It's questionable whether UV is required. The photosensitive film on PCBs can be activated visible violet in the 400nm to 430nm range. I don't know if actual UV is better, but LEDs are cheaper and more efficient at those longer wavelengths and it's safer too.
CharlieWorton:
Hey, Zero999 - DuPont specifies the peak photosensitive frequency of Riston at 365 nM.  The curve is pretty steep on either side - if i recall correctly, 400 nM is down several orders of magnitude in sensitivity.  Florescent tubes will work, but only because they're putting out some 365 nM radiation.  All the visible radiation is having almost no effect.  Here's a chart from Dupont, showing how dramatically film sensitivity falls off on either side of the 365 nM optimization point:



Source: "Safe Lighting for Riston* Photopolymer Dry Film Resists" at https://mafiadoc.com/safe-lighting-for-riston-photopolymer-dry-film-resists-dupont_5b6d1882097c47d12c8b45ca.html

I agree with the general 'safety first' theme of many of the responses.  But, as a former lab tech and process control tech, I believe in optimizing everything - and that includes the exposure.  The finished box will be designed so that all radiation remains in the box.  This optimizes - minimizes - the exposure.

Thanks, everyone - Charlie
Navigation
Message Index
Next page
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod