Electronics Developing PCB's with UV CCFL Possible?

I have been experementing with with developing PCB's using the UV Exposure Method. But to keep costs down and to be able to control the UV lights electronicly (I plan to build a timer for the lights later on) I chose to use CCFL lights since they run on 12V. I have tried several times to expose the copper board to the light but have up till now been unsucessful. The results that I got were always very faint, and sometimes very bad on one side of the board. The tracks just didnt seem to be properly defined. I am not sure if the developer is not strong enough or I am not leaving it under the light for enough time. I am using precoated photo etch PCBs under 4 CCFL lights running off 2 inverters. But recently, not sure how, both inverters stopped working (I have a feeling I blew the transistors).
I am asking if anyone has tried this method with cold cathode lights and if you have any advice? I need to know whether it is worth buying another set of CCFL lights.

 

CCFL blacklights don't emit the shortwave UV that this requires.

You can use an exposure box made with an array of small standard household fluorescent tubes, which is very commonly done and works well, or use a high energy UV-erasing short-wave tube - note that you'll need to keep the UV light confined within the box if you use these.

Or, depending on your location and climate, expose them in teh sun.

 

And then you can wire the lights to any normal house hold light timer.

 

the kind of uv bulbs that you need are commonly used in uv sanitizing lights. The kind of UV that uv cathodes emit is pretty close to regular violet light, so it may cause things for fluoresce, but its not the kind you need.

 

Thanks for all the help. Looks like ive been wasting alot of time and money on the CCFL lights.

 

I've seen PCBs developed with blacklights with OK results. You get bad edges if your design isn't pressed perfectly against the PCB. Even small gaps will ruin the results. Try applying a lot more pressure evenly across the board, and make sure your glass is not bent at all. (And that it doesn't filter UV!!)

 

Also, what are your exposure times?

A good idea is to make a test pcb by placing a cardboard baffle over the pcb and progressively removing it over the course time (however long of an exposure you're using, plus a little extra).

The end result is a pcb with graduated exposure that will give you an accurate idea of how long you have to develop it for.

 

I think the real problem here is amount of UV, not the wavelength. PCB resist (according to Wikipedia) is sensitive to the 'A' type UV that's fairly harmless as well as the shorter waves that kill stuff.

A UV CCFL is under half the wattage of a standard 8W 12" exposure box tube, and I bet the amount of UV is much less than half the proper tube, I doubt they use expensive phosphors in a cheap CCFL. Try much longer times, keep doubling up exposures on a step-wedge.

 

Buy one of these tanning devices at a local scrap yard (I bought three of them for 5-6 Euro each):

They contain 4x15W Cleo UV lamps, the lamp starters, lamp sockets and two ballasts.

Find an old scanner (with suitable room inside the housing):


Dismount the tanning device:


Empty the scanner housing:


Mount everything in the scanner housing (I made new aluminium L-profiles to hold the lamp starters and sockets):


Recable everything (I added my PIC16F84 countdown timer):


The cover with the keyboard/display board and the original scanner glass covered with diffused 3M-film


And voila, a cheap and compact UV-box:




CD

Edit: Added one more picture.

 

wow thats a real nice idea, it would make a cool eprom eraser as well.

 

But then you need UVC (short wave germicide) UV tubes.
The Cleo 15W tubes are UVA.

CD

 

Wow thank you very much for the help. I have been trying every now and again since Easter and have never got any good results. I used exposure times up to 10mins. Most of them came close but I never got a clear image. I will start looking for tanning device and give that a go. I will keep you updated with my progress.
Sami

 

Nordic's design for a PCB light-box using UV LEDs.

More info and LED specs here.

 

Finest trace that I can etch successfully (circle with the cross to indicate a drill hole):



I estimate the line width at about 1 to 2mil (0.0254 to 0.0508mm), the copper traces on the PCB are 20mil (0.508mm)
The left drill hole is fuzzy because I focussed onto the right one (640x480pixels, 200mm objective at 25cm (10") distance in macromode).

- 180 seconds of exposure (4x15 Watt Cleo lamp from Philips tanning device in scanner housing)
- 45 seconds of development (bag of development powder in 1 liter of water)
- 12 minutes of etching in FECL3 at 50-60°C.
- 15 minutes of chemical tinning

CD

 

I use a halogen bulbed Overhead projector and get very accurate results. These things are easy to find cheap, in fact I got mine free of charge.

A 650w bulb has an exposure time of 5 1/2 minutes... roughly.

[/IMG]

 

WOW! That PCB is completely totally flipping amazing!!! I plan to set up a proper PCB developing center in my shed after final examinations. My problem is that I use an overhead transparency that has been treated to allow inkjet printers to print on it and the transparency isn't 100% clear. The transparency has a sort of fuzziness from the layer that absorbs the ink. As a result, the exposed PCB doesn't react completely and properly with the light. Should I exposed it longer? Or maybe use a laser printer with a better transparency? My friend recently purchased a colour laser printer (I'M SO FREAKING JEALOUS!), perhaps I can use his.

Sputnik

 

I found no problems with using inkjet transparencies. Though I did experiment a bit.

If your getting fuzziness it is perhaps your printer settings that are wrong. Check its set for transparencies. The HP printer I use has a setting for HP premium transparency (from memory) which I find gives the most accurate result. I use a higher quality transparency sheet from my local art shop.

Other things you can try...

1. Print the image several times over the same transparency. I found with my HP printer I could print 3 times onto the same sheet. Makes the lines very black. I use a HP5550 which seems to be able to consistantly print in exactly the same place on the sheet. Make sure you give each print some time to dry before overprinting.

2. Use more than one transparency. Its pretty easy to line them up. One way to do it is to have a 16 or 20 pin soic somewhere on your negative. Then if they are out of line you can see very easily. The traces turn into a black blob.

3. Do a test run on a scrap bit of photoboard. I use a set of different width traces in a straight line. I expose them to differing amounts of light and use the one that gives the sharpest image (sorry if this is basic advice, I dont know your level of experiance)

 

Thanks!

I use a laserprinter @600dpi and clear laser transparencies (overhead sheets).
The black traces on the transparency are not perfectly black (more a dark grey) after printing but I treat the transparency with Chinese ink (apply the ink onto the toner traces with an ear cotton, let dry, and wipe the excess ink with a moisted kitchen paper).
After this treatment the black toner on the transparencies is really deep black.

To determine the correct UV exposure I use a piece of PCB sheet that I expose for different times (ex. 30,60,90,120,180,210,240 seconds) by moving a piece of aluminumfoil every 30secs of exposure.
This is the result after development and etching:



In my case, you can clearly see that 180secs of exposure is correct with the Bungard PCB sheet (I'm not able to make decent PCB's with those f***ing Velleman sheets).

CD

 

I have an HP OfficeJet G85. When I print on transparencies, I set the paper type to transparency and increase the ink volume and all that stuff to lay more ink down. I'll see what other makes of transparecies are available tomorrow.

I wish I hadn't dismantled that old laser printer I fished out the trash years back...

Sputnik

 

laser printers + transparencys = melted transparency round laser printe drum.

in other words, dont mix unless you have heat resistant transparencys.