UV Exposure box

[akis]

I manufacture PCBs at home, and have bought and used this one from ebay

http://www.ebay.co.uk/itm/UV-EXPOSURE-UNIT-10-2-X-8-3-FOR-SCREEN-PRINTING-HOT-STAMPING-PAD-PCB-CURING-/201436726253?hash=item2ee69083ed:g:vgIAAOSwL7VWjwWa

It arrived half broken etc but after I changed the top glass and a blown capacitor, it has been working OK.

It uses 6 * UV lamps of 8W each for a total of 48W (G8 T5 lamp). I have a light meter which I do not trust very much. Anyway it shows around 1800 LUX right at the top of the glass where the PCB sits.

These lamps develop black spots at the ends or break completely and are expensive to replace and hard to find plus they are UV-C the ones I could find (I need UV-A for safety).

I was thinking of throwing away those lamps and replacing with UV LEDs - for example I have a 5m LED strip of 300 5050 UV-A LEDs - it draws over 3A at 12V. As it is wound on a bobbin and not laid out I cannot measure its luminance.

Has anyone done something similar and are these LED strips any good for PCB development? Or shall I stick to my 6 8W lamps, even if some are UV-C ?

[sleemanj]

I use an array of eight 1W UV leds in my current home-made box.  Works well.  But I'm not doing anything incredibly challenging, the odd 10mil trace to squeeze in a tight spot is as fine as I care to bother with.  However due to the fairly tight beam of a typical LED, even wide angle ones, my box has to be quite tall and in fact I have a double layer of diffusion between the leds and the platen (leds, then about 10cm above it is a glass sheet with a diffuser on each side, then about 10cm above that is the platen), otherwise there would be bad hot spots in the exposure area.

I have them arranged (over a slightly larger than A4 area) in a 3:2:3 pattern.

Code: [Select]

* * *
 * *
* * *

I've seen others use the LED strip you describe, I think people have said it works well for them, but again I'd be concerned about hot-spots with a short distance between platen and leds.

[Siwastaja]

Spot source is better for resolution than a diffuse source, especially if your vacuum system is lacking. I use 20W Ebay UV LED at about 40 cm distance and it works great, it's never the limiting factor for resolution. I don't go below 8/8 mils, even though 6/6 seems to be OK in a pinch.

[akis]

There is no much depth between glass and bottom - maybe 8 cm - I cannot use spot lights - whatever I use must be well diffused within 8 cm.

With the 5m LED strip that I have, I could cut it in lengths of 30cm and place it all across the bottom, that would be around 16 strips and they'd pull over 3A @ 12V, maybe as much as 4A. It would be a total of 300 LEDs.

But am not sure if the intensity will be enough. Will I get 1800 LUX or will I get less (assuming my light meter is not rubbish)?

[kurt]

I have had good luck with LED strips in my DIY UV exposure box. My build log is here: http://kavionic.com/blog/DIY_1188_LED_Dual_Sided_UV_Radiation_Unit

[sleemanj]

But am not sure if the intensity will be enough. Will I get 1800 LUX or will I get less (assuming my light meter is not rubbish)?

Does it really matter?  Whatever you get you will have to recalibrate your processes to the new light sources anyway with appropriate exposure length step testing.  If the leds put out less useful UV, it's going to take a bit longer, if they put out more useful UV, it's going to take a bit shorter.

I'd be more concerned with cooling, unless there is a fan in that box.  You know leds don't like getting hot, and it's really annoying to have them start to go on the fritz in a UV rig...
 

[akis]

Did not think about cooling but with the existing tubes, the LUX starts at 1000 when they are "cold" (room temperature) and increases to 1800 after around 5-8 mins warm up time. This took me time to discover on double sided boards, why was one side so much better exposed than the other...

Regarding the intensity, I believe the 6*8W tubes are quite strong, and even then, I need like 8-9 mins exposure times. I do not particularly want to be waiting 20 minutes per PCB side, that is why I was thinking will the light be adequate?

[sleemanj]

Regarding the intensity, I believe the 6*8W tubes are quite strong, and even then, I need like 8-9 mins exposure times. I do not particularly want to be waiting 20 minutes per PCB side, that is why I was thinking will the light be adequate?

That is very long for an exposure, well, for dry film it would be, I don't know what you are using.  I wonder how much useful UV your bulbs are actually putting out... not much maybe.

Even when I was using 5mm leds (about 100 thereof from memory over an A4 area, with considerable distance to get the coverage from the narrow beams) I would only have a 2m30 to 3m exposure (for dry film).

[jpanhalt]

@akis
You do not need or want UVC.   For one thing, ordinary glass blocks most of that energy.  If your tubes are clear and glow with a blue-white light, they may be low-pressure mercury.   More than 90% of their energy will be blocked by ordinary glass.  Second, photoresists contain photosensitizers that effectively increase the wavelength of light that should be used for the chemical reaction.

The emission spectrum from fluorescent lamps is dependent to the phosphor that is used to coat the inside of the tube.  You can look up what that emission spectrum is for the tubes you are using.

Photoresists for making PCB's with which I am familiar depend on UVA and are designed for 380 nm or longer wavelength light. Some (positive types) work well with ordinary fluorescent bulbs.  In any event, you need bulbs or LED's ( I use straight fluorescent bulbs) that emit light in the 380 to 405 nm range.   DuPont's Riston negative resists recommend 350 to 380 nm.  The shorter wavelength doesn't offer any advantages and is harder to work with. 

Collimated light is best for sharp images, but is more difficult to produce.   I just use three bulbs with a reflector behind them.  Do not line the box with reflecting material.  It offers no advantage and increases scatter, which is not good.

John

[DTJ]

Regarding the intensity, I believe the 6*8W tubes are quite strong, and even then, I need like 8-9 mins exposure times. I do not particularly want to be waiting 20 minutes per PCB side, that is why I was thinking will the light be adequate?

Sounds like a long exposure time. I've got 2 x 4W (IIRC) lamps in my light box build in an old pine wood box.
I get away with 90 seconds exposure time. I used to use a Kinsten commercial light box years ago at work and the exposure time was also under 2 minutes.

[ion]

Regarding the intensity, I believe the 6*8W tubes are quite strong, and even then, I need like 8-9 mins exposure times.

That doesn't sound too strong - my DIY box uses a pair of 25W tubes 15cm above the board and I get exposure times of 75 seconds, even the dry film soldermask only takes 3.5 minutes.  Perhaps there's something wrong with yours seeing as they're almost the same power rating.

Also, if you're having a hard time finding UV-A tubes, search for insect killer tubes - you'll probably get more results.

[akis]

I use a laser printer to print on transparent 110gsm paper - for better defined lines with no "spotting" I use two papers, in that case the exposure time is 8-9 minutes per side, if I use one paper then the exposure is around 5 minutes.

It may be that the tubes have lost their intensity or it may be that my developing process is not great. I try to control developing as best as I can with temperature between 22C and 24C and even then some boards take 1.5 mins to develop, others take 3-4 mins...

In the end however they all etch great.

[ion]

I use a laser printer to print on transparent 110gsm paper - for better defined lines with no "spotting" I use two papers, in that case the exposure time is 8-9 minutes per side, if I use one paper then the exposure is around 5 minutes.

I use 90gsm tracing paper.  I'm not sure if it has better UV transparency, but that may explain the difference.

[akis]

Unfortunately 90gsm stretches too much - 150gsm is best, but my printer has developed an aversion to it, so now I use 112gsm which is in between.

This stretching becomes apparent when you try to put two or four sheets together and the holes do not align properly especially for medium size boards (eg 4"x4")

The paper stretches as it's being printed/pressed/heated inside the laser printer.

[ion]

Interesting - I never noticed any stretching, at least not anything noticable on the largest (1.5"x2") boards I'd made.

Not likely to be an issue in the future either as a few months ago I noticed that if I take my inkjet out of economy mode, it matches the resolution of the laser I was using.  I would say the masks actually look better (although it may be the toner wasn't the best quality).
The exposures are pretty good too.

[akis]

Of course the inkjet does not heat the paper. in contrast with the laserjet which actually cooks the paper (too hot to touch almost).