UV tubes for PCB exposure box

54 views
Skip to first unread message

Luke Weston

unread,
Nov 2, 2009, 4:45:49 AM11/2/09
to connected-commu...@googlegroups.com
Kalex in Melbourne sell three different UV tubes; NEC 8W, 15 W or 20 W tubes, for between $10.50 and $13.75 each depending on what size you want.

http://kalex.net.au/Text/1098428399218-9745/pC/1098424907968-1192/

Mike B said those are the tubes he used in his light box, so we know they work well. I'm not sure what the rating on those tubes is, though? 15 W I think? (Mike, can you confirm?)

There are also these guys:

http://www.nelsonlamps.com.au/images/pdf/C%20Fluorescent%20lamps.pdf

There are two types of UV tubes that they sell - the visible/UV ones (listed for insect control) - these emit both visible purple light and UV-A, and hence light up bright purple - I'm pretty sure these are the same as the ones Kalex sell and the same as the ones Mike is using.

One minor disadvantage - which we saw on Saturday - is that, well, they attract bugs, as you might expect.

They also sell the "blacklight blue" UV tubes, which are the exact same thing as above, except that the body of the tube is made from dark purple Wood's glass which absorbs essentially all of the visible light but lets the UV through - creating a so-called "black light". These are exactly the same thing as your familiar blacklights at a nightclub, etc.

Given that it's the UV that we're interested in and not the visible wavelengths, both types of tubes should work equally well, I think, and the blacklight tubes should have the added (slight) advantage that they do not attract bugs.

It should be noted that all these types of lamps are emitting relatively long wavelength UV-A - meaning that they are not dangerous to the eyes or skin. Short-wavelength UV-C emitting tubes - such as those used for germicidal lamps or EPROM erasers - can be potentially dangerous to the eyes and skin, but these are not required nor desirable just to expose photoresists on PCBs.

Unfortunately I can't find any retail information or prices on the Nelson Lamps site.

michaelc

unread,
Nov 2, 2009, 6:25:20 AM11/2/09
to Connected Community HackerSpace
I've done a little looking into this today too. I've found a few
discussions about photo resist (though not about Kinsten in
particular) that suggest the optimum wavelength is around 350nm, which
is certainly in the UVA range. My understanding of the BLB tubes is
that they are less efficient than a BL tube because they rely on a
subtractive approach as you've described - most of the input energy is
not output as UV.

The Kalex tubes are obviously very well suited to the Kinsten. I guess
a good configuration for us would be two short tubes top and possibly
bottom to get a good spread of light over boards. How many ballasts do
we have in that light fitting, and would they be suitable for the 15W
tubes? We could always just build the bottom half first to see how
things go.

An alternative I've been looking at is the tubes used for an acrylic
fingernail paint dryer. They output UV at 365nm, so they'd be at the
right wavelength. The tubes are very cheap, for example these -
http://cgi.ebay.com/4-x-9W-Nail-Art-UV-Lamp-Machine-365nm-Lamp-Bulb-Tube_W0QQitemZ320442641124QQcategoryZ45205QQcmdZViewItemQQ_trksidZp4340.m263QQ_trkparmsZalgo%3DSIC%26its%3DI%252BC%26itu%3DUCI%252BIA%252BUA%252BFICS%252BUFI%26otn%3D10%26ps%3D63
- are USD$6.70 for a set of four including shipping. My little
exposure box is adequate at 22W although not fast (about 5 minutes
exposure) so I'm guessing that 36W per side would be fine. I've also
got my tubes mounted too far away from the board. I've read a
recommendation of 5 to 8 cm, compared to the 12 cm I've used. Those
tubes are PL-S tubes, which are a bit like a regular compact fluoro
tube but without a built in electronic ballast. They have a built in
starter, but if we bought these tubes we'd still need to buy ballasts
and G23 sockets, which would get expensive. Instead, it would be
cheaper to buy a pre-assembled nail dryer like this
http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=180415690012&ssPageName=STRK:MEWAX:IT#ht_4424wt_1167
at USD$30 including shipping. If you take a look at the photo you'll
notice it has an open bottom, so it might be easy to put two back to
back for simultaneous double sided exposure if we want that.

Here's the two options I'd suggest for single side at a time exposure,
and how I see them stacking up:

1. Buy 2 x 15W tubes from Kalex

- Base cost of $25.85, plus the cost and effort of building a box and
a timer with 240V switching. It may be considerably more expensive if
the ballasts or sockets we have aren't suitable or if we decide to
build another one for 2 layer exposure and we don't have enough
ballasts.
- We get the fun of building our own enclosure, electrics and
electronics.
- It will take us some time to build something nice.
- Proven to work very well with Kinsten - no risk.
- May be able to use some parts we already have.
- Not as easy or cheap for people to replicate at home, especially if
they don't have parts already available or the same tools as us.

2. Buy a the nail dryer from the ebay dealer

- Total cost of USD$30 - no extra parts to buy. Quite likely to be
cheaper in total than option 1.
- Replacement tubes are ubiquitous and cheap.
- Slightly higher power than 2x15W NEC tubes (though that's input
power - who knows about the light output).
- Instant gratification - we order from Hong Kong, it arrives in a
week or two, plug it in and go.
- Doesn't give us the the joy of building our own (but I assume we can
re-house the innards if we want to).
- Doesn't use the parts we have available now: possible recycling
fail.
- Slight risk that it isn't suitable for Kinsten (very low risk in my
opinion given the wavelength of the tubes).
- Built in timer and control electronics for the tubes - I assume
we'll easily be able to adjust the timer if 3 minutes isn't right, or
replace it with $1.50 worth of micro.
- Fully reproducible by someone who wants one for home without needing
any special effort, tools or tubes.
- It's pink.


Michael Collas

michaelc

unread,
Nov 2, 2009, 6:51:57 AM11/2/09
to Connected Community HackerSpace
I've also done a little research on the transparency of glass and
plastic to UV at the kind of wavelengths we need. The summary is this:

- Glass is very transparent.
- Polycarbonate blocks a lot of UV, and would probably be quite
unsuitable.
- Acrylic has a steep cutoff in transmission at a wavelength that
depends on the type of acrylic. At its most transparent, it will
transmit a bit more light than glass, but the 'wrong' type of acrylic
will attenuate UVA very sharply. For example, see the chart on this
page: http://www.cityplastics.com.au/lighttrans.html

Tomorrow I'm going to test out the acrylic that I used for my exposure
holder by doing a test exposure with both one and two layers of the
acrylic in the same test run. If it's blocking a lot of the UV, this
should show me.

Michael Collas

On Nov 2, 8:45 pm, Luke Weston <reindeerfloti...@gmail.com> wrote:

Mitch Davis

unread,
Nov 2, 2009, 7:04:09 AM11/2/09
to connected-commu...@googlegroups.com
On Mon, Nov 2, 2009 at 10:25 PM, michaelc <michael...@gmail.com> wrote:
>
> I've done a little looking into this today too.

Indeed. quite a write-up.

> exposure) so I'm guessing that 36W per side would be fine. I've also
> got my tubes mounted too far away from the board.

Possibly, although your use of alfoil on the sides means the light
doesn't really have anywhere to go, so the r² law won't be so bad.

Yesterday at CCHS we saw a demonstration of what happens if the board
is too close: the middle of the board was exposed ok, but the ends of
the board were barely touched.

Somewhere in between is a happy medium.

Mitch.

Luke Weston

unread,
Nov 2, 2009, 7:05:47 AM11/2/09
to Connected Community HackerSpace


On Nov 2, 10:25 pm, michaelc <michael.col...@gmail.com> wrote:
> I've done a little looking into this today too. I've found a few
> discussions about photo resist (though not about Kinsten in
> particular) that suggest the optimum wavelength is around 350nm, which
> is certainly in the UVA range. My understanding of the BLB tubes is
> that they are less efficient than a BL tube because they rely on a
> subtractive approach as you've described - most of the input energy is
> not output as UV.

Yeah, but the coating is only attenuating the visible, it's not
supposed to be attenuating the UV much if at all, so I would have
thought it wouldn't be any different.

How many ballasts do
> we have in that light fitting, and would they be suitable for the 15W
> tubes? We could always just build the bottom half first to see how
> things go.

3 tubes, and 3 sets of the ballast, starter, and the pair of
"tombstone" tube end sockets.
I don't know what the designed tube power is, I'd have to read it off
the original tubes, but they're at Andy's place so I can't right now.
Can you use a fluoro ballast with tubes of less than the power rating
that it's designed for?

RE: transmission in plastics and glass. It's clear* that glass is
superior... but it's hard to cut it yourself.
But you can just go to a glass shop and ask for a bit of glass cut for
the dimensions you need.
Then just use a wood bead, or a bit of silicone, to hold it in place.

* Sorry.

michaelc

unread,
Nov 2, 2009, 8:33:41 AM11/2/09
to Connected Community HackerSpace
>
> Yeah, but the coating is only attenuating the visible, it's not
> supposed to be attenuating the UV much if at all, so I would have
> thought it wouldn't be any different.
>

The incandescent BLB bulbs are the inefficient and low output ones, as
they generate a broad spectrum of light and filter everything except
the UV. When it comes to fluoro tubes, you're right that it makes no
difference. The mercury vapor in the tube emits lots of short
wavelength UV which is converted to the UVA we want by the phosphor
coating, but also some visible light. The Woods glass coating on a BLB
fluoro is just useful when you really don't want the visible part of
the light emitted by the tube. In our case, that extra light doesn't
matter at all so we can use the cheaper and more common tubes without
the coating.

>
> RE: transmission in plastics and glass. It's clear* that glass is
> superior... but it's hard to cut it yourself.
> But you can just go to a glass shop and ask for a bit of glass cut for
> the dimensions you need.
> Then just use a wood bead, or a bit of silicone, to hold it in place.
>

Silicone would almost certainly work for mounting glass to a frame,
but I'm not yet convinced that glass is superior for this application.
I currently expose through acrylic, so I know its transparency to UV
is generally acceptable. Tomorrow I'll try to find out if this is a
significant factor in the exposing process or just an interesting
technical distraction. From what I've read today, the best case
transparency for acrylic is better than glass. In other regards
acrylic has some nice properties: considerably easier to work with
than glass, lighter, and less likely to break in an unpleasant way.

> * Sorry.

As you ought to be. :) Punning is a vice, and there's no vice versa.

Michael

michaelc

unread,
Nov 2, 2009, 8:40:34 AM11/2/09
to Connected Community HackerSpace
> Can you use a fluoro ballast with tubes of less than the power rating
> that it's designed for?
>

I'm not sure, but here's a document about fluoros that might be useful
if we decide to take the build our own route: http://members.misty.com/don/f-lamp.html

Michael

Michael Borthwick

unread,
Nov 2, 2009, 7:46:30 PM11/2/09
to connected-commu...@googlegroups.com


Here are some pics of the Pebble board in question. The exposure time
was only 30 seconds. I don't think it is fair to say that the ends
were "barely touched".

http://www.flickr.com/photos/65169791@N00/4067976623/in/photostream/
http://www.flickr.com/photos/65169791@N00/4067976393/in/photostream/

Arguably the board is salvageable although personally I would be
inclined to chalk it up to beer fuelled bravado and make another.

I originally built my light box for exposing single sided PCB's but
improvised the double-sided mode used on the weekend when Luke turned
up with double-sided artwork for the Artemis rocket PCB: http://www.flickr.com/photos/65169791@N00/4068664528/

I think that trying to expose both sides of the Pebble board
simultaneously by sandwiching it between the tubes failed because
Pebble is wider than Artemis and therefore suffers more from the
inverse square law fall-off of light flux at the extremities of the PCB.

The fact that there is any image at all at the edges suggests to me
that the process has quite a bit of latitude and as long as people
avoid the obvious boundary conditions I think that the technique of
making PCB's using pre-sensitised board is straightforward and fun.

Before I made the light box I exposed all my boards under a single
conventional Hanimex flouro desk lamp for 15 minutes (moving it around
for larger boards) and "back in the day" the board shown here was
exposed using sunlight:
http://www.flickr.com/photos/65169791@N00/3885904494/in/photostream/

Mike

Mitch Davis

unread,
Nov 3, 2009, 1:07:50 AM11/3/09
to connected-commu...@googlegroups.com
On Tue, Nov 3, 2009 at 11:46 AM, Michael Borthwick
<hol...@netspace.net.au> wrote:
>
> On 02/11/2009, at 11:04 PM, Mitch Davis wrote:
>
> Here are some pics of the Pebble board in question. The exposure time
> was only 30 seconds. I don't think it is fair to say that the ends
> were "barely touched".

Oh wow! Well, they look a lot different to when I last saw them!
Thank you for posting those pics. My comment wasn't meant as
criticism; I have a lot to learn about making circuit boards.

Mitch.

michaelc

unread,
Nov 3, 2009, 4:06:00 AM11/3/09
to Connected Community HackerSpace
On Nov 3, 12:33 am, michaelc <michael.col...@gmail.com> wrote:
> is generally acceptable. Tomorrow I'll try to find out if this is a
> significant factor in the exposing process or just an interesting
> technical distraction. From what I've read today, the best case

I've just completed a simple test. I've exposed 3 small (about 15 mm x
10 mm) boards with a pattern consisting of some text and traces from
16 mils down to 8 mils. All three boards were exposed at the same time
and developed at the same time. The first board was exposed in my
normal way as a reference, with one layer of acrylic between the light
and the board, the second was exposed with two pieces of acrylic
between the light and the board, and the third with one layer of
acrylic and an extra layer of drafting paper. The results:

- The first (reference) board was slightly under exposed. I don't know
why: it could be because the lights in my exposure box are no longer
as bright as they used to be and I need to re-calibrate my exposure
time, or it could be because I was using some very old Kinsten from
2002 (I wanted to use up the scraps).
- The second (2 x acrylic) board seems slightly less exposed than the
reference board but almost imperceptibly, so that might just be my
imagination.
- The third (extra layer of paper) board is significantly less well
exposed and would not be usable.

My conclusions are that the acrylic is not attenuating the UV to any
degree that is significant to the exposure process. The choice between
glass and acrylic does not seem to hinge around transparency, but
around other practical aspects that all lead me to choose acrylic. The
effect of the extra layer of drafting paper is interesting. One day
I'll try to find some transparencies that will work in my laser
printer and see if they produce any better outcomes (better resolution
would be nice) other than shorter exposure time. I have some laser
printer transparencies but my printer just doesn't fuse the toner on
to them.

Michael Collas
Reply all
Reply to author
Forward
0 new messages