Homebrew PCBs

Someone said that no professional board shops use ultraviolet LEDs.  Why 
is that?

-- 
David Griffith
dgriffi@...

A: Because it fouls the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing in e-mail?

On 28/01/2010 23:18, David Griffith wrote:
> Someone said that no professional board shops use ultraviolet LEDs.  Why
> is that?
>
>    
Much better resolution is achieved with UV tubes.

Leon

>Much better resolution is achieved with UV tubes.

OK, I'll bite.  Why do UV tubes give better resolution than UV LEDs?  

Stuart





[Non-text portions of this message have been removed]

Hi.

I have a feeling its because professional production have big 300mm x 
400mm size panels and the number of LEDs in one unit just makes it too 
large. It would have like 400 LEDs per side. Where as at present they 
just stick a few tubes in which can be replaced easily.

Trev

Since I know very very little about optics, I can not tell you a sceintific answer. However, the problem lies in how the light incide on to the superficial area of the laminate. If you place a LED matrix you can not guarantee that a uniform light will incide the board. Think of that light as cones emerging from each led. There is some equipment that can "see" how the light pattern is emerging. I do not remember which equipment is (maybe for scientific purposes) but you would see that there are "interferences" in the light. Something like water waves adjacent to others. And forgetting about that expensive scientific equipment and returning to PCB shop equipment, if you use a radiometer measuring different points you would encounter that the UV emission is different. If you try to collimate that LED matrix you could waste a lot of time, effort, and too much money just to obtain (if you can do it) almost the same performance that a flourescent or mercury lamp setup. Maybe someone with real optics knowledge could bring some real information, and not just the empirical question that is enough for me.

It might also have something to do with the UV wavelength. I think
LEDs have a longer wavelength, close to purple, while the tubes might
be further into the UV range. This may affect exposure time or
resolution.

> Someone said that no professional board shops use ultraviolet LEDs. Why
> is that?
>
> --
> David Griffith
> dgriffi@...
>

On Fri, 29 Jan 2010, Trevor White wrote:

> Hi.
>
> I have a feeling its because professional production have big 300mm x
> 400mm size panels and the number of LEDs in one unit just makes it too
> large. It would have like 400 LEDs per side. Where as at present they
> just stick a few tubes in which can be replaced easily.

I suspect too that light from a tube may be easier to diffuse.  Another 
thing may be that there are fewer parts to go bad.  Looking around town I 
see lots of LED traffic lights in various stages of partial failure.

-- 
David Griffith
dgriffi@...

A: Because it fouls the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing in e-mail?

I wonder why that is?

On Thu, Jan 28, 2010 at 5:42 PM, Leon Heller <leon355@...> wrote:

> On 28/01/2010 23:18, David Griffith wrote:
> > Someone said that no professional board shops use ultraviolet LEDs.  Why
> > is that?
> >
> >
> Much better resolution is achieved with UV tubes.
>
> Leon
>
>


[Non-text portions of this message have been removed]

> I have a feeling its because professional production have big 300mm x 
> 400mm size panels and the number of LEDs in one unit just makes it too 
> large. It would have like 400 LEDs per side. Where as at present they 
> just stick a few tubes in which can be replaced easily.
> 
> Trev

No, sorry but that is not the reason. A typical exposure UV unit for industrial use can cost USD$10,000.00 and that is just the beginning prices. Although thousands of LEDs per side would be required, the number (and price) of them it is not significant in the industrial shops. The other main reason to avoid LEDs that I forgot to explain in another post is that the UV emission is very poor compared to fluorescent and mercury lamps. You could fix that getting closer the LEDs to the PCB board but that is not desirable in any way. To obtain real small traces and details the solution is to put the UV source and the PCB so far that is practical, so you would need very high ouput in the light source. Some professional UV units have the PCB and UV source separated more than 90 centimeters. However they use very high mercury-lamps and the exposure only needs a few seconds. Typical semi-professional UV exposure units (like MegaUK and so) made from fluorescent tubes also have narrow distances between the PCB and UV tubes, but that is a reason of compromise, trying to balance cost and performance.

That makes sense. ... Thanks for the great explanation.  Basically the
quality (intensity and spectrum) of the UV is better.
><> ... Jack

On Thu, Jan 28, 2010 at 6:31 PM, jcarlosmor <jcarlosmor@...> wrote:

>
> > I have a feeling its because professional production have big 300mm x
> > 400mm size panels and the number of LEDs in one unit just makes it too
> > large. It would have like 400 LEDs per side. Where as at present they
> > just stick a few tubes in which can be replaced easily.
> >
> > Trev
>
> No, sorry but that is not the reason. A typical exposure UV unit for
> industrial use can cost USD$10,000.00 and that is just the beginning prices.
> Although thousands of LEDs per side would be required, the number (and
> price) of them it is not significant in the industrial shops. The other main
> reason to avoid LEDs that I forgot to explain in another post is that the UV
> emission is very poor compared to fluorescent and mercury lamps. You could
> fix that getting closer the LEDs to the PCB board but that is not desirable
> in any way. To obtain real small traces and details the solution is to put
> the UV source and the PCB so far that is practical, so you would need very
> high ouput in the light source. Some professional UV units have the PCB and
> UV source separated more than 90 centimeters. However they use very high
> mercury-lamps and the exposure only needs a few seconds. Typical
> semi-professional UV exposure units (like MegaUK and so) made from
> fluorescent tubes also have narrow distances between the PCB and UV tubes,
> but that is a reason of compromise, trying to balance cost and performance.
>
>


[Non-text portions of this message have been removed]

On 28/01/2010 23:49, Stuart Schaffert wrote:
>> Much better resolution is achieved with UV tubes.
>>      
> OK, I'll bite.  Why do UV tubes give better resolution than UV LEDs?
>
>    
Collimation.

Leon

--- In Homebrew_PCBs@yahoogroups.com, James Bishop <bishopaj@...> wrote:
>
> It might also have something to do with the UV wavelength. I think
> LEDs have a longer wavelength, close to purple, while the tubes might
> be further into the UV range. This may affect exposure time or
> resolution.
> 
>


Fluorescent tubes emit light in sharp peaks of fairly specific wavelenghts. There are several peaks produced by the mercury vapor discharge itself, and others produced by the phosphor coating which is excited by absorbing the short wave UV from the discharge. Ordinary glass absorbs shortwave UV, so special germicidal and ozone producing tubes are made of quartz.

LEDs tend to emit peaks as well, but the wavelengths are indeed longer, I don't think anyone has come up with a true UV LED, the closest they get is deep violet which will excite many UV sensitive materials but the intensity is much lower than that of a mercury discharge which is a very strong UV emitter.

>> Stuart Schaffert wrote:
>> OK, I'll bite.  Why do UV tubes give better
>> resolution than UV LEDs?

> Leon wrote:
> Collimation.

And intensity.

A 60 second exposure with a UV tube will give
harder edges to the resist than a 10 minute
exposure with a lower intensity source.

And uniformity as someone else also pointed
out.

On 29/01/2010 02:50, James wrote:
>
> --- In Homebrew_PCBs@yahoogroups.com, James Bishop<bishopaj@...>  wrote:
>    
>> It might also have something to do with the UV wavelength. I think
>> LEDs have a longer wavelength, close to purple, while the tubes might
>> be further into the UV range. This may affect exposure time or
>> resolution.
>>
>>
>>      
>
> Fluorescent tubes emit light in sharp peaks of fairly specific wavelenghts. There are several peaks produced by the mercury vapor discharge itself, and others produced by the phosphor coating which is excited by absorbing the short wave UV from the discharge. Ordinary glass absorbs shortwave UV, so special germicidal and ozone producing tubes are made of quartz.
>    

PCBs require long wave UV which isn't absorbed by ordinary glass.

Leon

"cunningfellow" <andrewm1973@...> writes:
> A 60 second exposure with a UV tube will give harder edges to the
> resist than a 10 minute exposure with a lower intensity source.

Would this affect etch-under at all?  My latest board had trace
thinning to the point where many were missing completely, I've been
"bloating" my films to compensate but a better process would be a
better solution.

--- In Homebrew_PCBs@yahoogroups.com, Leon Heller <leon355@...> wrote:
>
> On 29/01/2010 02:50, James wrote:
> >
> > --- In Homebrew_PCBs@yahoogroups.com, James Bishop<bishopaj@>  wrote:
> >    
> >> It might also have something to do with the UV wavelength. I think
> >> LEDs have a longer wavelength, close to purple, while the tubes might
> >> be further into the UV range. This may affect exposure time or
> >> resolution.
> >>
> >>
> >>      
> >
> > Fluorescent tubes emit light in sharp peaks of fairly specific wavelenghts. There are several peaks produced by the mercury vapor discharge itself, and others produced by the phosphor coating which is excited by absorbing the short wave UV from the discharge. Ordinary glass absorbs shortwave UV, so special germicidal and ozone producing tubes are made of quartz.
> >    
> 
> PCBs require long wave UV which isn't absorbed by ordinary glass.
> 
> Leon
>


I never said they didn't. As I mentioned, there are multiple wavelengths emitted by the unfiltered blacklight UV tubes, the strongest being in the long wave UV range which is emitted by the phosphor being excited by short wave UV inside the tube. UV is by definition invisible to the human eye, what you see are various visible wavelengths also emitted by the tube. "UV" LEDs are not really UV, they are deep violet, but close enough that many UV sensitive materials respond to some extent.

> 
> Would this affect etch-under at all?  My latest board had trace
> thinning to the point where many were missing completely, I've been
> "bloating" my films to compensate but a better process would be a
> better solution.
>


What are you using to hold your film to the blank PCB? You have to make sure the printed side is against the board, and it has to be absolutely flat. Use a piece of relatively thick glass, 3/16" or so that won't flex, and hold it against the board with a decent amount of pressure. If you're using a light box, it works well to set a brick on top of the board to hold it down flat on the film. Many printers have trouble getting a nice solid black on transparency film too, vellum paper tends to work better, it's transparent enough to UV, and a laser printer will lay down a nice dark print on it without any warp or ripple.

I've only made a handful of photo etch boards myself, but I've never had any thinning at all. I've always used a 15W bug zapper tube and exposed for 4-5 minutes.

James wrote:
> 
> --- In Homebrew_PCBs@yahoogroups.com, James Bishop <bishopaj@...> wrote:
>> It might also have something to do with the UV wavelength. I think
>> LEDs have a longer wavelength, close to purple, while the tubes might
>> be further into the UV range. This may affect exposure time or
>> resolution.
>>
> 
> Fluorescent tubes emit light in sharp peaks of fairly specific wavelenghts. There are several peaks produced by the mercury vapor discharge itself, and others produced by the phosphor coating which is excited by absorbing the short wave UV from the discharge. Ordinary glass absorbs shortwave UV, so special germicidal and ozone producing tubes are made of quartz.
> 
> LEDs tend to emit peaks as well, but the wavelengths are indeed longer, I don't think anyone has come up with a true UV LED, the closest they get is deep violet which will excite many UV sensitive materials but the intensity is much lower than that of a mercury discharge which is a very strong UV emitter.

The power level from array LEDs is high.

http://www.phoseon.com/
http://www.leduv.com/
http://www.jentonuvled.com/

http://www.google.com.au/#hl=en&source=hp&q=%22uv+led%22&btnG=Google+Search&meta=&aq=f&oq=%22uv+led%22&fp=4fcec8be4c0b548b

"James" <jamesrsweet@...> writes:
> What are you using to hold your film to the blank PCB?

Thin glass plate with 3-8 in Hg vacuum.  Holds it very tight against
the PCB.

> You have to make sure the printed side is against the board,

Always do.

> printers have trouble getting a nice solid black on transparency
> film too, vellum paper tends to work better,

I didn't have any luck with vellum, the laser printer didn't put down
much toner and ink just smeared.  I'm using Jetstar film with an
inkjet, I can do 100% coverage at 2880 dpi and get *fantastic* prints
off it (even under a microscope).  Gives about 8 steps on the 21-step
scale so I know the ink print is good.

The UV film after developing looks great too, it just makes traces
that are smaller than the film by about 1-2 mil on each edge, which is
enough on a 6 mil trace to not leave enough surface area for the film
to hold on, and it etches completely through.  I figured maybe the
film's edge had something to do with it.

DJ Delorie wrote:

> 
> Would this affect etch-under at all? My latest board had trace
> thinning to the point where many were missing completely, I've been
> "bloating" my films to compensate but a better process would be a
> better solution.


Just guessing but did you check either light intensity distribution (are
you using the leds setup) but most important (what at least Adam
Seychell calls) etch break point. He uses one non printed copper clad
board in the etch and sees if are copper areas etched sooner then
others. That means bad agitation. See his last comments about etch tank
setup.

Else just over expose it a bit...

David Griffith wrote:
> Someone said that no professional board shops use ultraviolet LEDs.  Why 
> is that?
> 

An LED is a point source of light. Fluorescent tubes emit a more-or-less 
constant intensity of fairly diffused UV along almost their entire length.

Thus it's easier to calculate the exposure for the boards, then once 
you've got everything set up, it's easily repeatable. On top of that, 
you don't get brightness peaks and troughs as you move around the 
exposure platform (glass plate or whatever).

-- 
Phil.
ygroups@...
http://www.philpem.me.uk/