Homebrew_PCBs

Hello...

Long time ago I remember that someone has UV laser mounted on inkjet printer and developing photoresist on pcb directly.
I'd like to know is someone has making some test with cheap 3W UV LED instead laser. I think the power from LED can be focused on pretty small spot to cure photoresist. But don't have any idea how long should take to.
My idea is to use focused LED to aprox 1mm spot and then to use some mask (like aperture whell) to do finer trace. And to use vector instead raster imaging.

If you are interested in direct imaging a photo sensitized board search back
through the posts around Nov of 09 for Volkan. He built a UV laser diode direct
writer that did spectacular images. The 400 nm blue-ray lasers are available on
ebay. He has some images posted under the Files secton of the forum. These UV
lasers are very strong >> 100mW.

I took him a board I had coated in AQ3000 (water based photo resist) to try an
image. Once we got the exposure set the image was amazing crisp.

Volkan also built a drum photo plotter.

Both these project on my list of cool things to build.

Craig

I found that posts but there are 150mW laser involved. And this thing
isn't cheap. In other side the UV power led is cheap. And I think that
3W UV led should compare to 150mW laser. (20x more power)

And my intention is to make that beam aprox paralel to with size let's
say 1mm squared. And then to use aperture whell wit round holes from
8mil up to 40mil, and used gerber stile data to work as old day
photoploter. Ie not raster scan but vectors...

Just don't know if LED is good enought...
Slavko

On 09/14/2013 09:05 PM, designer_craig wrote:
> If you are interested in direct imaging a photo sensitized board search back
> through the posts around Nov of 09 for Volkan. He built a UV laser diode direct
> writer that did spectacular images. The 400 nm blue-ray lasers are available on
> ebay. He has some images posted under the Files secton of the forum. These UV
> lasers are very strong >> 100mW.
>
> I took him a board I had coated in AQ3000 (water based photo resist) to try an
> image. Once we got the exposure set the image was amazing crisp.
>
> Volkan also built a drum photo plotter.
>
> Both these project on my list of cool things to build.
>
> Craig
>
>

Just out of curiosity, why the vector writing an image. Sounds like it would take a long time to draw out your circuit pattern.

I've gone from pulsar toner transfer paper (digi key part # 182-1003-nd) to a Puretch photo resist film (from http://www.capefearpress.com/puretch.html). I like my sources to be consistent and reliable. No offense to those who go Ebay route.

I print my pattern on Highland Transparency film #78-6969-8593-2 (from office depot) and use my HP 2600 laser printer set on B&W only.

Puretch film is applied to a clean copper clad board. set the printed image on top. Make sure there are no spaces between the image and the board. I use a 50W halogen light about 6" away for 2"x2" board. Expose for 15 minutes.

remove protective mylar film, develop the resist in soda ash solution, pat dry. then expose the film for another 10 minutes under the halogen light to cure (harden) the resist.

Etch.

Seems less of a hassle than creating a vector printer. Image resolution is down to the resolution of my printer.

Puretch claim they can resolve down to less than a mil line width. But my printer can't. :P

You might want to give this a try. I use it to pattern TQFP and 0.5mm pitched parts along with 0402 components.

Just a thought.

And just follow the direction carefully on capefear youtube video. Lastly, the image has to be reversed. where you want the copper, you want a clear view. Where you don't want the copper the image on the transparency has to be black.

Either way, I wish you lots of fun tinkering with solid state laser.

--- In Homebrew_PCBs@yahoogroups.com, Slavko Kocjancic <eslavko@...> wrote:
>
> I found that posts but there are 150mW laser involved. And this thing
> isn't cheap. In other side the UV power led is cheap. And I think that
> 3W UV led should compare to 150mW laser. (20x more power)
>
> And my intention is to make that beam aprox paralel to with size let's
> say 1mm squared. And then to use aperture whell wit round holes from
> 8mil up to 40mil, and used gerber stile data to work as old day
> photoploter. Ie not raster scan but vectors...
>
> Just don't know if LED is good enought...
> Slavko
>
>
> On 09/14/2013 09:05 PM, designer_craig wrote:
> > If you are interested in direct imaging a photo sensitized board search back
> > through the posts around Nov of 09 for Volkan. He built a UV laser diode direct
> > writer that did spectacular images. The 400 nm blue-ray lasers are available on
> > ebay. He has some images posted under the Files secton of the forum. These UV
> > lasers are very strong >> 100mW.
> >
> > I took him a board I had coated in AQ3000 (water based photo resist) to try an
> > image. Once we got the exposure set the image was amazing crisp.
> >
> > Volkan also built a drum photo plotter.
> >
> > Both these project on my list of cool things to build.
> >
> > Craig
> >
> >
>

Hello...
With this type (photo with transparency) of operation I can make board
for SMD, TQFP... down to 0.5mm pin pitch. But I'm not able to make one
big board with 20cm long connector. The difference in length is random
from death true to the 1mm! You can try to draw 20cm x 20cm rectangle
and print it. Then measure size of rectangle and diagonals. Printed
with LJ1018, LJ1020, (HL2030 little better) is just rubbish!. It's not
problem for board up to aprox 10x10cm as difference is small enought to
force connector into holes. And of course as I dril with CNC the wrong
size is not good thing either.

On 09/16/2013 05:56 AM, smilingcat90254 wrote:
> Just out of curiosity, why the vector writing an image. Sounds like it would take a long time to draw out your circuit pattern.
>
> I've gone from pulsar toner transfer paper (digi key part # 182-1003-nd) to a Puretch photo resist film (from http://www.capefearpress.com/puretch.html). I like my sources to be consistent and reliable. No offense to those who go Ebay route.
>
> I print my pattern on Highland Transparency film #78-6969-8593-2 (from office depot) and use my HP 2600 laser printer set on B&W only.
>
> Puretch film is applied to a clean copper clad board. set the printed image on top. Make sure there are no spaces between the image and the board. I use a 50W halogen light about 6" away for 2"x2" board. Expose for 15 minutes.
>
> remove protective mylar film, develop the resist in soda ash solution, pat dry. then expose the film for another 10 minutes under the halogen light to cure (harden) the resist.
>
> Etch.
>
> Seems less of a hassle than creating a vector printer. Image resolution is down to the resolution of my printer.
>
> Puretch claim they can resolve down to less than a mil line width. But my printer can't. :P
>
> You might want to give this a try. I use it to pattern TQFP and 0.5mm pitched parts along with 0402 components.
>
> Just a thought.
>
> And just follow the direction carefully on capefear youtube video. Lastly, the image has to be reversed. where you want the copper, you want a clear view. Where you don't want the copper the image on the transparency has to be black.
>
> Either way, I wish you lots of fun tinkering with solid state laser.
>
>
> --- In Homebrew_PCBs@yahoogroups.com, Slavko Kocjancic <eslavko@...> wrote:
>> I found that posts but there are 150mW laser involved. And this thing
>> isn't cheap. In other side the UV power led is cheap. And I think that
>> 3W UV led should compare to 150mW laser. (20x more power)
>>
>> And my intention is to make that beam aprox paralel to with size let's
>> say 1mm squared. And then to use aperture whell wit round holes from
>> 8mil up to 40mil, and used gerber stile data to work as old day
>> photoploter. Ie not raster scan but vectors...
>>
>> Just don't know if LED is good enought...
>> Slavko
>>
>>
>> On 09/14/2013 09:05 PM, designer_craig wrote:
>>> If you are interested in direct imaging a photo sensitized board search back
>>> through the posts around Nov of 09 for Volkan. He built a UV laser diode direct
>>> writer that did spectacular images. The 400 nm blue-ray lasers are available on
>>> ebay. He has some images posted under the Files secton of the forum. These UV
>>> lasers are very strong >> 100mW.
>>>
>>> I took him a board I had coated in AQ3000 (water based photo resist) to try an
>>> image. Once we got the exposure set the image was amazing crisp.
>>>
>>> Volkan also built a drum photo plotter.
>>>
>>> Both these project on my list of cool things to build.
>>>
>>> Craig
>>>
>>>
>
>
>
> ------------------------------------
>
> Be sure to visit the group home and check for new Links, Files, and Photos:
> http://groups.yahoo.com/group/Homebrew_PCBsYahoo! Groups Links
>
>
>
>

Hello...
With this type (photo with transparency) of operation I can make board
for SMD, TQFP... down to 0.5mm pin pitch. But I'm not able to make one
big board with 20cm long connector. The difference in length is random
from death true to the 1mm! You can try to draw 20cm x 20cm rectangle
and print it. Then measure size of rectangle and diagonals. Printed
with LJ1018, LJ1020, (HL2030 little better) is just rubbish!. It's not
problem for board up to aprox 10x10cm as difference is small enought to
force connector into holes. And of course as I dril with CNC the wrong
size is not good thing either.

On 09/16/2013 05:56 AM, smilingcat90254 wrote:
> Just out of curiosity, why the vector writing an image. Sounds like it would take a long time to draw out your circuit pattern.
>
> I've gone from pulsar toner transfer paper (digi key part # 182-1003-nd) to a Puretch photo resist film (from http://www.capefearpress.com/puretch.html). I like my sources to be consistent and reliable. No offense to those who go Ebay route.
>
> I print my pattern on Highland Transparency film #78-6969-8593-2 (from office depot) and use my HP 2600 laser printer set on B&W only.
>
> Puretch film is applied to a clean copper clad board. set the printed image on top. Make sure there are no spaces between the image and the board. I use a 50W halogen light about 6" away for 2"x2" board. Expose for 15 minutes.
>
> remove protective mylar film, develop the resist in soda ash solution, pat dry. then expose the film for another 10 minutes under the halogen light to cure (harden) the resist.
>
> Etch.
>
> Seems less of a hassle than creating a vector printer. Image resolution is down to the resolution of my printer.
>
> Puretch claim they can resolve down to less than a mil line width. But my printer can't. :P
>
> You might want to give this a try. I use it to pattern TQFP and 0.5mm pitched parts along with 0402 components.
>
> Just a thought.
>
> And just follow the direction carefully on capefear youtube video. Lastly, the image has to be reversed. where you want the copper, you want a clear view. Where you don't want the copper the image on the transparency has to be black.
>
> Either way, I wish you lots of fun tinkering with solid state laser.
>
>
> --- In Homebrew_PCBs@yahoogroups.com, Slavko Kocjancic <eslavko@...> wrote:
>> I found that posts but there are 150mW laser involved. And this thing
>> isn't cheap. In other side the UV power led is cheap. And I think that
>> 3W UV led should compare to 150mW laser. (20x more power)
>>
>> And my intention is to make that beam aprox paralel to with size let's
>> say 1mm squared. And then to use aperture whell wit round holes from
>> 8mil up to 40mil, and used gerber stile data to work as old day
>> photoploter. Ie not raster scan but vectors...
>>
>> Just don't know if LED is good enought...
>> Slavko
>>
>>
>> On 09/14/2013 09:05 PM, designer_craig wrote:
>>> If you are interested in direct imaging a photo sensitized board search back
>>> through the posts around Nov of 09 for Volkan. He built a UV laser diode direct
>>> writer that did spectacular images. The 400 nm blue-ray lasers are available on
>>> ebay. He has some images posted under the Files secton of the forum. These UV
>>> lasers are very strong >> 100mW.
>>>
>>> I took him a board I had coated in AQ3000 (water based photo resist) to try an
>>> image. Once we got the exposure set the image was amazing crisp.
>>>
>>> Volkan also built a drum photo plotter.
>>>
>>> Both these project on my list of cool things to build.
>>>
>>> Craig
>>>
>>>
>
>
>
> ------------------------------------
>
> Be sure to visit the group home and check for new Links, Files, and Photos:
> http://groups.yahoo.com/group/Homebrew_PCBsYahoo! Groups Links
>
>
>
>

Does anyone have a link to Volkan msg of the 9th of November? yahoo search just give me error msg's

On Sun, Sep 15, 2013 at 12:05 AM, Slavko Kocjancic <eslavko@...> wrote:

I found that posts but there are 150mW laser involved. And this thing
isn't cheap. In other side the UV power led is cheap. And I think that
3W UV led should compare to 150mW laser. (20x more power)

And my intention is to make that beam aprox paralel to with size let's
say 1mm squared. And then to use aperture whell wit round holes from
8mil up to 40mil, and used gerber stile data to work as old day
photoploter. Ie not raster scan but vectors...

Just don't know if LED is good enought...
Slavko

On 09/14/2013 09:05 PM, designer_craig wrote:
> If you are interested in direct imaging a photo sensitized board search back
> through the posts around Nov of 09 for Volkan. He built a UV laser diode direct
> writer that did spectacular images. The 400 nm blue-ray lasers are available on
> ebay. He has some images posted under the Files secton of the forum. These UV
> lasers are very strong >> 100mW.
>
> I took him a board I had coated in AQ3000 (water based photo resist) to try an
> image. Once we got the exposure set the image was amazing crisp.
>
> Volkan also built a drum photo plotter.
>
> Both these project on my list of cool things to build.
>
> Craig
>
>

The 150mw or even higher power UV lasers are available on Ebay, they are used in BlueRay DVD players. An LED would work but it will take a very long time to expose a board. You need lots of joules/sq mm to activate the resist. The exposure spot area is going to be around .05mm to get good pattern definition and its just not going to be possible to mask and focus and LED to that level with any kind of power. Most higher power LED's have multiple junctions which means they are not point sources and such will not focus to a point.

The laser Volkan was using was in the $30 range from Ebay, its best to use one with a glass lenses for better focus and less loss.

The laser module with heat sink was mounted into an ink jet printer shuttle, the board moved in the Y direction on a custom built bed below the laser. The bed was driven by a stepper motor driving a lead screw. He built custom electronics to provide the laser video modulation it was was timed by the plastic head reader strip inside the ink jet printer. As the head moved across the carriage the laser was turned on and off. Once the head scanned across the board the bed was moved for the next pass. He only wrote in one direction because of technical issues with reading the plastic sensing strip.

Even with the carriage moving at full speed in both directions it took about 15 to 20 minutes to write a 150mm x 75mm board, but the image was spectacular.
Craig

--- In Homebrew_PCBs@yahoogroups.com, Slavko Kocjancic <eslavko@...> wrote:
>
> I found that posts but there are 150mW laser involved. And this thing
> isn't cheap. In other side the UV power led is cheap. And I think that
> 3W UV led should compare to 150mW laser. (20x more power)
>
> And my intention is to make that beam aprox paralel to with size let's
> say 1mm squared. And then to use aperture whell wit round holes from
> 8mil up to 40mil, and used gerber stile data to work as old day
> photoploter. Ie not raster scan but vectors...
>
> Just don't know if LED is good enought...
> Slavko
>
>
> On 09/14/2013 09:05 PM, designer_craig wrote:
> > If you are interested in direct imaging a photo sensitized board search back
> > through the posts around Nov of 09 for Volkan. He built a UV laser diode direct
> > writer that did spectacular images. The 400 nm blue-ray lasers are available on
> > ebay. He has some images posted under the Files secton of the forum. These UV
> > lasers are very strong >> 100mW.
> >
> > I took him a board I had coated in AQ3000 (water based photo resist) to try an
> > image. Once we got the exposure set the image was amazing crisp.
> >
> > Volkan also built a drum photo plotter.
> >
> > Both these project on my list of cool things to build.
> >
> > Craig
> >
> >
>

Hi Buckwheat,

maybe you should refrain from condescending remarks which I have highlighted in bold. It doesn't add any useful information. I could easily remark "Perhaps its not just your Mandarin that you are slaving over and perhaps you should have read it tad more carefully."

I did say print not plot. I also said HP2600 laser printer again not a plotter. I also did not say it was inkjet but rather a laser printer. I was informed that majority on this forum and other tech forum, members are not necessarily familiar with printing technology nor with electronics. I tried to keep it simple for the new person. If you are still not sure I print the gerber image onto a transparency using my HP laser printer. It can print images and not just text.

Slavkok had no problem understanding my concern and was able to reply as to why he was going down the road he chose. And someone else did quip that it will take 15 to 20 minutes to write a moderately sized board.

I do have lot to offer technically. I do not need flaming from participants. It is my choice to be here to be helpful. So if you wish to feel like a big man by flaming people, go right ahead and others here will not get any technical information. Not some false information often found on internet. e.g. can't store copper chloride in plastic container. Copper chloride is a metal salt and will not attack plastic.

There were few others wrong information floating around.

If you are having inconsistent exposure, you may need to be using a calibrated exposure scale from Stouffer http://www.stouffer.net/Specialtyguides.htm ... You may be over exposing.

Sincerely,

Ms. Smilingcat

--- In Homebrew_PCBs@yahoogroups.com, brane2 wrote:
>
> Hmm... I'm not trying to be language fascist as god knows I make many
> mistakes, but here I have hard time following what you are trying to say.
>
> Maybe quick brush-up with some online english course would be in order ?
> Damn it, if I can slave with mandarin, you should be able to endure an
> english lesson or two... ;o)
>
> Anyway, if I understood you correctly, you:
>
> - are using classical approach with photoresist and UV exposure
>
> - i don't understand what you mean by "photo". Is it "photoplotter"
> transparency, done in some graphic shop ( on a something like Linotronic
> imagesetter)
> OR
> .. you mean transparency printed on inkjet
> OR
> ... you mean transparency, printed on laser
>
> It is also not clear how does your transparency differ from drill
> coordinates. Does it get randomly stretched or it is just pseudorandom-
> seemingly random but exactly repeatable with repeated print of the same
> content ?
>
> I tried it with Epson's R800 and results seem at first glance very
> repeatable, but not enough dark. This is why I'm checking cups drivers
> on Linux to see whether I can persuade printer to make several passes
> over each point of the page without reinserting the page as printer
> never seems to grab it exactly the same way and edges get fuzzied and
> multiplied.
>
>
> Best regards,
>
>
> Branko

...so to clean up the problem, and not to argue...

I can make small board with toner transfer metod. Fine traces aren't
problem.
I can make small board with sensitived board and printed transparency on
laser Fine traces aren't problem.

... But I have problem with accuarcy. If I draw rectangle 200mm x 200mm
and then print that and then measure that, the dimennsion is not right.
With few prints made with same settings I measure the side from 199.5mm
up to 201.3mm. I tested 3 printers I own, and LaserJet 1018 and 1020
works identicaly. The Brother HL2030 is little better (from 200mm to 201.1)

So If I pick some big SOIC or TQFP with side of 20mm the error is aprox
aprox 0.1mm. It's bare visible with naked eye. So no problem with that.

Then I pick some chip in DIL40. The last pin is aprox 0.4mm off the
hole. If holes are drilled oversized then package should fit, else wil NOT.

Now I have terminal blocks on the edge of board. I have 20 pins stacked
(200 mils) and in last pin the difference can be 1mm and it's not
possible to fit that component.

And other thing is that I drill with CNC. And of course if one side of
board is ok the other is not.

Why the print's isnt consistent I don't know.
For toner transfer (printing on paper) It can be humidity, but when
printing on plastic transparency I think the humidity can't be problem.
As image is more distorted in height I assume that it's printer
mechanics problem.

As I know that my CNC is way under 0.1mm precission It's logical for me
to try to use it as photoploter.
The process is slower but got accuracy. So If I need accuracy (large
board) I should go with photoploter style, for small board's the toner
transfer is better choice...

I know that my english isn't perfect, but think that majority of people
will understand me. ...ker ne govorim kitajsko.


Slavko.

Hi Slavko,

It is very common for laser printers to have errors in vertical and horizontal dimensions. It's probably the same for Inkjet printers and plotters.
Laser printers can also have some distortion across the width of the page because of errors in the lenses and mirrors.
As you mention, the feed mechanisms may slip a little, or the rollers may change dimensions due to temperature or accumulation of dirt, or rubber rollers may shrink.

Good PCB editing software tools should allow you to adjust the X and Y scales to compensate for errors in the printer.
Some laser printer drivers also allow for fine adjustment in scale, although normally apply an equal scale factor to both height and width.
Some Gerber tools also allow you to expand or contract the X & Y scales to compensate for different output devices.
If you don't have tools to do this, you may be able to manually add a few extra commands to the start of the PCB Gerber file using a text editor as I seem to recall there are scale factor commands in the Gerber standard (???).

Using a CNC machine for photo plotting sounds like a good idea if it's fast enough. Have you tried using a ink pen in place of the milling bit? I have been tempted to try this but I haven't found a good ink that will stick to copper and is waterproof, and doesn't run.
It may be easier than photo resist and would work on plain PCB.

Cheers,
Mark


-----Original Message-----
From: Homebrew_PCBs@yahoogroups.com [mailto:Homebrew_PCBs@yahoogroups.com] On Behalf Of Slavko Kocjancic
Sent: Tuesday, 17 September 2013 06:15 AM
To: Homebrew_PCBs@yahoogroups.com
Subject: Re: [Homebrew_PCBs] Re: Imaging PCB with LED

...so to clean up the problem, and not to argue...

I can make small board with toner transfer metod. Fine traces aren't
problem.
I can make small board with sensitived board and printed transparency on
laser Fine traces aren't problem.

... But I have problem with accuarcy. If I draw rectangle 200mm x 200mm
and then print that and then measure that, the dimennsion is not right.
With few prints made with same settings I measure the side from 199.5mm
up to 201.3mm. I tested 3 printers I own, and LaserJet 1018 and 1020
works identicaly. The Brother HL2030 is little better (from 200mm to 201.1)

So If I pick some big SOIC or TQFP with side of 20mm the error is aprox
aprox 0.1mm. It's bare visible with naked eye. So no problem with that.

Then I pick some chip in DIL40. The last pin is aprox 0.4mm off the
hole. If holes are drilled oversized then package should fit, else wil NOT.

Now I have terminal blocks on the edge of board. I have 20 pins stacked
(200 mils) and in last pin the difference can be 1mm and it's not
possible to fit that component.

And other thing is that I drill with CNC. And of course if one side of
board is ok the other is not.

Why the print's isnt consistent I don't know.
For toner transfer (printing on paper) It can be humidity, but when
printing on plastic transparency I think the humidity can't be problem.
As image is more distorted in height I assume that it's printer
mechanics problem.

As I know that my CNC is way under 0.1mm precission It's logical for me
to try to use it as photoploter.
The process is slower but got accuracy. So If I need accuracy (large
board) I should go with photoploter style, for small board's the toner
transfer is better choice...

I know that my english isn't perfect, but think that majority of people
will understand me. ...ker ne govorim kitajsko.


Slavko.


------------------------------------

Be sure to visit the group home and check for new Links, Files, and Photos:
http://groups.yahoo.com/group/Homebrew_PCBsYahoo! Groups Links

Hello...
I do use compensation (ie image scalling to 99.8%) as in that case I got
best result.
The anyoning part is that error is not constant. It's varies from sheet
to sheet. And error seems to be sum of paper stretching (humidity and
temperature change) slipping in printer and optics difference. I found
that lines (width in paper) is far more consistent than height. So
scanning mirror in laser doesn't make big error, but paper transport
mechanism does.
Year's ago I try with pen in CNC but all pen's tested doesn't have
precission I wanted. The consistent width of line is problem. As I want
line in range of 8 - 10 mils I can't find the such pen.. For lines over
1mm there are no problem...

On 09/17/2013 06:57 AM, Mark Harrison wrote:
> Hi Slavko,
>
> It is very common for laser printers to have errors in vertical and horizontal dimensions. It's probably the same for Inkjet printers and plotters.
> Laser printers can also have some distortion across the width of the page because of errors in the lenses and mirrors.
> As you mention, the feed mechanisms may slip a little, or the rollers may change dimensions due to temperature or accumulation of dirt, or rubber rollers may shrink.
>
> Good PCB editing software tools should allow you to adjust the X and Y scales to compensate for errors in the printer.
> Some laser printer drivers also allow for fine adjustment in scale, although normally apply an equal scale factor to both height and width.
> Some Gerber tools also allow you to expand or contract the X & Y scales to compensate for different output devices.
> If you don't have tools to do this, you may be able to manually add a few extra commands to the start of the PCB Gerber file using a text editor as I seem to recall there are scale factor commands in the Gerber standard (???).
>
> Using a CNC machine for photo plotting sounds like a good idea if it's fast enough. Have you tried using a ink pen in place of the milling bit? I have been tempted to try this but I haven't found a good ink that will stick to copper and is waterproof, and doesn't run.
> It may be easier than photo resist and would work on plain PCB.
>
> Cheers,
> Mark
>

The 150mW laser cost 50eur, the 3W led is 2eur.
I don't know how hard is to focus LED but I select 3W as this is made on
one die, so it's more close to point source. I respect Volkan job as
it's perfect result, and waiting 20 minutes is not bad if result is
known to be good.
About apertures.. For raster scanning the aperture should be very fine,
as you say in range of 0.05mm or les. But for vector scanning the finest
aperture is the same width as finest trace (0.2mm for example)
So seems that noone was already try to use hi power led.. (I found few
experiment on net with low power UV leds in range of few miliwats to be
unsucess..) Seems that I need to try what is all about. (LED already
ordered...)

On 09/16/2013 08:11 PM, designer_craig wrote:
> The 150mw or even higher power UV lasers are available on Ebay, they are used in BlueRay DVD players. An LED would work but it will take a very long time to expose a board. You need lots of joules/sq mm to activate the resist. The exposure spot area is going to be around .05mm to get good pattern definition and its just not going to be possible to mask and focus and LED to that level with any kind of power. Most higher power LED's have multiple junctions which means they are not point sources and such will not focus to a point.
>
> The laser Volkan was using was in the $30 range from Ebay, its best to use one with a glass lenses for better focus and less loss.
>
> The laser module with heat sink was mounted into an ink jet printer shuttle, the board moved in the Y direction on a custom built bed below the laser. The bed was driven by a stepper motor driving a lead screw. He built custom electronics to provide the laser video modulation it was was timed by the plastic head reader strip inside the ink jet printer. As the head moved across the carriage the laser was turned on and off. Once the head scanned across the board the bed was moved for the next pass. He only wrote in one direction because of technical issues with reading the plastic sensing strip.
>
> Even with the carriage moving at full speed in both directions it took about 15 to 20 minutes to write a 150mm x 75mm board, but the image was spectacular.
> Craig
>
> --- In Homebr

If you're thinking of vector drawing on your CNC then it might be worth checking out the pcb-gcode Yahoo group. The software is an extension to Eagle for isolation routing your board. It is intended for use with a mill but should work just as well for a laser / focused LED and allow for compensation for the beam size. Of course if you're not using Eagle it's no use to you.

I've got a 40W CO2 laser and might try this at some point for burning paint off the copper rather than using it with a photosensitive board. Too many other things in the pipeline at th emoment though.

http://nothinglabs.blogspot.co.uk/2013/05/2-layer-circuit-boards-using-laser.html

This also might be of interest

http://hackaday.com/2012/08/09/exposing-pcbs-with-a-home-made-laser-printer/

Slavlo: regarding your print dimensions,..

I had a very simliar problem, printed copper and solder mask, etched the board and and then the solder mask did not line up anymore, had up to 3mm on 200mm deviations. And it turned out to be the humidity, makes the paper shrink and expand. Anywho a quick test you can do is draw/print marks on the paper heat it up in the oven for 10 minutes ~80C and measure again

So what I do now is dry the paper in an oven then print the mask and dry it again every time I use them. BTW humidity is generally over 80% here all the time

On Mon, Sep 16, 2013 at 2:15 PM, Slavko Kocjancic <eslavko@...> wrote:

...so to clean up the problem, and not to argue...

I can make small board with toner transfer metod. Fine traces aren't
problem.
I can make small board with sensitived board and printed transparency on
laser Fine traces aren't problem.

... But I have problem with accuarcy. If I draw rectangle 200mm x 200mm
and then print that and then measure that, the dimennsion is not right.
With few prints made with same settings I measure the side from 199.5mm
up to 201.3mm. I tested 3 printers I own, and LaserJet 1018 and 1020
works identicaly. The Brother HL2030 is little better (from 200mm to 201.1)

So If I pick some big SOIC or TQFP with side of 20mm the error is aprox
aprox 0.1mm. It's bare visible with naked eye. So no problem with that.

Then I pick some chip in DIL40. The last pin is aprox 0.4mm off the
hole. If holes are drilled oversized then package should fit, else wil NOT.

Now I have terminal blocks on the edge of board. I have 20 pins stacked
(200 mils) and in last pin the difference can be 1mm and it's not
possible to fit that component.

And other thing is that I drill with CNC. And of course if one side of
board is ok the other is not.

Why the print's isnt consistent I don't know.
For toner transfer (printing on paper) It can be humidity, but when
printing on plastic transparency I think the humidity can't be problem.
As image is more distorted in height I assume that it's printer
mechanics problem.

As I know that my CNC is way under 0.1mm precission It's logical for me
to try to use it as photoploter.
The process is slower but got accuracy. So If I need accuracy (large
board) I should go with photoploter style, for small board's the toner
transfer is better choice...

I know that my english isn't perfect, but think that majority of people
will understand me. ...ker ne govorim kitajsko.

Slavko.

Most printers ( and copiers)) enlarge the output to 102% to prevent the edges from having a corona.

Try scaling back to 98% and check your measurements.

I know that my english isn't perfect, but think that majority of people
will understand me. ...ker ne govorim kitajsko.

Majority is irrelevant here. What is important, is whether you will be understood by anyone that can offer you useful answer.

The problem with the LED is the light is pattern is most likely 110 to 140 degree. Even if you had a 20X objective lens below your aperture mask the mask hole for a .128mm spot is only going to be 2.5mm. Without some serious low loss collimination of the LED energy you are not going to get very many joules of LED light energy on the Riston. Riston publishes the exposure energy in joules required, do the calculations. The nice thing about the laser is all the energy is sort of in the same direction to start with.
Craig

--- In Homebrew_PCBs@yahoogroups.com, Slavko Kocjancic <eslavko@...> wrote:
>
> The 150mW laser cost 50eur, the 3W led is 2eur.
> I don't know how hard is to focus LED but I select 3W as this is made on
> one die, so it's more close to point source. I respect Volkan job as
> it's perfect result, and waiting 20 minutes is not bad if result is
> known to be good.
> About apertures.. For raster scanning the aperture should be very fine,
> as you say in range of 0.05mm or les. But for vector scanning the finest
> aperture is the same width as finest trace (0.2mm for example)
> So seems that noone was already try to use hi power led.. (I found few
> experiment on net with low power UV leds in range of few miliwats to be
> unsucess..) Seems that I need to try what is all about. (LED already
> ordered...)
>
>
> On 09/16/2013 08:11 PM, designer_craig wrote:
> > The 150mw or even higher power UV lasers are available on Ebay, they are used in BlueRay DVD players. An LED would work but it will take a very long time to expose a board. You need lots of joules/sq mm to activate the resist. The exposure spot area is going to be around .05mm to get good pattern definition and its just not going to be possible to mask and focus and LED to that level with any kind of power. Most higher power LED's have multiple junctions which means they are not point sources and such will not focus to a point.
> >
> > The laser Volkan was using was in the $30 range from Ebay, its best to use one with a glass lenses for better focus and less loss.
> >
> > The laser module with heat sink was mounted into an ink jet printer shuttle, the board moved in the Y direction on a custom built bed below the laser. The bed was driven by a stepper motor driving a lead screw. He built custom electronics to provide the laser video modulation it was was timed by the plastic head reader strip inside the ink jet printer. As the head moved across the carriage the laser was turned on and off. Once the head scanned across the board the bed was moved for the next pass. He only wrote in one direction because of technical issues with reading the plastic sensing strip.
> >
> > Even with the carriage moving at full speed in both directions it took about 15 to 20 minutes to write a 150mm x 75mm board, but the image was spectacular.
> > Craig
> >
> > --- In Homebr
>

I know that this can (will) be problematic. As angle of led is around
120 degrees it's unusable as is. If I make aperture of 0.2mm at the led
then beam 1mm far from led is over 3mm wirdh. So I need to colimate that
light to be more paralel. But the 1'st test to be made is if LED can
develop emulsion at all. So if "flash" of 0.5 second from 2-3mm distance
is enought the project may sucess..

On 09/18/2013 06:43 PM, designer_craig wrote:
> The problem with the LED is the light is pattern is most likely 110 to 140 degree. Even if you had a 20X objective lens below your aperture mask the mask hole for a .128mm spot is only going to be 2.5mm. Without some serious low loss collimination of the LED energy you are not going to get very many joules of LED light energy on the Riston. Riston publishes the exposure energy in joules required, do the calculations. The nice thing about the laser is all the energy is sort of in the same direction to start with.
> Craig
>
>