NavigationUser login |
Hold the phones - inkjetThanks to Paul I just remembered another idea. Since the UV laser pulses are too slow for laser use (on further reading, 50% duty cycle is massively optimistic), use an inkjet printer. Have the output of the UV air laser connected to a fibre, and stick the other end of the fibre where the ink cartridge lives. Add some focusing optics on the end, and presto. Need to modulate the output of the laser according to the make-ink-go-now signal, and we're away. So the next question is power density... Say a UV box has two 8W tubes which shine over an A4 sheet for two minutes. Let's assume the tubes are 50% efficient, which they almost certainly aren't. So that's 8W of light over a 0.29m*0.21 m area ~ 131Wm-2. Over two minutes the energy dissipated in the photoresist is about 120*131=16kJm-2 Lets say our laser can put out 1mW. This is a conservative estimate, since some guy guesstimated his nitrogen TEA laser was putting out 200mW. Say we focus this down to a 20um (~1200dpi) diameter circle. That's a power density of 0.001 / (pi x (10 x 10-6)2) = 3.2MWm-2. So how fast does an inkjet scan? Lets say it does one line in half a second. So how long does it dwell on a pixel? Let's go for 1200dpi. That's a pixel size of 4.5x10-10m2. The laser spends 0.5/1200=400us over a pixel. 400x10-6 x 3.2x106 = 1.3kJm-2 So it looks like a 1mW laser isn't powerful enough, but it's pretty close. 10-15mW would be better. However, it looks like this is doable with a nitrogen laser, Crap, one problem - I think inkjets print a row of dots with each sweep of the head. Perhaps old ones don't.
Submitted by jeff on Tue, 08/07/2007 - 17:26. categories [ ]
|
Error?
It looks to me like you divided the half a second only by one inch's worth of pixels (1200). Should really be 0.5/( 0.21/(0.0254/1200) ) = 50us over a pixel...
Doesn't make an amazingly large difference - but still.
correct
yep, you're quite right. so we need an even more powerful laser, or just bolt one on the new CNC table(!)
blog posts on the CNC table to follow.
uv photoplotter
here is a page of a very nice guy :
http://pagesperso-orange.fr/francois.dubrulle/aproposdemoi/index.html
You'll see few pictures of a modified plotter (uw led + optic fiber ) to plot to a film. still the speed is an issue
Awesome
Thanks for the links - both very interesting!
Diffraction grating
I'm interested in this for an entirely different application. I want to use a UV lamp and optomechanical system to project lines on a plate coated with a UV-erasable ink. (Most organic dyes would work) Haven't even started building anything yet but I've done quite a bit of research including on patents for holography and lithography decades ago. Your site gave me some ideas.
Even if I could create fine enough lines from the UV source (LEDs or lamp), I'd still be running into the limitations of the surface/dye/vibration/etc. One thing I found that is neat, is someone converted a CD/DVD from a reflection grating to a transmission grating, thus giving literally 10x better image quality. (the lens of a disc will have internal reflection and such) I just wanted to see how good I could go on my own.
The idea:
1) Have a focus of the UV into a line using either some lens (expensive), or reflective chamber, baffles and a pair of razor blades (cheap but complex).
2) Have an octagonal mirror or such reflect the light to different columns on the sheet.
3) Pulse the light at high frequency to get finer lines until the resolution gets too close to the definition. Read: power/aperature issues will limit this
Not exactly useful for circuits, but highly useful for printing parallel lines. Yes, I checked - everyone of these holographic ideas was patented decades ago or more except the LED part which is covered by some non-holographic patents. It was frustrating to find I had no new ideas. :P They now use a far superior method involving holograms or surface scoring for professional gratings. I just have my mind set on finding a way using UV LEDs or sterilizing lamps.