Electronics > Beginners

How does photoresist attach to a PCB during manufacture?

(1/7) > >>

Richardcavell:
Hi,

I’ve been watching YouTube videos of people handmaking their own PCBs. In some cases, it seems as though the copper is put under a UV light with a mask in the way (artwork printed on transparent plastic). Then the board is immersed in “photoresist” fluid.

What happens to the copper under the influence of UV light? How does the photoresist fluid know where to attach to the board?

Richard

Siwastaja:
You got it the wrong way. Board (all of it) is coated with the photoresist before exposure. Exposure to light softens or hardens (depending on the photoresist type; positive or negative) those parts of photoresist where the light hits. The liquid the boards are dipped into after exposure as you have noticed is called "developer", it simply washes out the soft photoresist. Hardened photoresist remains and protects the copper from acid during etching.

EPAIII:
Just to further clarify, the photoresist is like a spray paint. I have purchased it in rattle cans, just like Krylon or Rust-Oleum spray paint. You spray it on and let it dry - in the DARK, of course. It clings to the copper just like paint would. Then the board is ready for the exposure with a positive or negative photographic image of the circuit.

One more thing. Ordinary photographic film is not dense enough. The UV light will penetrate what are black areas for normal photographic purposes. A high contrast film is needed. Or a high contrast image from a laser style printer. Ink jet printers do not make a dense enough black.

Siwastaja:

--- Quote from: EPAIII on March 28, 2023, 04:13:19 am ---Just to further clarify, the photoresist is like a spray paint. I have purchased it in rattle cans, just like Krylon or Rust-Oleum spray paint. You spray it on and let it dry - in the DARK, of course. It clings to the copper just like paint would. Then the board is ready for the exposure with a positive or negative photographic image of the circuit.
--- End quote ---

Additionally to sprayed or dipped photoresist formulations, a roll-type photoresist is very common. It comes in big rolls, where protective film is removed, and photoresist laminated on the PCB using pressure and heat. (I have successfully used just an office laminator. I find that working with the laminated photoresist is much easier and more repeatable than the spray types where you have to get the spraying process just right, and it affects everything else.)


--- Quote ---One more thing. Ordinary photographic film is not dense enough.
--- End quote ---

What do you mean by ordinary photographic film? Who has access to large format photographic film these days, even? If you could, I would place a bet that a standard B/W photographic film of any type would be dense enough, no problem at all. People did use exposed film to look at the Sun directly (not that I'd recommend doing it).


--- Quote ---Or a high contrast image from a laser style printer. Ink jet printers do not make a dense enough black.
--- End quote ---

It's usually the opposite. Laser printers tend to suck colossally in getting enough density, in fact I have never seen a high-quality laser-printed film, they vary between unusable and barely OK. There are sprays that improve the density of laser prints and they tend to work wonders. On the other hand, many inkjet printers can be configured (in the driver settings) to increase the amount of ink used, which will get you much better density than most standard office laser printers. But oh boy those inkjet transparency films are expensive.

When I did this PCB stuff regularly and wanted to do it properly, the only solution that really worked well was an expensive Epson photographic inkjet printer, those expensive inkjet transparencies, laminated dry film (negative) photoresist, point-source UV light, ½oz copper clad, and vigorously agitated (large amounts of compressed air) cupric chloride etching tank to minimize edge effects. That combination was enough to get 6/6 mil and near-100% yield, and demonstrate 4/4 mil with acceptable yield.

Another issue I noticed with laser printers was unrepeatable geometrical distortions, even within page, due to the heat used. (Cheap) inkjets also produce distorted images but at least it keeps the same for every page printed so you can compensate in the artwork (I just used GIMP shear effect and resize functions for this.)

Also, you usually don't need to work in dark, as the photoresist materials are designed to be only sensitive to near-UV. Just remember that fluorescent tubes release a bit of near-UV, and photoresist materials have a little bit of sensitivity for deep blue visible light as well. Sunlight is also no-no. You can simply reduce the amount of light, preferably some incandescent or low power warm white LED lights, which is still enough for you to work conveniently and safely.

donlisms:
(Large format photographic film is not *entirely* gone these days!)

Just as a detail, I believe it's the developer that does the hardening, not the exposure.  Then I would expect another wash of some kind to remove the unwanted coating.  This is very similar to the dye transfer photographic process, which was the most difficult and complex, but also the most beautiful. 

Anyway, what I was going to say... if you're using a material that isn't dense enough to control the exposure, you may be able to double it up.  Use two layers and align them carefully, or tweak the printing process somehow, something like that.

If I was still doing that rather than having boards made for me VERY inexpensively, the first thing I'd do is some testing.  Use all the advice about this or that light and material and time and such as a starting point, and the invest in some tests to figure out what's really going on, and deal with the frustration of bad results once and for all.  A great help for exposure testing is a "step tablet," a graduated gray scale that you can expose through.  Every step of .3 in density is the equivalent of cutting the exposure time in half, so you expose *once* at your longest estimate, and you get results for every time from there on down.  Develop. Pick the best step. Use that time. 

Navigation

[0] Message Index

[#] Next page