Some info I read re UV exposure resist undercutting, with questions..

[FuZZ1L0G1C]

Hi. I've been doing a lot of web research lately, regarding the best method of creating clear, sharp PCB traces at fine resolution.
Currently, I have no sophisticated UV equipment or materials, which is why I'm researching what to get.
My current modus operandi is the old-fashioned and clumsy Dalo-resist pen and rub-down pad / track transfers, and Ferric Chloride etchant.
A (clean) plastic cat-litter tray is used for larger boards, and Tupperware "lunchbox" for mini-PCB's.
These are silicone'd with perforated fish-tank airlines, fed by a fish-tank filter oscillating air-pump.
This 'bubble-tank' matrix helps the etching process.
An unused (inherited) dinner-server tray warmer plate thingy on low temp keeps etchant at lukewarm temperature.
A home-made 12 V solenoid is used for auto-rocking the entire warmer / tray, but due to resonance wave buildup splashing, I generally use occasional hand rocking.
Reading E.P. posts and other websites, the Laser toner iron-on transfer method seems to create fine details.
I'm unsure whether my two Samsung Laser printers have suitable toner, or if the print transfer materials mentioned are even available.
UV, on the other hand, seems (to me) to be a more reliable method, without 'dissing' the iron-on toner method.
(As a noob to advanced PCB production, these are just my personal assumptions).
Many eons ago, I had a B&W photographic darkroom, with Enlarger, trays, developers, etc.
Therefore I grasp the concept of the UV / photo-resist method.
I wonder if an enlarger / slide projector would work? Or bright sunlight?
Suppose that UV lamps are not available for the former, and the latter (sun) would be too 'spread', causing undercutting.
Maybe the sun method could utilize a black 'hood' to block side radiation?
We have a PCB service via a nearby Electronics Component Supplier, but my idea is to make my own in the evenings, on weekends, or just to create anytime 'on-the-fly'.
I was thinking of building my own UV source using several portable 'money detector' (3 W?) UV tubes stripped from their plastic "torch" housings.
Another UV source is the 220 V AC 60 W "black-light" globe sold at some audio / Hi-Fi retailers as part of "DJ/disco" accessories.
A website I was browsing (?Wiki How-To?) also mentioned face-tanning UV lamps.
The articles also mentioned the problem of undercutting due to offset light sources "creeping" under the non-exposeable masks (see my explanatory sketches).
Now I have a few questions which I didn't see answers to elsewhere (regarding UV exposure method):
Does one need a blacked-out darkroom, or will photo-resist board be safe under normal indoor lighting (combo of 1 x 100 W incandescent standing shaded floor-lamp and 6 x energy saver 'fluorescent' ceiling globes, individually switchable).
The article re undercutting makes me wonder - is a flipped printed image placed face down on photo-resist, or is there a clear acetate thickness between (positive image up)?
Regards,
Clive.

 

[Harald Kapp]

Does one need a blacked-out darkroom, or will photo-resist board be safe under normal indoor lighting (combo of 1 x 100 W incandescent standing shaded floor-lamp and 6 x energy saver 'fluorescent' ceiling globes, individually switchable).

It depends on the duration of the exposure to the indoor light. A few minutes are usually no problem. For storage keep the boards in a dark place - last time I saw such board it came in a black plastic pouch well suited for storage.

The article re undercutting makes me wonder - is a flipped printed image placed face down on photo-resist, or is there a clear acetate thickness between (positive image up)?

You can do it both ways. Obviously havin the print diretly next to the photosensitive layer minimizes undercut. Undercut is also a matter of distance between the lamps and the pcb: higher distance means less undercut as you can easily imagine from the picture you posted.
With traces >= 200 mil wide, transparent film of the kind used in photocopiers and a distance >30 cm I have had negligible undercut (that's been years ago when I still made my own pcbs ).

 

[kellys_eye]

You also get undercutting if you take too long to etch the board so fresh solution is very important if you want to etch finer lines.

For transparencies it is always preferable to have the exposure (print) side against the resist.

Most pre-sensitised boards I use come with peel-off black plastic covers but pre-sensitised boards also have a 'use by' date so always purchase as-and-when rather than keeping stock.

My own exposure unit was made from 6 tubes from a face tanning unit (mounted in an old flatbed scanner case)..... I've seen good results from UV LED units (care taken to get the distance from LED to PCB correct based on beam angle of the LED itself etc).

Bigclivedotcom (Youtube) has done some videos on experiments in track width using various forms of exposure etc - worth Googling/watching.

 

[FuZZ1L0G1C]

Thank you re valuable advice - off to buy transparency sheets tomorrow from dealers.
Tried a test-print via black laser on clear lamination sleeve - amazing clarity / sharp edges, but no good for true transparency, as it warps due to (middle) glue setting in a curved (paper-path) configuration!
Still, it demonstrated that laser should work well on proper OHP.

 

[Bluejets]

With reference to long etching time resulting in undercut, some of the older board (not coated) seemed to have a thicker copper layer than ones I bought recently.

 

[kellys_eye]

Copper coating is identified by the weight - 1oz, 2oz (1 ounce, 2 ounce etc) and IIRC this refers to the weight per square foot. Obviously a 2oz copper is twice as thick as 1oz copper.

You might specify 2oz copper for boards designed for higher current handling but, generally, 1oz copper coated is the usual stuff.

 

[hevans1944]

One possible remediation to the under-cutting problem during copper etching of the PCB is to use a slightly pressurized stream of heated etching solution directed perpendicular to the board via spray nozzles. IIRC, we purchased a heated ferric chloride tank from Kepro with this arrangement in the 1960s or early 1970s.

Kepro, alas, is no more, but the concept is simple enough for a DIY build. This approach will speed up the etching process considerably, and minimize under-cutting by limiting the amount of time the under-cut edges are exposed to etching solution. Another possibility is to use a soft "acid brush" to manually agitate the etching solution against the circuit board, exposing fresh copper to the etching solution and thereby speeding up the etching process.

An even simpler and less expensive solution is to e-mail the artwork to a commercial PCB fabricator. Turnaround time, even to Asia, is usually less than a week. Of course this eliminates the "instant gratification" and broken carbide drill bits that result from a purely DIY PCB fabrication.

 

[Hopup]

I did have very good results with laser printer and good transparencies when I did some boards myself years ago.

Spray etching machines seem to be very good for proto use and they could be easily made diy. I'm not sure if heavans referred them in his post.

 

[(*steve*)]

Another very important consideration for undercutting is the thickness of the copper layer.

Using thin 1/2 oz copper will allow you to etch faster and this undercut proportionately less.

Commercially made boards are not etched and don't suffer this problem.

 

[kellys_eye]

Commercially made boards are not etched

??? - new one on me! Explain!

 

[(*steve*)]

Take a look at this:

The interesting part runs from 14:30 to 18:20

Whilst etching is involved, it is not used to etch the full thickness of the copper.

 

[hevans1944]

Whilst etching is involved, it is not used to etch the full thickness of the copper.

Indeed. The "magic chemicals" that the drilled board is dipped in to form a conductive film inside the vias can also be used to form a conductive film over both sides of a bare FR4 substrate. The electroplating of copper in the vias then also simultaneously plates a thin layer of copper onto both sides of the board. It is this thin layer that will eventually be etched away.

IIRC, the initial plating (or foil lamination if that is used) is very thin and serves only to support the photo-lithographic mask that defines where the conductive traces will appear. And of course the initial plating defines where the conductive vias are. After the mask is applied and developed, electroplating of copper continues until the customer-specified thickness (or weight in ounces per square foot) of copper is obtained. Some empirical data regarding via hole drill size is used to ensure that the correct final inside diameter of the via is obtained at various specified copper thicknesses.

In essence, only the initial thin film of plated copper must be etched away. Etching occurs after the over-plating of tin on the copper traces and inside the vias, said plating acting as a etchant resist after the chemical removal of the original photo-lithographc exposed and developed resist coating, whose only purpose was to define where electroplating deposits of copper and tin were made.

Starting with bare substrate stock has some advantages, in terms of process and quality control, although it does add a layer of complexity since the panels have to be rinsed, dried, and coated with photo-resist after the initial "flash" electroplating, before electroplating to final copper thickness followed by electroplating with tin can occur.

Whether the savings in not having to stock copper foil-laminated substrates is worth the extra rinsing and drying step would depend on the volume of product produced. I doubt it was a factor in my local PCB fab in Dayton, OH, going out of business in the 1980s. Rumor has it the EPA regulations regarding used chemistry recovery and disposal were the main factors that made their operation unprofitable. Gold-over-nickel plating for edge connectors was also part of the PCB fab process and the chemistry for that is pretty nasty.