[Repost] A Home Photoetching Operation (long)
Randy Gordon-Gilmore
Thursday.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Based on the email replies I have gotten, there is an obvious interest in my
home photoetching article. It is to be published in the 2mm Scale Association
Magazine (see the website at http://www.luna.co.uk/~gevans/pages/2mm.htm for
information about the Association), and Nigel Cliffe, the editor, has given
his blessing on a public posting. So without further ado...
_A Home Photoetching Operation_
Making the Burlington Pioneer Zephyr in N-scale is what brought me back into
railroad modeling, after an absence of almost twenty years. The Pioneer went
into service in 1934. With its corrugated stainless steel bodyshell and
streamlined design, it was a radical and distinctive departure from the
heavyweight passenger trains of the day. There have been two HO-scale
Pioneers that I know of, both in brass from Japan, but no one has made an
N-scale Pioneer. I realized that, to adequately model the Pioneer, I would
have to photoetch the body panels.
-=Background=-
I started in the fall of 1992 by studying the few articles I could find on
home photoetching, together with a few more articles on DIY printed circuit
boards. (I was not to learn of the extensive British model railroad
photoetching industry for almost two years.)
My first experiments were with hand agitation in a small plastic tray, using
sprayed-on liquid photoresist from a local electronics supply store. I
quickly found this was not satisfactory. The etching was very uneven, and by
the time part outlines were etched fully through, there was severe
undercutting of the resist in areas.
I then built, from acrylic plastic, a tank in which I could hang the brass
sheet vertically, still submerged in the acid, with a pump to circulate the
liquid quickly. (It was a centrifugal pump made for chemicals, with a
magnetic coupling to the external motor.) I added an aquarium heater to keep
the etchant at 40øC. This improved matters, but there was still undercutting
of the resist pattern, and on relief etching, there was a very rough surface
finish.
There were two basic problems. The first was in the design of the etching
tank. I realized that the acid was not circulating quickly enough (even
though the pump was about 4liters/min and the tank only held about 1 liter of
acid). With inadequate circulation, there is a thin surface layer of acid
containing dissolved etching by-products. This "contaminated" liquid seems
to be the main cause of the poor surface finish, which also slows down the
etching process and requires the workpiece to be kept in the liquid longer,
leading to the increased undercutting.
Another problem was with the resist itself. I was not able to get an even,
thin coating, and it was even harder to get a coating without pinholes. From
an article on commercial printed circuit board manufacture, I learned of dry-
film resist. This is a partially-cured, but still tacky, plastic film which
is laminated onto the brass using heat and pressure, and fully cures on
exposure to ultraviolet light. I obtained a sample roll of this resist to
try out.
I also made a tank with a higher-pressure pump, and plastic nozzles to spray
the ferric chloride onto the sheet, with a clip to suspend the sheet
vertically in the tank above the surface of the acid. This worked much
better (about 4 times faster than the circulating tank!) and I even had to
dilute the ferric chloride about 1:1 with distilled water to slow down the
etching action. The surface finish was much smoother also. But, the etching
was "directional" due to the acid running off the surface in one direction.
I had to keep stopping the pump, and rotating the sheet 1/4 turn every so
often to even out the directionality.
Which brings us to the present...
-=The Equipment=-
For applying the resist, I use a Canon "PC Fuser I" office laminating machine.
The fuser is, as far as I can tell, a fixed hollow roller containing a
halogen tube bulb (like in floor lamps) to generate the heat, and a motorized
pressure roller springloaded against it. Brass up to .020"(.5mm) is not a
problem with it, but when I tried running a .062"(1.6mm) PCB through it, it
creaked and groaned, but did successfully laminate the resist. I will not do
that again, though.
My exposure frame consists of two sheets of .06" (1.5mm) thick acrylic with
thick acrylic perimeter frames. The clear aperture is about 5" x 7"
(125 x 175 mm). Clearance holes through one perimeter frame, and tapped holes
through the other, allow me to clamp the two using thumbscrews. There is a
thin rubber gasket between the sheets along the inner edge of the perimeter
frame. Near one edge of the aperture, there is a vent hole which leads out
to one side of the perimeter frame on the "base" half. Airtight plastic
fittings and vinyl hose connect this vent to a "vacuum reservoir" made from
acrylic tube and sheet. This in turn is connected, through a stopcock, with
another fitting which allows attachment of a 60ml plastic syringe. An
expanded polystyrene "plug", covered with dark fabric, fits in the aperture
of each perimeter frame, and acts as a shutter.
My etching tank is made from a piece of 300mm diameter PVC plastic sewer pipe
(which I bought unused), and holds 4 liters of ferric chloride, diluted 1:1
with distilled water from the stock solution bought from the electronics
store. It stands about 1200mm tall, is closed permanently at the bottom, and
has a removable cover. The tank floor is clear acrylic, and is "V" shaped.
(I used clear acrylic on the theory that I could monitor the state of the
ferric chloride, which turns from a rusty brown color to greenish as it
accumulates dissolved brass. However, the liquid is too dense to pass
light...) A 150W aquarium heater near the apex of the "V" keeps the etchant
at 40øC.
There are two nozzles (equivalent to #H-83251-00 on pg. 1123 of the 1995
Cole-Parmer laboratory products catalog) in the wall of the tank directly
across from one another, each fed by an oscillating pump. These pumps
(equivalent to #H-07101-01 on pg. 1113 of the Cole-Parmer catalog) are made
for laboratory use, and have all rubber/plastic internal parts.
Unfortunately, the pumps do not put out enough pressure to completely atomize
the etchant, but the spray is fine enough for the finest etching I do. I can
turn each pump on individually, to fine-tune the double-sided etching. All
plumbing is vinyl tubing with plastic fittings and hose clamps.
A shaft (titanium, to resist the acid, as even stainless steel will dissolve
in the ferric chloride eventually) comes in the side next to one nozzle, so I
can hold the sheet in the middle of the tank, and turn it continuously while
etching. To hold the brass while etching, I drill a ~2mm hole near each
corner. A four-armed titanium wire "spider" with hooked ends engages these
holes, and applies outward tension to keep the sheet from bowing. The
plastic hub of this spider presses onto the titanium shaft.
-=The Materials=-
I have been buying K&S brand brass at a local hobby shop. This is half-hard
material, comes in 4" x 10" sheets (100mm x 250mm), and is available in .005",
.010", .015", .025" and .032" thickness (.12mm, .25mm, .38mm, .62mm, .80mm).
It is very flat, dent and scratch free, but a little on the expensive side.
I have been searching for a metals supplier who will sell small quantities of
bulk material. I would like to try full-hard temper for pieces that do not
require a lot of forming.
I use Laminar LM dry-film photoresist from Morton International. This
material is a soft, .001" (.025mm) thick, light translucent blue partially-
cured plastic film. It has transparent cover sheets on both sides to protect
it before laminating. On exposure to ultraviolet light, it cures to a hard,
dark blue coating. It is resistant to the etching acid, but the uncured
resist is soluble in sodium carbonate (washing soda), a weak base, and the
cured material is soluble in sodium hydroxide (lye), a much stronger base.
-=Cleaning the Brass=-
For cleaning the brass, I use a Scotchbrite (TM) pad, made by the 3M company,
and available at hardware or building supplies stores. It is like a nylon
scouring pad with imbedded abrasive. I use the dark maroon type, which is
intended to smooth varnish finishes on fine furniture. I cut the pad into
squares about 2" x 3" (50mm x 75mm) and use a new piece for every etching
session.
This avoids the mess associated with the generally recommended pumice powder,
which is very hard to completely rinse off the brass.
General Notes:
DO NOT use any cleaner that contains chlorine. It will react with the zinc
in the brass and create gas which will make pinholes in the coating. In fact,
I do not use any cleaner per se.
Only and always use distilled water for all cleaning, diluting, etc.
associated with etching. Initially I did not realize this, and the minerals
in the tap water interfered with both the cleaning and etching action.
Wear rubber gloves for the whole process, not so much to protect your hands,
but to protect the brass from skin oils.
Make a solution of the ferric chloride, diluted about 1:100 with distilled
water from the etching solution. Have it in a glass or plastic container big
enough to immerse the brass piece.
The Cleaning Process:
1. Scrub both sides of the brass under running tap water with the Scotchbrite
pad until the surface is evenly clean and shiny. I put it on a piece of
acrylic plastic during scrubbing so I don't abrade the kitchen sink.
2. Rinse off the brass with distilled water from a spray or squirt bottle.
3. Dip the brass in the weak FeCl solution for 10 or 15 seconds. This will
very slightly etch the surface and make it chemically clean.
4. Rinse off the brass again with distilled water.
5. Blot off most of the water with a fresh, clean paper towel. Do not let
the brass sit stationary on the towel, or it will pick up a pattern like
a watermark from the towel.
6. Finish drying the brass with a blowdryer or other source of warm air.
There should be no watermarks or fingerprints visible.
7. Don't touch the brass with bare hands until it is coated with the resist.
-=The Artwork=-
I do all my drafting on a PC, using VersaCAD software. From the model
drawings, I develop flat patterns for the etched pieces, making allowances
for bends where required. I plot the output in Postscript. Areas to be
etched are black, while the areas to be protected are left clear.
For trial etching and non-critical pieces, I print the artwork directly onto
NewsCraft LGN, a "high-tech translucent laser printer paper for negative
images to use in creating printing plates", which is "specially coated to
improve toner adhesion" and "dimensionally stable under heat and pressure of
laser printing", using a 1200DPI (dot per inch) laser printer.
For the final etches, I send the Postscript files out to be printed on film
using a Linotronics phototypesetter at 2500DPI.
The laser printed artwork does have a slight drawback. Due to the way the
paper is pulled through the printer, no two printouts are exactly alike. The
misregistration between the artwork for the two sides might be .1mm over a
10cm span. I live with this for my test etchings. The film artwork is always
dead on.
I do all my etching from both sides, even on material as thin as .1mm. But
most of my work so far has surface details. The part outlines, and window
openings, are on both the front and back artworks, and detail to only be
etched halfway through is on only one artwork. By the time the windows are
etched fully through, the surface detail is etched halfway through.
For my etching, lining up the two artworks by eye, on a light table (or
sunlit window) is sufficiently accurate. I leave a 1cm or so clear edge
around the artwork, and tape the front and back pieces together along one
edge, making a folder into which the brass can be placed.
-=Laminating the Photoresist=-
I laminate the brass one side at a time. I cut a piece of the resist about
1"(25mm) wider and 2"(50mm) longer than the brass, and remove the protective
sheet from one side. I then lay the brass onto a sheet of plain paper (I
learned the hard way NOT to use old laser prints or photocopies!), and place
the leading edge of the resist slightly ahead of the leading edge of the
brass. On the very clean brass, the resist will start to stick on contact
and produce wrinkles, so I get the leading edge of the resist tacked down
onto the paper, and then hold the rest up off the brass as the sandwich is
pulled through the rollers. It is important to let the brass cool to room
temperature before coating the second side. When the sandwich is cool, I
trim around the edges of brass with a scissors (the resist will be laminated
to the paper all around), and repeat the process with the other side.
I do my laminating at night, with all the lights turned out except for a
yellow incandescent "bug light" in a lamp across the room as a safelight.
This is probably a little darker than necessary, seeing that the commercial
etcher I visited had fluorescent lights, and only a yellow-tinted plastic
tent around their laminating area to filter the ultraviolet from the light.
But I'd rather be safe than sorry.
-=Exposing the Resist=-
I place the coated brass between the front and back artwork, and clamp the
sandwich in the exposure frame. I do this in a room with the curtains drawn,
to minimize inadvertent exposure of the photoresist. I then pull a vacuum in
the exposure frame.
The 60ml syringe is used as a vacuum pump, by opening the stopcock, drawing
the syringe plunger back, closing the stopcock, detaching the syringe,
pushing the plunger home, reattaching the syringe, and repeating. Six to
seven cycles evacuate the exposure frame to ensure close contact between the
artwork and the photoresist-coated brass sheet. The vacuum reservoir ensures
that any small leaks in the gasket do not significantly affect the vacuum for
at least several minutes.
I then take the exposure frame outside and use sunlight to expose the
photoresist. I do all the exposing within an hour of midday. The resist
has a fairly wide latitude of exposure time, but 60 seconds' exposure per
side, assuming a clear day, is about right. This obviously has its
limitations, requiring clear weather on a weekend, but I have yet to find an
adequately intense artificial UV source that is not exorbitantly expensive.
(Commercial photoetchers and PC board houses usually use arc lamps for their
exposures.)
After exposing both sides, I bring the exposure frame inside, remove the
brass, strip off the remaining protective clear sheet from each side, and
soak the piece in a 1% solution of sodium carbonate. Within a couple of
minutes, the unexposed photoresist is dissolved, and I use a soft nail brush
to ensure it is all removed from the brass. I then rinse the piece with
distilled water, dry and wrap in tissue paper.
-=Etching the Brass=-
I preheat the ferric chloride for an hour or so before etching. I use this
time to drill the holes near each corner of the brass. I also use this time
to examine the resist closely, and repair any flaws with lacquer, which dries
quickly and is resistant to the ferric chloride.
I wear a Tyvek coverall, rubber gloves, and goggles when etching. The ferric
chloride is not particularly dangerous, and will not harm skin if washed off
promptly, but it will stain clothing fiercely, and I would not want to get
any in my eye! In case of spills or splatters, a solution of oxalic acid (a
mild organic acid) in water will soak out ferric chloride stains from both
clothing and surfaces. It is mild enough to not harm even cotton shirts
(much to the relief of my wife!) Maybe 5g acid mixed in 50ml of distilled
water.
I monitor the actual etching progress through the clear acrylic cover of the
tank. Etching double sided, .010" (.25mm) brass will etch through in 5-6
minutes. I turn the workpiece a quarter turn every quarter minute or so to
ensure even etching, and turn off both pumps when the etching is complete.
I place the completed etching in a weak solution of sodium bicarbonate
(baking soda) to neutralize any residual acid, while I etch the remaining
workpieces (I usually do 5 to 10 pieces in a session).
To strip the resist, I place the workpiece in a 1-2% solution of sodium
hydroxide. The resist will turn purple, swell and wrinkle, and eventually
lift off the brass. The sodium hydroxide will also tarnish the brass, so I
soak the now uncoated brass in vinegar with salt added (maybe a tablespoon
per liter). This acts as a brightening solution, and within a minute or so,
the brass is clean and shiny. A final distilled water rinse and drying, and
the etching job is finished.
-=Conclusion=-
It has been a long and sometimes hard learning process, but I can now
routinely make etchings of consistent quality. Hopefully, I'll have my
Zephyr model done by the time the prototype Pioneer (which is undergoing
cosmetic restoration) is back on display in Chicago next year. I'm just
finishing patterns now for the compound-curved areas (engineroom roof, and
observation car end), which will be lost-wax cast. I'm planning to do the
final model in nickel silver (which etches just like brass), to simulate the
stainless steel without having to get it plated.
A second major project is a shell for Burlington's streamlined 4-6-4 "Aeolus",
to go on a Kato mechanism. Other than that, I've made window frames, frames
and overlays for Commonwealth tender trucks (bogies!), footsteps, wagon-wheel
antennas, etc. In the works are a windmill, cattle pen fences and gates, and
some Burlington heavyweight passenger carsides, along with shells for a
Milwaukee F7 and NYC "Commodore Vanderbilt", both to go on the Kato Hudson
chassis, for members of the Bay Area NTRAK, to which I belong. I'm sure there
are many other possibilities. I'm starting to model the late 1930's, so I
think in N scale, I could even etch cars and trucks (lorries!) to fold up,
since most road vehicles in those days were still very angular.
I have only recently rejoined the Association, and have corresponded with
several members who sell etched kits. They all have been a great help in
showing me the state of the art in etching (and caused me to go back and
revise many of my in-process drawings!). I'd like to thank Andrew Cox, Nick
Dearnaley, Ted Humber, and Tim Watson.
-=Postscript=-
I am presently working on a larger tank, made from 18"(450mm) pipe, having
found an all-plastic diaphragm pump that will put out 45 psi(.3MPa), to give a
more even spray pattern over a wider area.
Also in the works is a UV exposure unit, based on plans of Think and Tinker,
Ltd. (http://www.thinktink.com/stack/volumes/volvi/pcbproto.html). This will
alleviate the "seasonality" of depending on clear sunlight for the exposures,
and let me etch through the rainy season.