Q's about making PCBs

[Active8]

Hi Mick...

I've been down the road that you are embarking upon. Years ago, I
starting making my own PCBs in house here for low runs on odd projects
and such. And it sounds to me you are on the right path. I'll give
you a few tips that I found out through trial and error that may be
helpful for you to look for while your going about the whole process.

Sodium persulfate is definately the way to go just because it isn't
nearly as messy and toxic as ferric chloride. Although, I did find
that the sodium persulfate has a shelf life. I bought mine in a
chrystalized form and even mixing it fresh from a bag, I noticed it's
cutting power deteriorated with each new batch I did from month to
month. And as a previous poster pointed out, it is necessary to heat
it to where it is somewhat steaming, but not boiling. The
manufacturer will have an exact temperature to use. Use an accurate
thermometer and a big glass baking dish on a small hot plate to get it
heated.

We got one from a hardware store (Pyrex!) to heat our Tin solution
and it shattered on the hot plate, so whatch out! It must have been
cooled wrong at the factory. Ask Rob Neilson the pyrex spraying
specialist. He should be blathering on at alt.guitaramps or whatever

Read up as much as you can and then just go for it. Be prepared to
screw up some boards.

Think positive. My only screwups have been the artwork. Forgetting
to manually add/remove a trace in CAD, or more frustrating, drawing
the resist on a double sided board with a marker and connecting to
the wrong pin - like connecting power to ground.

I've had great success with the etching, but I've always used FeCl
III since it was easier and is storable. That toner stuff took some
practice, but the boards can be as pristine as those produced from
photo methods.

We all have done it and in the learning
process, you're prone to do it as well.

Think positive. Assume success but think ahead to avoid screwups.

Have fun!

Yes. If you do screw up, remember that someone said PCB fab is a
bore, so it's not like you missed anything. When you have a pristine
board in your hands that soldered up as smooth as silk and the
circuit works, then you've had fun.

 

[Active8]

Yes! I've thought of that myself. I'm thinking, a device designed to
"count" UV light (ie amount of light * time) could determine exposure time
accurately enough so that you could use ordinary sunglight (or, hell,
anything) instead of a UV box. Anybody tried this? What kind of exposure
times? (I don't have a backyard so I'd have to keep my eye on the thing)

Thanks for the info

Mick.

You can calculate the appoximate times knowing the wattage of the
light, the area exposed, and the resist's spec.


2 * resist_E_spec_in_mJ_per_sq_cm / (Watts / cm^2)

that's an estimate of the seconds required.

Here's one guy's method. Use a test board. Print a series of letters
on a transparency. Cover all but one letter with something opaque
and turn on the UV. Expose one extra letter each second or so after
the minimum estimated time for up to the max time. Develope the
exposure, etch, and you'll have your answer. Better yet, instead of
letters, use lines of various widths and pads - little test coupons.

 

[Active8]

See www.electricstuff.co.uk/pcbs.html for tons of info

.

I wouldn't bother with the router - a Dremel won't cope with the axial load anyway. If you can't
afford a shear, a hacksaw is adequate. For small cuts, heavy duty scissors/tinsnips can be used.

A good paper shear works, too.

 

[Active8]

What are the router bits for then?

Sissy thin wood projects Actually, a good Dremel can be pretty
tough. I've used the cutoff disks on pcbs but it isn't great if you
cutt all the way through. You can score the board and snap it. You
can even do that by scoring with a utility knife. You know your
pushig your tool to the limits if your feed rate is so fast that the
RPMs drop a bunch. Cutting tools get hot, etc.

What kind of axial load can a dremel cope with? How do you prevent from
damaging it by loading it too much?

I wouldn't put so much force on it that the cutting tool shaft
flexes or the RPMs drop more than I'm comfortable with. Dremels
aren't too expensive and they're tough. You could try breaking one

I have an old high speed drill here
that I've ruined by pushing sideways too hard with it

Try using the right sized bit next time. Ovals? Elongated slots? Not
a job for a drill.

You can get 0.4mm and 0.6mm thick boards. I'm thinking you could etch a
single-sided 0.6mm, a double-sided 0.6mm and a single-sided 0.4mm and
somehow stick them together to make a 4-layer 1.6mm board. Is this
feasible?

Sure, but you'd have to plate the holes because you can't solder the
inner layers. "Sticking together" is done in industry with a prepreg
layer and a heat press.

Okay, maybe I'm cutting ahead of myself a bit here since I HAVEN'T MADE A
SINGLE BOARD YET, but I'm still curious.

Cool. Try Coomb's "Printed Circuits Handbook" for the masochist in
you.

 

[Active8]

I've seen that page, but thanks anyway

Okay, I had a bit more of a read about plating vias, and essentially it
sounds like too much of a pain.

I've read old posts of people saying they'd solder bits of wire into each
hole on both sides, or weave a big long bit of wire through all of them
and solder it in one go. Sounds tedious and would probably look crappy
too.

What I wouldn't mind is some way of getting a blob of solder in the holes
and sticking to the via pad on each side.

Have fun.

I think this is what they used
to do a couple of decades ago (I vaguely recall seeing boards like this).
Anyone tried this? Is there a clever way to go about it?

Yeah. It's called plated through hole and wave soldering.

 

[Active8]

I've seen that page, but thanks anyway

Despite his hatred of NaOH as a developer, many still do it. The PC
Handbook, IIRC lists sodium/potassium carbonate as a developer and
sodium/potassium hydroxide as the stripper for aqueous processible
resists. Thus the warnings about not developing in NaOH for too long
and not touching the resist.

 

[Wayne]

Whats wrong with Ammonium Persulfate? Its clear, so its easy to see the
board as its being etched.

Al

getting the temperature of the solution just so .... if it's not quite right
it won't work, it's availability is in small 400g containers, it's a use
once and discard

 

[JeffM]

plated vias

Have fun.
Active8

Yup. The only reliable way is to poke a copper wire thru
and tack a significant length of it on each side.
....and, as Active8 pointed out in another part of this metathread,
this only works for 2-sided boards.

 

[Active8]

Yup. The only reliable way is to poke a copper wire thru
and tack a significant length of it on each side.
...and, as Active8 pointed out in another part of this metathread,
this only works for 2-sided boards.

Actually, someone else mentioned the thru wires. I just said, "Sure,
but you'd have to plate the holes because you can't solder the
inner layers."

As I'm sure you know, the PTH is done after the stackup is assembled
and desmearing and etchback.

 

[budgie]

A good paper shear works, too.

for the first couple of times :-(

 

[Fritz Schlunder]

Whats wrong with Ammonium Persulfate? Its clear, so its easy to see the
board as its being etched.

Al

Greetings.

I don't know what the chemistry involved with the ammonium persulfate as
etchant is, but I've observed that little gas bubbles form on the PCB where
the etchant contacts the copper. The gas bubbles shield the copper
underneath them from the etchant and so that particular little region of the
board does not etch while the rest (not covered by a bubble) still etches.
The net result is you end up with highly uneven etching if you just stick
the board in the solution and just leave it there for awhile. So in order
to get reasonable results either you have to manually be very religious at
agitating the solution enough to shake the bubbles loose as soon as they
form, or else your etchant tank needs some mechanism to do this for you.
Ferric chloride doesn't seem to produce these gas bubbles, so that is a
distinct advantage.

That said I'm sure both etchants (ferric chloride and ammonium persulfate)
are quite capable of producing good results provided the proper etching
techniques are mastered. My personal preference is ferric chloride.

 

[Animesh Maurya]

Despite his hatred of NaOH as a developer, many still do it. The PC
Handbook, IIRC lists sodium/potassium carbonate as a developer and
sodium/potassium hydroxide as the stripper for aqueous processible
resists. Thus the warnings about not developing in NaOH for too long
and not touching the resist.

Hello Mike,

Have you tried sodium/potassium carbonate as a developer? Does it
works as fine as silicate based developer? I am having plenty of it,
so eager to put it in some useful work now.

Best regards,
Animesh Maurya

 

[Eugene Rosenzweig]

Hello Mike,

Have you tried sodium/potassium carbonate as a developer? Does it
works as fine as silicate based developer? I am having plenty of it,
so eager to put it in some useful work now.

Best regards,
Animesh Maurya

We use sodium carbonate and it works like a charm.

Eugene.

 

[Mikael Ejberg Pedersen]

getting the temperature of the solution just so .... if it's not quite right
it won't work, it's availability is in small 400g containers, it's a use
once and discard

It isn't that bad. I admit that the temperature is important, but I
can't recognise the "use once and discard". I have used it for years
without problems. When I make a fresh solution, it is good for at
least 8-10 euro sized boards. And when not in use, I can keep it
bottled for over a year, and it's still ready to go when heated again.

 

[Active8]

for the first couple of times :-(

Oh. Good thing I didn't spend the bucks. The guy who posted that
didn't mention that it wore out.

Rephrase: I've read that a heavy duty paper shear from the office
supply store works.

 

[Active8]

Hello Mike,

Have you tried sodium/potassium carbonate as a developer? Does it
works as fine as silicate based developer? I am having plenty of it,
so eager to put it in some useful work now.

Best regards,
Animesh Maurya

No. It's cited in the handbook as the developer for aqueous
processible resists. There's no referencc to the silicates that I
can recall finding in the chapter on imaging, but it rings a bell.
Maybe it was mentioned in some web resource.

 

[Active8]

We use sodium carbonate and it works like a charm.

Eugene.

Where can I get one of those charms for comparison purposes?
There's some Wiccans around.

 

[Terry Given]

We got one from a hardware store (Pyrex!) to heat our Tin solution
and it shattered on the hot plate, so whatch out! It must have been
cooled wrong at the factory. Ask Rob Neilson the pyrex spraying
specialist. He should be blathering on at alt.guitaramps or whatever




Think positive. My only screwups have been the artwork. Forgetting
to manually add/remove a trace in CAD, or more frustrating, drawing
the resist on a double sided board with a marker and connecting to
the wrong pin - like connecting power to ground.

I've had great success with the etching, but I've always used FeCl
III since it was easier and is storable. That toner stuff took some
practice, but the boards can be as pristine as those produced from
photo methods.


Think positive. Assume success but think ahead to avoid screwups.

Yes. If you do screw up, remember that someone said PCB fab is a
bore, so it's not like you missed anything. When you have a pristine
board in your hands that soldered up as smooth as silk and the
circuit works, then you've had fun.

After you've spent weeks trying to discover elusive, temperature-sensitive
faults that turn out to be hairline (often literally) gaps in tracks, or
thru-hole links soldered on one side only, or other dodgy "vias" (Ive used
thos DIY via kits, they can work but are expensive and IMO prone to
failure), ie the damn circuit actually worked in the first place, the joy of
pcb manufacture begins to wear pretty thin, and the $200-$500 you saved on
not getting a professional PTH or multilayer PCB manufactured is revealed
for the false economy that it really is. Yet numerous companies persist in
this approach. Very different if youre a hobbyist, but if thats the case,
dead-bug the circuit on a sheet of copper-clad pcb.....few hobbyist circuits
CANT be made this way, and they usually work better than most DIY pcb
layouts anyway (its a ground-plane thing)

cheers
Terry

 

[Terry Given]

See www.electricstuff.co.uk/pcbs.html for tons of info


.

I wouldn't bother with the router - a Dremel won't cope with the axial load anyway. If you can't
afford a shear, a hacksaw is adequate. For small cuts, heavy duty

scissors/tinsnips can be used.


DANGER WILL ROBINSON!

you do ***NOT*** want to breathe fiberglass, even very small amounts. Think
Asbestos. Years ago I worked for a company that bought an LPKF PCB router.
We read the F*ing manual (yeah, I know its wrong , and immediately
invested in a fancy vacuum filtration system with fancy HEPA filter.

A simple technique, used by Ostlers for centuries, is to hum while you work,
ie constantly expel air.....a vacuum cleaner taped to the machine is not a
bad idea too.

Personally, I use my hacksaw. I could use my bandsaw, or my sawbench, or my
radial arm mitre saw, or my jigsaw, or my chainsaw, or my circular saw, or
my tenon saw, or my keyhole saw, or my coping saw, or my cross-cut saw.

cheers
Terry

 

[Active8]

scissors/tinsnips can be used.

DANGER WILL ROBINSON!

you do ***NOT*** want to breathe fiberglass, even very small amounts. Think
Asbestos. Years ago I worked for a company that bought an LPKF PCB router.
We read the F*ing manual (yeah, I know its wrong , and immediately
invested in a fancy vacuum filtration system with fancy HEPA filter.

A simple technique, used by Ostlers for centuries, is to hum while you work,
ie constantly expel air.....a vacuum cleaner taped to the machine is not a
bad idea too.

Personally, I use my hacksaw. I could use my bandsaw, or my sawbench, or my
radial arm mitre saw, or my jigsaw, or my chainsaw, or my circular saw, or
my tenon saw, or my keyhole saw, or my coping saw, or my cross-cut saw.

cheers
Terry

What did they see?