Followups set to s.e.design.

In sci.electronics.basics Jeff Liebermann wrote:
Yep. I learned the hard way NOT to run the layout and blueprint paper
through the rollers on the Diazit(?) machine.

Ah, a diazo copier. I first, and last, used one in high school drafting
class, around 1989. I remember that it was the first time I had seen a
peristaltic pump, and that after several guys had made copies on it, the
ammonia fumes would start to fill the room. The usual remedy was to
open the windows, even in January in Kansas City. Who needs OSHA or a
MSDS, anyway.

At one company, we did have a UV light, but it was constantly being
"borrowed" by the CEO's son for his psychedelic light show parties.

At a job about 7 years ago, the Panasonic 32" professional flat-screen
monitor (full HD, metal case, serial port, HD-SDI inputs - a few grand
at the time) was absent one Monday morning. After a search of the
building and general denials, the boss was preparing to phone the
police. Just then, a person in the same position showed up, and
admitted that he had taken it home over the weekend to watch the ball
game.

I found one of my layouts from 1985:
http://802.11junk.com/jeffl/PCB-Layout/

In the flesh, but as Rev D, as opposed to the Rev B artwork above:
http://www.amazon.com/Ftg-Data-Syste.../dp/B00456W0L2

In June 1986, you could get this board, an FT-156 light pen, *and* a
free copy of Windows (about what it was worth, then), for the low low
price of $349! Roughly $760 today, according to the bls.gov calculator,
so probably even more than that. Source: ad in the June 10, 1986 "PC
Magazine", probably not available at
http://books.google.com/books?id=pDGnxFyejN4C&lpg=PA313&ots=DAVcJi5l05&pg= PA313#v=onepage&f=false
or maybe even http://is.gd/0iHSD5 .

I've always built breadboards of everything I've done.

At the few places I've worked where they made boards from scratch, it
depended on the board. Simple stuff (microcontroller, a few LEDs)
didn't get breadboarded. Complex stuff (like a several-hundred-watt
1 GHz transmitter) had some sections breadboarded, like the final
amplifier and maybe some of the filtering. Then they would order three
boards, stuff and test, and usually end up doing one spin of the board.
This was recent enough that it was all CAD.

Matt Roberds

On 9/20/2014 11:01 PM, Jeff Liebermann wrote:
On Sat, 20 Sep 2014 18:25:05 -0700, John Larkin
wrote:

I found one of my layouts from 1985:
http://802.11junk.com/jeffl/PCB-Layout/
It's a light pen interface card for the IBM PD as a 16 bit ISA card. I
did a lousy job and am not very proud of it. However, it does show
what was typical of 1970's PCB layout technology. If anyone wants
details or more drawings, please say something as all of this is going
into the trash in a few daze.

Ooh, curved traces. I was taught to never do that, on the theory that
the tape would eventually creep in the corners.

That does happen if one stretches the tape when laying a trace. It's
especially bad with narrow traces. Traces will move, especially if
the layout is left in the sun. I used a rubber roller from my wet
photography kit, to flatten the traces and make sure they're properly
stuck to the mylar. (Incidentally, note that I used acetate instead
of mylar in the above layout. Not a good idea and I forgot why I did
it).

For RF, rounded corners are a problem due to impedance bumps.

From the source you cite:
"If you use a radius greater than three times the line width, you will
have a transmission line that is almost indistinguishable in impedance
characteristics from a straight section."

So where is the problem?


Sharp
corners are equally bad due to reflection problems. The compromise is
a chamfered corner (mitered bend):
http://www.microwaves101.com/microwave-encyclopedia/480-mitered-bends
which unfortunately also makes a tolerable fuse at the bend.

My understanding the reflection idea is also a myth but rather the real
issue is the impedance change due to the added capacitance of the
corner, which is also supported by your reference. Impedance changes
will also cause reflections, but the signal does not reflect from the
corner itself like a light beam.

--

Rick

In sci.electronics.repair Jeff Liebermann wrote:
On Sat, 20 Sep 2014 11:00:09 -0700, John Larkin
wrote:

The best layout people I have worked with were women. True today.

Agreed.

"How It Was: PCB Layout from Rubylith to Dot and Tape to CAD"
http://www.eetimes.com/author.asp?section_id=14&doc_id=1285442

A good light table, and a young body, were necessary for hand-taped
layout.

Yep. I brought in a NuArc light table that I inherited from a
previous print shop adventure. The lighting was superb, fairly cool,
and the table big enough for most PCB's. Something like this:
http://www.ebay.com/itm/321208321135
I dragged it through 2 employers, several long term consulting jobs,
and two home business ventures.

At the time, leaning over the table for hours was not particularly
difficult. Today, it would give me back pains in about 15 minutes.
Yep, a young body was a requirement. As I vaguely recall, the oldest
PCB layout person I knew that did layout on mylar was about 25 years
old.

You also needed a flat table with an overhead UV light, for burning
sepia assembly and fab drawings from the various mylar layers. And a
blueline machine of course.

Yep. I learned the hard way NOT to run the layout and blueprint paper
through the rollers on the Diazit(?) machine. Destroying the mylar
original was not a good thing. I had a sheet of plywood and a loose
glass plate. I would pile everything between the plywood and glass
plate, and take it outside for the exposure. Most of the time, the
registration was tolerable. At one company, we did have a UV light,
but it was constantly being "borrowed" by the CEO's son for his
psychedelic light show parties.

We didn't have trouble with multilayers. We just checked the layouts
(and the film!) a lot. Most boards worked first time; still do. A
couple days of overboard checking pay off.

That was suggested many times. However, the schedule never permitted
it. Management tended to prefer doing things over rather than getting
it right the first time. I was not in a position to change that even
though the damage it caused was obvious to everyone involved.

I still have a few mylar layouts around. I'll post pics if anyone is
interested.

I found one of my layouts from 1985:
http://802.11junk.com/jeffl/PCB-Layout/
It's a light pen interface card for the IBM PD as a 16 bit ISA card. I
did a lousy job and am not very proud of it. However, it does show
what was typical of 1970's PCB layout technology. If anyone wants
details or more drawings, please say something as all of this is going
into the trash in a few daze.

Please do tell more. you just don't see things like this, anywhere else.

On 21/09/14 04:01, Jeff Liebermann wrote:
As I previously noted, my attempts to do one product perfectly the
first time, failed because the PC fab house reversed the component and
circuit sides of the PCB.

I saw a neat variant of that. The circuit board was correct but the
designer hadn't realised that the documentation showed the IC from
the "wrong side", i.e. the way the chip packaging people viewed it.
First samples of ICs, and all that.

Magic smoke left the chips.

Solution: through holes, so mount them on the other side of the
board - problem solved.


Lesson learned: Trust but verify.

Always true!

On Sat, 20 Sep 2014 18:25:05 -0700, John Larkin
wrote:

On Sat, 20 Sep 2014 17:22:29 -0700, Jeff Liebermann
wrote:

On Sat, 20 Sep 2014 11:00:09 -0700, John Larkin
wrote:

The best layout people I have worked with were women. True today.

Agreed.

"How It Was: PCB Layout from Rubylith to Dot and Tape to CAD"
http://www.eetimes.com/author.asp?section_id=14&doc_id=1285442

A good light table, and a young body, were necessary for hand-taped
layout.

Yep. I brought in a NuArc light table that I inherited from a
previous print shop adventure. The lighting was superb, fairly cool,
and the table big enough for most PCB's. Something like this:
http://www.ebay.com/itm/321208321135
I dragged it through 2 employers, several long term consulting jobs,
and two home business ventures.

At the time, leaning over the table for hours was not particularly
difficult. Today, it would give me back pains in about 15 minutes.
Yep, a young body was a requirement. As I vaguely recall, the oldest
PCB layout person I knew that did layout on mylar was about 25 years
old.

You also needed a flat table with an overhead UV light, for burning
sepia assembly and fab drawings from the various mylar layers. And a
blueline machine of course.

Yep. I learned the hard way NOT to run the layout and blueprint paper
through the rollers on the Diazit(?) machine. Destroying the mylar
original was not a good thing. I had a sheet of plywood and a loose
glass plate. I would pile everything between the plywood and glass
plate, and take it outside for the exposure. Most of the time, the
registration was tolerable. At one company, we did have a UV light,
but it was constantly being "borrowed" by the CEO's son for his
psychedelic light show parties.

You didn't literally need a UV light. A 250-watt warehouse-type
mercury vapor lamp, maybe 6 feet above a table, worked fine.

A plain old "sun lamp" bulb also worked well. Dunno if
those are still readily available these days, since you'd
think that by now most everybody would be clued into the
skin cancer bit. On the other hand you can still buy a Big
Mac on most any street corner, so who knows?

Best regards,


Bob Masta

DAQARTA v7.60
Data AcQuisition And Real-Time Analysis
www.daqarta.com
Scope, Spectrum, Spectrogram, Sound Level Meter
Frequency Counter, Pitch Track, Pitch-to-MIDI
FREE Signal Generator, DaqMusiq generator
Science with your sound card!

On Sat, 20 Sep 2014 20:47:09 -0700, John Larkin
wrote:

I sometimes put test traces with SMAs on boards and TDR them. For a 50
ohm microstrip on a normal board, 90 degree bends are invisible on a
20 GHz TDR.

Oddly, if you say "90 degree bend" to someone doing a "tape up" (using
real tape), it assumes that they are actually producing a radiused
line with the tape, instead of a sharp 90 degree corner. Those were
called "cut corners" to allow for the distinction. Of course with
computah layout software a 90 degree bend is whatever the software
uses for such things, which could be anything from a sharp corner to a
pre-defined radiused curve.

I might be guilt of accepting conventional wisdom without
verification. I've looked at various microstrip designs with a TDR in
the past, but have always concentrated on material transistions
(connectors, component leads, transfomers, couplers, etc) and not
glitches along transmission lines. When I had to make a right angle
turn with a microstrip, I simply chamfered the corner in the
conventional manner and never tested if it made any difference. Were
I do a microstrip tomorrow, it would probably use chamfered corners on
the assumption that it's safer to follow conventional wisdom.

However, it's interesting that you found no reflections from a 90
degree bend (or corner)? I've never even bothered to look, mostly
because I've never had a TDR with enough sensitivity or bandwidth. I
did some Googling for examples and found this:
http://www.embedded.com/design/audio-design/4013429/4/The-HDMI-Design-Guide-A-short-compendium-for-successful-high-speed-PCB-design-in-HDTV-receiver-applications
http://www.embedded.com/design/audio-design/4013429/5/The-HDMI-Design-Guide-A-short-compendium-for-successful-high-speed-PCB-design-in-HDTV-receiver-applications
which shows a TDR display, and follows with layout suggestions
including chamfered corners on traces. The text of the article
indicates that there are corner reflections, but I don't see them on
the scope trace. That's what I would expect, as most of the
reflections are coming from material transistions (connectors, leads,
components, etc).

Vias are visible. What's tough is the transition from
microstrip to an edge-launch SMA. We've spent a lot of time getting
that good, ATLC sims and such.

When I had to go through the board, I sometimes added extra vias to
simulate the width of the strip line. I never tested if it made any
difference.

I had things much easier in the 1970's. Components and products were
much larger. Higher power consumption was tolerated. Standards and
environmental were more liberal or didn't exist. Product life cycles
were longer. Computers were still a design aid and not a necessity.
Designs and protocols just weren't that sensitive to reflections for
it to be an issue worth investigating.

As I previously noted, my attempts to do one product perfectly the
first time, failed because the PC fab house reversed the component and
circuit sides of the PCB.

We used to put PARTS SIDE and SOLDER SIDE in copper text. They need to
read right on the finished board.

Good point and another reason I wouldn't use my PCB for a job
interview. See:
http://802.11junk.com/jeffl/PCB-Layout/Circuit-and-Component-Side.jpg
Note that they have large "CIRCUIT" and "COMP" labels, with the
correct orientation. However, they are outside the board outline,
which means they might disappear from the negative, and certainly will
disappear from the step-and-repeat negative used to produce the actual
PCB's. These labels should have been on the actual PCB. However,
that was easier said than done because of lack of room. I try to put
them underneath a large size component, where hopefully, there are no
traces. That's usually not the case. When I put the labels in the
trace area, the size of the labels are usually too small for the PCB
fab shop to easily see. Despite them allegedly looking for such
labels, they often missed them. When I stupidly used "CIR" and
"COMP", someone managed to misread them, and reversed them anyway. For
this board, I ran out of board space, so I gave up and put the labels
outside of the board area, and hoped for the best.

Notice that the shorting bar needed to electroplate the gold contact
fingers is missing. My fault, but easily fixed by the PCB fab shop
(for a price).

I added some more photos:
http://www.11junk.com/jeffl/PCB-Layout/

Notice the silk screen at:
http://802.11junk.com/jeffl/PCB-Layout/Silk-screen.jpg
Those are NOT computer generated letters and patterns. They were done
with an India ink pen and a collection of templates. Sometimes, I
would use stick type and stick on component outlines, but mostly pen
and ink. Examples of some of the templates used:
http://www.11junk.com/jeffl/PCB-Layout/slides/templates-01.html
http://www.11junk.com/jeffl/PCB-Layout/slides/templates-02.html
Notice the 1:1 templates. Those were used to make a cardboard mockup
of the PCB to make sure that the big parts would fit. When using pen
and ink to make the silk screen, it was necessary to elevate the
template, to prevent ink from running under the template. I had
various schemes for doing that, but mostly it was several layers of
masking tape on the bottom of the templates. One had to be very
careful not to let the masking tape touch recent ink lines or they
would smear.

My examples of stick type or rub on letters is long gone.
https://www.google.com/search?q=rub+on+letters&tbm=isch
There were also machines that would produce such rub on letters in
strips.

For the silk screen, I used Koh-I-Noor pens.
https://www.google.com/search?&tbm=isch&q=koh-i-noor+rapidograph+pens
The pen tips are similar to plotter pens. Fast dry ink was nice, but
clogs to quickly. Slow dry meant that one had to be very careful not
to touch the lines before they were completely dry. The pen also had
to be held perfectly vertical, or it would dump a blob of ink on the
mylar.

--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

On Sun, 21 Sep 2014 01:32:22 -0400, rickman wrote:

On 9/20/2014 11:01 PM, Jeff Liebermann wrote:
For RF, rounded corners are a problem due to impedance bumps.

From the source you cite:
"If you use a radius greater than three times the line width, you will
have a transmission line that is almost indistinguishable in impedance
characteristics from a straight section."

So where is the problem?

The real problem is with using cut corners. Many companies and
schools taught different techniques of making 90 degree turns with
traces. I was taught to use the Xacto knife to cut half way across
the trace, before "bending" the tape 90 degrees to make a corner. That
eliminated a messy looking corner produced then the trace is cut all
the way across, and overlaid with tape at 90 degrees to make the
corner. That produced a radius of about 1/3rd the width of the trace.
However, such corner cutting takes time, and it's much easier to make
sweeping turns, which is what I did on this PCB. This also has the
advantage in RF where it produces shorter trace lengths than with
corners. However, none of this is important for this PCB. It was
designed for the IBM PC ISA bus, where the highest frequency it might
encounter would be about 14.3 MHz.

Sharp
corners are equally bad due to reflection problems. The compromise is
a chamfered corner (mitered bend):
http://www.microwaves101.com/microwave-encyclopedia/480-mitered-bends
which unfortunately also makes a tolerable fuse at the bend.

My understanding the reflection idea is also a myth but rather the real
issue is the impedance change due to the added capacitance of the
corner, which is also supported by your reference. Impedance changes
will also cause reflections, but the signal does not reflect from the
corner itself like a light beam.

Agreed. I should have said "impedance bump problem" instead of
"reflection problem".

--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

On Sat, 20 Sep 2014 11:00:09 -0700, John Larkin
wrote:


Not really, at least for RF. The real PCB would be fair accurate
clone of the hand made prototype board. The ground plane was always
on top of the PCB. Where the prototype used routed clearances for
non-grounded areas, the PCB layout used rubylith with those areas cut
out with a swivel knife compass. It was a bit tedious, but not very
difficult. The hard part was reconnecting the "islands" of ground
with Brady black tape.

I tried to find examples of such layouts using Google image search and
found nothing. I'll see if I dig out some old photos.

I still have a few mylar layouts around. I'll post pics if anyone is
interested.


You bet. Interesting history as some might say.

?-)

On Mon, 22 Sep 2014 23:56:33 -0700, josephkk
Gave us:

On Sat, 20 Sep 2014 11:00:09 -0700, John Larkin
wrote:


Not really, at least for RF. The real PCB would be fair accurate
clone of the hand made prototype board. The ground plane was always
on top of the PCB. Where the prototype used routed clearances for
non-grounded areas, the PCB layout used rubylith with those areas cut
out with a swivel knife compass. It was a bit tedious, but not very
difficult. The hard part was reconnecting the "islands" of ground
with Brady black tape.

I tried to find examples of such layouts using Google image search and
found nothing. I'll see if I dig out some old photos.

I still have a few mylar layouts around. I'll post pics if anyone is
interested.


You bet. Interesting history as some might say.

?-)



I remember 4X layout work, and the first XT and 286 PCs with AutoCAD
for doing 2 layer printer plotted 4X artwork (we were small and could
not afford UNIX workstation class CAD hardware back then).

We had our own camera in the engineering lab, and could send
production ready photo-resist masks, and gerber files to the PCB house,
saving those costs, which back then, mattered.

Then, some idiot formatted the 10MB XT drive. Oh Joy.

"DecadentLinuxUserNumeroUno" wrote in message
...
On Mon, 22 Sep 2014 23:56:33 -0700, josephkk
Gave us:

On Sat, 20 Sep 2014 11:00:09 -0700, John Larkin
wrote:


Not really, at least for RF. The real PCB would be fair accurate
clone of the hand made prototype board. The ground plane was always
on top of the PCB. Where the prototype used routed clearances for
non-grounded areas, the PCB layout used rubylith with those areas cut
out with a swivel knife compass. It was a bit tedious, but not very
difficult. The hard part was reconnecting the "islands" of ground
with Brady black tape.

I tried to find examples of such layouts using Google image search and
found nothing. I'll see if I dig out some old photos.

I still have a few mylar layouts around. I'll post pics if anyone is
interested.


You bet. Interesting history as some might say.

?-)



I remember 4X layout work, and the first XT and 286 PCs with AutoCAD
for doing 2 layer printer plotted 4X artwork (we were small and could
not afford UNIX workstation class CAD hardware back then).

We had our own camera in the engineering lab, and could send
production ready photo-resist masks, and gerber files to the PCB house,
saving those costs, which back then, mattered.

Then, some idiot formatted the 10MB XT drive. Oh Joy.

A 4.7Gb DVD-R would've mind blowing back then.

It wasn't *SO* long ago that you could set a 600Mb CD-RW formatting, and go
for lunch while you wait.