Why Are There No Right Angle Traces?

[Tom Woodrow]

Had to do that on a 6-layer board once.

Hand taping was so much fun. Actually I enjoyed
it most of the time. I even thought in 4x for a
long while after I laid out my last tapeup PCB.

Tom Woodrow
www.dacworks.com

 

[Frank Miles]

This an old thread but I can't resist adding my 2 cents.

1. Don't know about when they did PCB artwork with Pen and Ink, but when
we started using tape to layout PCBs (for me 1967), the tape would roll
up at 90 degree corners. Bas thing for storing taped artworks.

[snip]

Ah, but at least with tape you can bevel those nasty right-angle corners,
virtually eliminating the minor TDR glitch.

-frank
--

 

[John Larkin]

---
Also gone, and good riddance: The trip to the photo lab, making sure
whether the solder side was right-reading or not, making sure the whole
damn thing didn't get printed upside-down, emulsion side down or up,
pinholes in the negative that had to be opaqued before the positive
filmwork could be made, Aarghhh...

Now it's Gerbers emailed out and PCB's back in a few days. Heaven!

Yeah. I had to drive down to Mountain View to Lorry Ray Photography,
the only people I could trust to shoot everything right. Then I had to
call to see when it would be done, and drive down again to pick up the
film. And we were always losing film. Their camera was basicly two
rooms with a lens mounted in the wall between them... they could shoot
1:1 D-size art, accurate to a couple of mils. Out of business now,
along with those bastards at Bishop Graphics.

John

 

[John Larkin]

The red and blue tape was even less fun than the black crepy stuff. We
usually sent the whole artwork out to the PCB manufacturer and let
them worry about the photography. I don't recall an original ever
getting lost.

It was pretty fast for low density single-sided boards, especially if
the donuts were on 0.1" grid (which most of them were in those days).
Unless you ran out of the 0.060" tape or the 0.030" tape, or the 0.14"
donuts..

Best regards,
Spehro Pefhany

Nobody I knew liked the color stuff. We did 4 pieces of pin-aligned
clear mylar: padmaster, top traces, bottom traces, and silkscreen;
occasionally we did a 4-layer, with the plane art cut out of rubylith.
Lorry Ray would sandwich layers to shoot the film, and would do tricky
stuff to make ground planes and solder masks from the padmaster art...
they could expand, contract, and, or, all sorts of stuff
photographically.

We'd make assembly and fab drawings by exposing the mylars directly
onto sepia paper. By playing with layer exposure times, you could make
cool faded-layer composites, even burn the title blocks in while you
were at it.

John

 

[Fast Eddy]

I was told it was because the electronicals would fly off the edge of the
track
due to the speed they are travelling.

Rather like a car on a racetrack going really fast and coming to a corner
and sliding off.


You may have to clean out the box that the circuit is in because of the mess
it will make.

PS you will need a good eye cos you can't see them too well, but they are
there somewhere.


Edward B Tweedel
Oxford
UK

 

[John Larkin]

Are you reading Divinity? Tweedel DD has a nice ring to it.

That was Dum.

John

 

[Pat Quinn]

I was told the other day that there are no right angle traces on PCBs (the
traces go to something like 45 degrees between perpendicular traces) because
of something to do with something or other. Could something please explain
this or point me to some resources so I can book up.

Thanks

-BM

Here are my 2 cents:

1) across diagonals, you can pack traces closer
2) current density in the corner is fairly low, thus it is only
semi-functional metal
3) breakdown voltage will be lower off of sharp edges
4) a square corner will lower the impedance of a controlled-impedance
line at the corner and may lower signal integrity

-pq
Portland, Oregon

 

[Harry Conover]

I was told the other day that there are no right angle traces on PCBs (the
traces go to something like 45 degrees between perpendicular traces) because
of something to do with something or other. Could something please explain
this or point me to some resources so I can book up.

Thanks

-BM

It may very well have to do with the reason that fillets are used
whenever connecting to pads.

When I was first learning PC board layout 30 years ago, I was taught
that this reduces the problem of solder bridges when wave soldering.
This made sense to me, and so I still always use fillets when
connecting to pads.

I suspect that this is the same reason that right angle turn on PCB
laysouts should be avoided, in that they attract the formation of a
solder bridge that could very well short to other nearby conductors.

Harry C.

 

[Eric Immel]

I was told it was because the electronicals would fly off the edge of the
track
due to the speed they are travelling.

Rather like a car on a racetrack going really fast and coming to a corner
and sliding off.


You may have to clean out the box that the circuit is in because of the mess
it will make.

PS you will need a good eye cos you can't see them too well, but they are
there somewhere.


Edward B Tweedel
Oxford
UK

Don'cha hate it when electrons spill everywere?

I seem to recall, through a FeCl3 haze, something about acute angles
causing undercuts in the etching process.

EI

 

[Bob Pownall]

On 12/26/2010 4:33 PM, Phil Hobbs wrote:

AFAIK the only place where fill patterns are used in ICs is in chem-mech
polish. The issue there is that the polishing rate depends on the
composition of the surface, so you can get dishing (overpolishing) of
some areas.

Fill patterns (a.k.a. dummy patterns) are also used for some critical
etch levels, if isolated lines etch at a different rate than dense lines.

Bob Pownall

 

[[email protected]]

It's a matter of style. There's nothing wrong with right angle traces.
Most CAD software makes sorta soft corners anyway.

If there is some significant voltage on the trace relative to other
traces, a high filed strength is at the corner, which could cause a
flashover.

With high frequency signals or signals with sharp edges, the trace
should be treated as a transmission line, when the trace is longer
than about 10 % of the wavelength. In the worst case, the corner can
cause a reflection and the trace leaving at 90 degrees can act as a
stub, even if the transmission line is terminated properly. With a
smooth curvature, the trace works properly as a single transmission
line.

 

[Joerg]

It's a matter of style. There's nothing wrong with right angle traces.

Unless you are dealing with a hi-rel design. Right angles are prone to
stress tears. Ideally there should be no angles at all.

Most CAD software makes sorta soft corners anyway.

I sure wish that was so. Sometimes when I advised engineers to set the
CAD to round all corners and do teardrop entries at connectors and such
they responded that their CAD couldn't do that. At least not without
some model-making efforts.

Gimmee my Rubylith back

 

[Joerg]

I've never understood that concept. A tiny layer of dead-soft copper
is epoxied to a bunch of fiberglass. Why would there be stress at
corners? ...

One of several reasons is board flexing. Every board flexes unless you
epoxy the whole thing onto a chunk of granite. Think of it this way: Why
are bush pilots drilling holes at both ends of a metal surface tear?

... The real stress-reliability issue isn't traces, it's vias.

ftp://jjlarkin.lmi.net/BadVia2.jpg

Yep, vias, too. Which is why I'll never understand why so many layouters
leave the CAD at a very small default for via size and trace width. Not
on my boards ...

Most traces are drawn with a round Gerber aperture. That radiuses the
outside of 90 degree turns. Most traces are a little rounded on the
inside, too, on my boards. Maybe the board houses do that.

The process itself does that already, to some extent. But it is still a
very small radius. Why expose something to more stress than necessary,
especially if the remedy costs zero cents?

One thing I sure don't miss is tape on mylar. Checking a board against
the schematic could take two people two days. I fell in love once,
doing exactly that.

The first layouter I worked with in my career was a very attractive
red-haired lady, and she drove a Corvette.

Now a connectivity check takes a few seconds. Less potential for
romance, but worth it overall.

Connectivity, romance, hmm ...

 

[[email protected]]

We did fail all of the clearance checks.

IOW, you checked but didn't get clearance.

 

[[email protected]]

I can assure you that nobody was board.

No one was on board.

 

[Joerg]

I can assure you that nobody was board.

But one became like a wild boar?