ic pinouts

[John Larkin]

Tube pin numbers were clockwise as seen from below, so I guess the
people who first packaged ICs followed the CCW-from-above convention.

Big mistakes were physical symmetry and putting the ground pin in a
corner [1]. TI introduced a logic series with Vccs and grounds on
middle pins, but that didn't catch on.

I believe that every possible permutation of 1-2-3 pin numbers, CW and
CCW, has been seen on various SOT-23 data sheets. And probably every
possible SOT-143/343. People can't even decide on the physical
location of the big pin, or what to number it. Some data sheets name
the pins but don't number them. Some people offer part xxx and xxxR,
with pins bent in opposite directions.

I'm doing a proto 4-layer board layout to test some oscillator
circuits, and the SAV551 footprint just didn't look right. It isn't. I
figure that there are 48 possibilities.

[1] with modern pick-and-place and inspection tools, rotated part
errors hardly ever happen any more.


--

John Larkin Highland Technology, Inc

lunatic fringe electronics

 

[bitrex]

On 5/19/19 12:12 PM, John Larkin wrote:

Tube pin numbers were clockwise as seen from below, so I guess the
people who first packaged ICs followed the CCW-from-above convention.

Big mistakes were physical symmetry and putting the ground pin in a
corner [1]. TI introduced a logic series with Vccs and grounds on
middle pins, but that didn't catch on.

Putting ground and power in the corners alone has probably cost billions
of dollars in lost productivity/routing
frustration/why-doesn't-this-shit-work-decoupling-issues over the past
50 years or so I figure, to prevent, what, destruction of the chip by
installing it backwards? a pretty uncommon and immediately obvious
error. A costly one too, maybe but only circa 1965.

I believe that every possible permutation of 1-2-3 pin numbers, CW and
CCW, has been seen on various SOT-23 data sheets. And probably every
possible SOT-143/343. People can't even decide on the physical
location of the big pin, or what to number it. Some data sheets name
the pins but don't number them. Some people offer part xxx and xxxR,
with pins bent in opposite directions.

I'm doing a proto 4-layer board layout to test some oscillator
circuits, and the SAV551 footprint just didn't look right. It isn't. I
figure that there are 48 possibilities.

[1] with modern pick-and-place and inspection tools, rotated part
errors hardly ever happen any more.

 

[olaf]

John Larkin <jjlarkin@highlandtechnology.com> wrote:

with modern pick-and-place and inspection tools, rotated part
errors hardly ever happen any more.

I have seen a pcb where the inductor was placed 90degree wrong. The
stepdown converter was working, but the efficency was only 10%. Very
funny.

Olaf

 

[John Larkin]

On Sun, 19 May 2019 18:30:42 +0200, olaf <olaf@criseis.ruhr.de> wrote:

John Larkin <jjlarkin@highlandtechnology.com> wrote:

[1] with modern pick-and-place and inspection tools, rotated part
errors hardly ever happen any more.

I have seen a pcb where the inductor was placed 90degree wrong. The
stepdown converter was working, but the efficency was only 10%. Very
funny.

Olaf

If you short the inductor, you get a linear regulator!


--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

 

[Winfield Hill]

John Larkin wrote...

I'm doing a proto 4-layer board layout ... and
the SAV551 footprint just didn't look right.

I recommend examining an actual part first.


--
Thanks,
- Win

 

[John Larkin]

On 20 May 2019 13:17:12 -0700, Winfield Hill
<hill@rowland.harvard.edu> wrote:

John Larkin wrote...

I'm doing a proto 4-layer board layout ... and
the SAV551 footprint just didn't look right.

I recommend examining an actual part first.

My layout guy just got the pin mumbers tangled. He called the big pin
1, and it's actually 2.

I have played with a real part, and the data sheet is correct. Weird
but correct.


4s 3g

1d 2s


where 2 is the big pin.


--

John Larkin Highland Technology, Inc

lunatic fringe electronics

 

[Gerhard Hoffmann]

Am 21.05.19 um 04:18 schrieb John Larkin:

On 20 May 2019 13:17:12 -0700, Winfield Hill
hill@rowland.harvard.edu> wrote:

John Larkin wrote...

I'm doing a proto 4-layer board layout ... and
the SAV551 footprint just didn't look right.

I recommend examining an actual part first.

My layout guy just got the pin mumbers tangled. He called the big pin
1, and it's actually 2.

I have played with a real part, and the data sheet is correct. Weird
but correct.


4s 3g

1d 2s


where 2 is the big pin.

Some Avago Schottky diode rings etc had a weird way of counting.
(No longer important, extinct)

I have made a test layout Thursday evening and got the boards
from PCBway in China on Monday morning. The world is shrinking, it
seems.

First thing I noted was that my decal for the LT3094 is too small.
Mea culpa.

I have made up for it with very artsy soldering.
All in all $4 for PCBway and $26 for the transport by DHL, 10 boards.
There have been worse times.

Cheers,
Gerhard

 

[boB]

On Sun, 19 May 2019 09:12:26 -0700, John Larkin
<jjlarkin@highlandtechnology.com> wrote:

Tube pin numbers were clockwise as seen from below, so I guess the
people who first packaged ICs followed the CCW-from-above convention.

Big mistakes were physical symmetry and putting the ground pin in a
corner [1]. TI introduced a logic series with Vccs and grounds on
middle pins, but that didn't catch on.

I believe that every possible permutation of 1-2-3 pin numbers, CW and
CCW, has been seen on various SOT-23 data sheets. And probably every
possible SOT-143/343. People can't even decide on the physical
location of the big pin, or what to number it. Some data sheets name
the pins but don't number them. Some people offer part xxx and xxxR,
with pins bent in opposite directions.

I'm doing a proto 4-layer board layout to test some oscillator
circuits, and the SAV551 footprint just didn't look right. It isn't. I
figure that there are 48 possibilities.

with modern pick-and-place and inspection tools, rotated part
errors hardly ever happen any more.

And when the wrong reel of parts is put on the machine, ALL of the
boards are wrong in the same way.

Haven't had that problem in a very long time now.