Copperless military PCBs

[Piotr Wyderski]

The "only" true advantage of a PCB-based design is its HUGE potential
for automated production and testing, with all the obvious implications
of this fact, such as a cheap final product and repeatabe parameters. Other
factors (cooling, impedance of power planes, signal integrity,
reliability) are inferior to what can be achieved using older
or more exotic techniques. For most of the time it is not worth the
effort, as the PCB is good enough, sometimes going full custom pays off.
Printed Wiring Boards, for one example, they still seem to be available
from Hitachi. A piece of
good coax connecting various points on the same PCB is another
example of a superior hybrid solution. With a planar trace you just
can't get even close to the performance level offered by a dedicated
solution. This is all clear.

But, at the other extreme, I occasionally see military-grade electronics
mounted on an FR4 substrate, but without the use of *any*
copper layers. Absolutely everything is manually wired. This is probably
the best example of this wiring obsession:

https://youtu.be/kKjchciO_wc?t=364

but these are good too:

https://youtu.be/IY6tOxWgwW4?t=34
https://youtu.be/IY6tOxWgwW4?t=97

What were those guys trying to achieve, if even a single copper layer
would have allowed them to route all the obvious stuff and apply wiring
only where truly needed? These are avionics/military designs, so I
presume they were made by real experts, not a bunch of loonies.

Best regards, Piotr

 

[John Robertson]

On 2019/09/24 2:52 p.m., Piotr Wyderski wrote:

The "only" true advantage of a PCB-based design is its HUGE potential
for automated production and testing, with all the obvious implications
of this fact, such as a cheap final product and repeatabe parameters. Other
factors (cooling, impedance of power planes, signal integrity,
reliability) are inferior to what can be achieved using older
or more exotic techniques. For most of the time it is not worth the
effort, as the PCB is good enough, sometimes going full custom pays off.
Printed Wiring Boards, for one example, they still seem to be available
from Hitachi. A piece of
good coax connecting various points on the same PCB is another
example of a superior hybrid solution. With a planar trace you just
can't get even close to the performance level offered by a dedicated
solution. This is all clear.

But, at the other extreme, I occasionally see military-grade electronics
mounted on an FR4 substrate, but without the use of *any*
copper layers. Absolutely everything is manually wired. This is probably
the best example of this wiring obsession:

https://youtu.be/kKjchciO_wc?t=364

but these are good too:

https://youtu.be/IY6tOxWgwW4?t=34
https://youtu.be/IY6tOxWgwW4?t=97

What were those guys trying to achieve, if even a single copper layer
would have allowed them to route all the obvious stuff and apply wiring
only where truly needed? These are avionics/military designs, so I
presume they were made by real experts, not a bunch of loonies.

    Best regards, Piotr

Single layer/side (SS) PCBs have a problem with cracking/cold solder
connections on heavier components (1W & up resistors for example), it
takes very little motion for SS PCB parts to fail at the solder
connections or rip the trace off the board and then break the copper.

I find that the 70s vintage commercial grade US tech that used single
sided PCBs had a number of failures due to poorly supported parts...

Those 1970s Russian units look like they could handle a fair bit of
abuse. I have no experience with US military grade 1970s tech, but
assume they would be also very robust and am guessing that it was
unlikely that the US military would use SS PCBs either.

John :-#)#

 

[bitrex]

On 9/24/19 5:52 PM, Piotr Wyderski wrote:

The "only" true advantage of a PCB-based design is its HUGE potential
for automated production and testing, with all the obvious implications
of this fact, such as a cheap final product and repeatabe parameters. Other
factors (cooling, impedance of power planes, signal integrity,
reliability) are inferior to what can be achieved using older
or more exotic techniques. For most of the time it is not worth the
effort, as the PCB is good enough, sometimes going full custom pays off.
Printed Wiring Boards, for one example, they still seem to be available
from Hitachi. A piece of
good coax connecting various points on the same PCB is another
example of a superior hybrid solution. With a planar trace you just
can't get even close to the performance level offered by a dedicated
solution. This is all clear.

But, at the other extreme, I occasionally see military-grade electronics
mounted on an FR4 substrate, but without the use of *any*
copper layers. Absolutely everything is manually wired. This is probably
the best example of this wiring obsession:

https://youtu.be/kKjchciO_wc?t=364

but these are good too:

https://youtu.be/IY6tOxWgwW4?t=34
https://youtu.be/IY6tOxWgwW4?t=97

What were those guys trying to achieve, if even a single copper layer
would have allowed them to route all the obvious stuff and apply wiring
only where truly needed? These are avionics/military designs, so I
presume they were made by real experts, not a bunch of loonies.

    Best regards, Piotr

Decadent capitalist PCB-machine expensive and slave labor was cheap

 

On Tuesday, 24 September 2019 23:52:52 UTC+1, John Robertson wrote:

On 2019/09/24 2:52 p.m., Piotr Wyderski wrote:
The "only" true advantage of a PCB-based design is its HUGE potential
for automated production and testing, with all the obvious implications
of this fact, such as a cheap final product and repeatabe parameters. Other
factors (cooling, impedance of power planes, signal integrity,
reliability) are inferior to what can be achieved using older
or more exotic techniques. For most of the time it is not worth the
effort, as the PCB is good enough, sometimes going full custom pays off..
Printed Wiring Boards, for one example, they still seem to be available
from Hitachi. A piece of
good coax connecting various points on the same PCB is another
example of a superior hybrid solution. With a planar trace you just
can't get even close to the performance level offered by a dedicated
solution. This is all clear.

But, at the other extreme, I occasionally see military-grade electronics
mounted on an FR4 substrate, but without the use of *any*
copper layers. Absolutely everything is manually wired. This is probably
the best example of this wiring obsession:

https://youtu.be/kKjchciO_wc?t=364

but these are good too:

https://youtu.be/IY6tOxWgwW4?t=34
https://youtu.be/IY6tOxWgwW4?t=97

What were those guys trying to achieve, if even a single copper layer
would have allowed them to route all the obvious stuff and apply wiring
only where truly needed? These are avionics/military designs, so I
presume they were made by real experts, not a bunch of loonies.

    Best regards, Piotr

Single layer/side (SS) PCBs have a problem with cracking/cold solder
connections on heavier components (1W & up resistors for example), it
takes very little motion for SS PCB parts to fail at the solder
connections or rip the trace off the board and then break the copper.

I find that the 70s vintage commercial grade US tech that used single
sided PCBs had a number of failures due to poorly supported parts...

Those 1970s Russian units look like they could handle a fair bit of
abuse. I have no experience with US military grade 1970s tech, but
assume they would be also very robust and am guessing that it was
unlikely that the US military would use SS PCBs either.

John :-#)#

The way that board is done the solder joints aren't likely to be broken by mechanical shock. With regular PCBs they are. The wires don't look well restrained though.


NT

 

[Piotr Wyderski]

bitrex wrote:

> Decadent capitalist PCB-machine expensive and slave labor was cheap

No, the capitalist pigs were enjoying their own wiring orgies:

https://www.youtube.com/watch?v=RpB5JqGo1co

The comrades wouldn't trust their slaves (that is, the GULAG prisoners)
enough to perform this kind of job well. These devices were designed by
experts and built at the top-notch facilities of the era by well paid
and skilled workers. Sometimes you could see a stellar level wire
lacing. They certainly wanted to achieve something, but I am not sure what.

Best regards, Piotr

 

[Jan Panteltje]

On a sunny day (Tue, 24 Sep 2019 23:52:07 +0200) it happened Piotr Wyderski
<peter.pan@neverland.mil> wrote in <qme365$mrj$1@gioia.aioe.org>:

The "only" true advantage of a PCB-based design is its HUGE potential
for automated production and testing, with all the obvious implications
of this fact, such as a cheap final product and repeatabe parameters. Other
factors (cooling, impedance of power planes, signal integrity,
reliability) are inferior to what can be achieved using older
or more exotic techniques. For most of the time it is not worth the
effort, as the PCB is good enough, sometimes going full custom pays off.
Printed Wiring Boards, for one example, they still seem to be available
from Hitachi. A piece of
good coax connecting various points on the same PCB is another
example of a superior hybrid solution. With a planar trace you just
can't get even close to the performance level offered by a dedicated
solution. This is all clear.

But, at the other extreme, I occasionally see military-grade electronics
mounted on an FR4 substrate, but without the use of *any*
copper layers. Absolutely everything is manually wired. This is probably
the best example of this wiring obsession:

https://youtu.be/kKjchciO_wc?t=364

but these are good too:

https://youtu.be/IY6tOxWgwW4?t=34
https://youtu.be/IY6tOxWgwW4?t=97

What were those guys trying to achieve, if even a single copper layer
would have allowed them to route all the obvious stuff and apply wiring
only where truly needed? These are avionics/military designs, so I
presume they were made by real experts, not a bunch of loonies.

Best regards, Piotr

Yes military stuff,
I can buy this here for 20 Euro (about 20 USD):
https://www.youtube.com/watch?v=JHk7wy-qmZQ
It is old, an antenna tuner, but shows how to do mechanics,
it will probably survive a little banging.

PCBs are vulnarable, not flexible on impact etc.
I have build complex wired stuff that still works 100% after 30 years...

What is even worse is connectors soldered on PCBs, just plug in often enough
and the soldering fails.

It is a different world, from the buy a new '?phone' every 2 years...

 

[nuny@bid.nes]

On Tuesday, September 24, 2019 at 11:12:49 PM UTC-7, Jan Panteltje wrote:

On a sunny day (Tue, 24 Sep 2019 23:52:07 +0200) it happened Piotr Wyderski
peter.pan@neverland.mil> wrote in <qme365$mrj$1@gioia.aioe.org>:

The "only" true advantage of a PCB-based design is its HUGE potential
for automated production and testing, with all the obvious implications
of this fact, such as a cheap final product and repeatabe parameters. Other
factors (cooling, impedance of power planes, signal integrity,
reliability) are inferior to what can be achieved using older
or more exotic techniques. For most of the time it is not worth the
effort, as the PCB is good enough, sometimes going full custom pays off.
Printed Wiring Boards, for one example, they still seem to be available
from Hitachi. A piece of
good coax connecting various points on the same PCB is another
example of a superior hybrid solution. With a planar trace you just
can't get even close to the performance level offered by a dedicated
solution. This is all clear.

But, at the other extreme, I occasionally see military-grade electronics
mounted on an FR4 substrate, but without the use of *any*
copper layers. Absolutely everything is manually wired. This is probably
the best example of this wiring obsession:

https://youtu.be/kKjchciO_wc?t=364

but these are good too:

https://youtu.be/IY6tOxWgwW4?t=34
https://youtu.be/IY6tOxWgwW4?t=97

What were those guys trying to achieve, if even a single copper layer
would have allowed them to route all the obvious stuff and apply wiring
only where truly needed? These are avionics/military designs, so I
presume they were made by real experts, not a bunch of loonies.

Best regards, Piotr

Yes military stuff,
I can buy this here for 20 Euro (about 20 USD):
https://www.youtube.com/watch?v=JHk7wy-qmZQ

His accent is so thick I turned on the subtitles.

A sample from 1:18-

"You have to be a nun with choker bells..."

I turned them back off.

It is old, an antenna tuner, but shows how to do mechanics,
it will probably survive a little banging.

I expect it would survive a twenty-foot drop onto concrete.

Or two.

PCBs are vulnarable, not flexible on impact etc.
I have build complex wired stuff that still works 100% after 30 years...

We called those "post boards" when I was in the USAF in the early 1970s.

What is even worse is connectors soldered on PCBs, just plug in often enough
and the soldering fails.

Don't get me started.

> It is a different world, from the buy a new '?phone' every 2 years...

The article in the video is like a lot of military gear- designed to survive worst-case "users" who had no personal investment in the equipment surviving.


Mark L. Fergerson

 

[Tim Williams]

"Piotr Wyderski" <peter.pan@neverland.mil> wrote in message
news:qmevtp$gva$1@gioia.aioe.org...

The comrades wouldn't trust their slaves (that is, the GULAG prisoners)
enough to perform this kind of job well. These devices were designed by
experts and built at the top-notch facilities of the era by well paid and
skilled workers. Sometimes you could see a stellar level wire lacing. They
certainly wanted to achieve something, but I am not sure what.

The proletariat needs work!

Tim

--
Seven Transistor Labs, LLC
Electrical Engineering Consultation and Design
Website: https://www.seventransistorlabs.com/

 

[Adrian Tuddenham]

Piotr Wyderski <peter.pan@neverland.mil> wrote:

What were those guys trying to achieve, if even a single copper layer
would have allowed them to route all the obvious stuff and apply wiring
only where truly needed? These are avionics/military designs, so I
presume they were made by real experts, not a bunch of loonies.

Copper tracks bonded to a board will crack if the board is flexed too
much, wiring is much more forgiving. By separating out the 'support'
function and the 'connection' function, the designers can optimise both.
One long-lasting brand of domestic radio used an aluminium panel with
the components soldered on both sides to copper pins through plastic
insulating rivets; that company also did Government work.

Wired circuits are a lot more repairable than printed ones, although
that might not be a factor in military use, where failure can result in
total destruction from external causes.


--
~ Adrian Tuddenham ~
(Remove the ".invalid"s and add ".co.uk" to reply)
www.poppyrecords.co.uk

 

"nuny@bid.nes" <alien8752@gmail.com> wrote in
news:a43e52b2-18b4-469a-b2f0-81a6cc0011cb@googlegroups.com:

On Tuesday, September 24, 2019 at 11:12:49 PM UTC-7, Jan Panteltje
wrote:
On a sunny day (Tue, 24 Sep 2019 23:52:07 +0200) it happened
Piotr Wyderski <peter.pan@neverland.mil> wrote in
qme365$mrj$1@gioia.aioe.org>:

The "only" true advantage of a PCB-based design is its HUGE
potential for automated production and testing, with all the
obvious implications of this fact, such as a cheap final product
and repeatabe parameters. Other factors (cooling, impedance of
power planes, signal integrity, reliability) are inferior to
what can be achieved using older or more exotic techniques. For
most of the time it is not worth the effort, as the PCB is good
enough, sometimes going full custom pays off. Printed Wiring
Boards, for one example, they still seem to be available from
Hitachi. A piece of good coax connecting various points on the
same PCB is another example of a superior hybrid solution. With
a planar trace you just can't get even close to the performance
level offered by a dedicated solution. This is all clear.

But, at the other extreme, I occasionally see military-grade
electronics mounted on an FR4 substrate, but without the use of
*any* copper layers. Absolutely everything is manually wired.
This is probably the best example of this wiring obsession:

https://youtu.be/kKjchciO_wc?t=364

but these are good too:

https://youtu.be/IY6tOxWgwW4?t=34
https://youtu.be/IY6tOxWgwW4?t=97

What were those guys trying to achieve, if even a single copper
layer would have allowed them to route all the obvious stuff and
apply wiring only where truly needed? These are
avionics/military designs, so I presume they were made by real
experts, not a bunch of loonies.

Best regards, Piotr

Yes military stuff,
I can buy this here for 20 Euro (about 20 USD):
https://www.youtube.com/watch?v=JHk7wy-qmZQ

His accent is so thick I turned on the subtitles.

A sample from 1:18-

"You have to be a nun with choker bells..."

I turned them back off.

It is old, an antenna tuner, but shows how to do mechanics,
it will probably survive a little banging.

I expect it would survive a twenty-foot drop onto concrete.

Naaah... It would have a CSA cert sticker on it then.

snip

 

[Piotr Wyderski]

Tim Williams wrote:

> The proletariat needs work!

The proletariat wanted to bring you the flame of revolution and these
were the tools.

That's obvious, but the details of the construction are far from
obvious. Winding a toroidal transformer through the PCB is another
of their quirks. Integrated planar magnetics done their way.

Best regards, Piotr

 

[Jon Elson]

On Tue, 24 Sep 2019 23:52:07 +0200, Piotr Wyderski wrote:

The "only" true advantage of a PCB-based design is its HUGE potential
for automated production and testing, with all the obvious implications
of this fact, such as a cheap final product and repeatabe parameters.
Other factors (cooling, impedance of power planes, signal integrity,
reliability) are inferior to what can be achieved using older or more
exotic techniques. For most of the time it is not worth the effort, as
the PCB is good enough, sometimes going full custom pays off. Printed
Wiring Boards, for one example, they still seem to be available from
Hitachi. A piece of good coax connecting various points on the same PCB
is another example of a superior hybrid solution. With a planar trace
you just can't get even close to the performance level offered by a
dedicated solution. This is all clear.

But, at the other extreme, I occasionally see military-grade electronics
mounted on an FR4 substrate, but without the use of *any*
copper layers. Absolutely everything is manually wired. This is probably
the best example of this wiring obsession:

https://youtu.be/kKjchciO_wc?t=364
This one looks like it could be somewhat high-voltage circuitry.

but these are good too:

https://youtu.be/IY6tOxWgwW4?t=34 https://youtu.be/IY6tOxWgwW4?t=97

In all cases, these are likely VERY low-volume production aerospace
designs, from WAY back, when any new technology was not trusted.
They might even be one-off prototypes, and after testing were reduced to
more practical construction techniques. I have certainly seen similar
construction in US-made aerospace prototypes, too. One issue was in the
"old days" it could take months to get a PC board designed, with hand-
made artwork with Rubylith and crepe tape, while such prototypes like
your videos could be whipped up in a shop in a couple days.

I have a Honeywell Alert airborne computer that was originally designed
for the X15 project (supersonic plane, 1959 - 1968) that was only flown
on the last few X15 flights. It used multilayer PC boards with all
signal traces buried in the interior, with IC's on both sides of the
board. Also, the 6 boards are connected to the motherboard by flex-print
ribbon cables. So, the boards cannot be removed from the main chassis.
But, they can be unclamped from the thermal conduction system and opened
like a book. They were apparently worried about intermittent contacts
between daughter boards and motherboard. But, then the CPU is connected
to the memory with D-style connectors. Maybe since that was done with
resilient cables, there would not be the stresses between rigid
components pressing on the connectors. Anyway, it just seemed like a
pretty strange way to build things.



Jon

 

[Tom Del Rosso]

John Robertson wrote:

Single layer/side (SS) PCBs have a problem with cracking/cold solder
connections on heavier components (1W & up resistors for example), it
takes very little motion for SS PCB parts to fail at the solder
connections or rip the trace off the board and then break the copper.

Is that because plated holes hold the pads down? Or because solder on
both sides holds the lead better, in which case it doesn't have to be
plated through?

 

[piglet]

On 26/09/2019 8:24 pm, Tom Del Rosso wrote:

John Robertson wrote:

Single layer/side (SS) PCBs have a problem with cracking/cold solder
connections on heavier components (1W & up resistors for example), it
takes very little motion for SS PCB parts to fail at the solder
connections or rip the trace off the board and then break the copper.

Is that because plated holes hold the pads down? Or because solder on
both sides holds the lead better, in which case it doesn't have to be
plated through?

The hole plating is what turns a fragile joint into a very rugged strong
joint.

piglet

 

[bitrex]

On 9/25/19 3:14 PM, Jon Elson wrote:

On Tue, 24 Sep 2019 23:52:07 +0200, Piotr Wyderski wrote:

The "only" true advantage of a PCB-based design is its HUGE potential
for automated production and testing, with all the obvious implications
of this fact, such as a cheap final product and repeatabe parameters.
Other factors (cooling, impedance of power planes, signal integrity,
reliability) are inferior to what can be achieved using older or more
exotic techniques. For most of the time it is not worth the effort, as
the PCB is good enough, sometimes going full custom pays off. Printed
Wiring Boards, for one example, they still seem to be available from
Hitachi. A piece of good coax connecting various points on the same PCB
is another example of a superior hybrid solution. With a planar trace
you just can't get even close to the performance level offered by a
dedicated solution. This is all clear.

But, at the other extreme, I occasionally see military-grade electronics
mounted on an FR4 substrate, but without the use of *any*
copper layers. Absolutely everything is manually wired. This is probably
the best example of this wiring obsession:

https://youtu.be/kKjchciO_wc?t=364
This one looks like it could be somewhat high-voltage circuitry.

but these are good too:

https://youtu.be/IY6tOxWgwW4?t=34 https://youtu.be/IY6tOxWgwW4?t=97

In all cases, these are likely VERY low-volume production aerospace
designs, from WAY back, when any new technology was not trusted.
They might even be one-off prototypes, and after testing were reduced to
more practical construction techniques. I have certainly seen similar
construction in US-made aerospace prototypes, too. One issue was in the
"old days" it could take months to get a PC board designed, with hand-
made artwork with Rubylith and crepe tape, while such prototypes like
your videos could be whipped up in a shop in a couple days.

I have a Honeywell Alert airborne computer that was originally designed
for the X15 project (supersonic plane, 1959 - 1968) that was only flown
on the last few X15 flights. It used multilayer PC boards with all
signal traces buried in the interior, with IC's on both sides of the
board. Also, the 6 boards are connected to the motherboard by flex-print
ribbon cables. So, the boards cannot be removed from the main chassis.
But, they can be unclamped from the thermal conduction system and opened
like a book. They were apparently worried about intermittent contacts
between daughter boards and motherboard. But, then the CPU is connected
to the memory with D-style connectors. Maybe since that was done with
resilient cables, there would not be the stresses between rigid
components pressing on the connectors. Anyway, it just seemed like a
pretty strange way to build things.



Jon

That piece is probably from the 1980s.

 

[Phil Hobbs]

On 9/26/19 5:44 PM, piglet wrote:

On 26/09/2019 8:24 pm, Tom Del Rosso wrote:
John Robertson wrote:

Single layer/side (SS) PCBs have a problem with cracking/cold solder
connections on heavier components (1W & up resistors for example), it
takes very little motion for SS PCB parts to fail at the solder
connections or rip the trace off the board and then break the copper.

Is that because plated holes hold the pads down?  Or because solder on
both sides holds the lead better, in which case it doesn't have to be
plated through?





The hole plating is what turns a fragile joint into a very rugged strong
joint.

piglet

Plus the fillet at the bottom of the hole. Works almost like an eyelet
without the unreliability.



Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

 

[Adrian Tuddenham]

Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 9/26/19 5:44 PM, piglet wrote:
On 26/09/2019 8:24 pm, Tom Del Rosso wrote:
John Robertson wrote:

Single layer/side (SS) PCBs have a problem with cracking/cold solder
connections on heavier components (1W & up resistors for example), it
takes very little motion for SS PCB parts to fail at the solder
connections or rip the trace off the board and then break the copper.

Is that because plated holes hold the pads down?  Or because solder on
both sides holds the lead better, in which case it doesn't have to be
plated through?





The hole plating is what turns a fragile joint into a very rugged strong
joint.

piglet


Plus the fillet at the bottom of the hole. Works almost like an eyelet
without the unreliability.

If it moves, or if the board flexes, the track will break away from it.
Wire connections will flex and remain connected - as long as they aren't
flexed so much that they work-harden.

--
~ Adrian Tuddenham ~
(Remove the ".invalid"s and add ".co.uk" to reply)
www.poppyrecords.co.uk