Is single layer PCB really worse reliability than double lay

[Piglet]

On 11/05/2019 19:51, Rick C wrote:

On Saturday, May 11, 2019 at 1:48:52 PM UTC-4, piglet wrote:
On 11/05/2019 15:28, bitrex wrote:
On 5/11/19 3:23 AM, Piglet wrote:
On 10/05/2019 23:28, klaus.kragelund@gmail.com wrote:
Hi

So, trying to save money as always has led me down a treacherous path
of trying to do a design on 1 layer PCB

I have been told, by reasonable people, that single layer PCB will
have lower reliability due to the soldering giving up. Clinching of
pins, and gluing can mitigate this somewhat

Anyone studied 1 layer vs 2 layer reliability issues?

Cheers

Klaus

As others have noted the problem is not so much the number of layers
as the absence of a plated hole. Without the strength of the plated
hole you become reliant on the weak adhesive glueing the copper to the
substrate. So make the pads around holes as large as possible. Staking
and clinching also help. IME it is better to use right angle
connectors instead of vertical pin headers. With careful design and
build quality single sided reliability is good.

piglet


The reliability of the single-layer non-plated-thru PCBs of e.g.
synthesizer manufacturers like Roland and Oberheim products from the
1980s is not good. Well I mean it's probably OK until you actually want
to repair something else on the board. they're brittle and fragile and
the pads want to drop and lift at the slightest amount of heat from an iron

Plenty of hi-rel aerospace stuff in the 1950s and 1960s used single
sided before plated holes became mainstream.

You are right about the adhesive being the weak link. You see that
nowadays with smd rework too.

I'm not following. If the parts are mounted from one side and the copper is on the other side, how could there be any stress pushing the copper away from the substrate? The chips could only be pulled off the board, not into it.

Bitrex was complaining about the adhesive failing causing pad lift due
to excessive soldering iron heat during repair/rework. You'll have to
ask him for his horror-stories

piglet

 

[Rick C]

On Saturday, May 11, 2019 at 1:48:52 PM UTC-4, piglet wrote:

On 11/05/2019 15:28, bitrex wrote:
On 5/11/19 3:23 AM, Piglet wrote:
On 10/05/2019 23:28, klaus.kragelund@gmail.com wrote:
Hi

So, trying to save money as always has led me down a treacherous path
of trying to do a design on 1 layer PCB

I have been told, by reasonable people, that single layer PCB will
have lower reliability due to the soldering giving up. Clinching of
pins, and gluing can mitigate this somewhat

Anyone studied 1 layer vs 2 layer reliability issues?

Cheers

Klaus

As others have noted the problem is not so much the number of layers
as the absence of a plated hole. Without the strength of the plated
hole you become reliant on the weak adhesive glueing the copper to the
substrate. So make the pads around holes as large as possible. Staking
and clinching also help. IME it is better to use right angle
connectors instead of vertical pin headers. With careful design and
build quality single sided reliability is good.

piglet


The reliability of the single-layer non-plated-thru PCBs of e.g.
synthesizer manufacturers like Roland and Oberheim products from the
1980s is not good. Well I mean it's probably OK until you actually want
to repair something else on the board. they're brittle and fragile and
the pads want to drop and lift at the slightest amount of heat from an iron

Plenty of hi-rel aerospace stuff in the 1950s and 1960s used single
sided before plated holes became mainstream.

You are right about the adhesive being the weak link. You see that
nowadays with smd rework too.

I'm not following. If the parts are mounted from one side and the copper is on the other side, how could there be any stress pushing the copper away from the substrate? The chips could only be pulled off the board, not into it.

--

Rick C.

+ Get a 5,000 miles of free Supercharging
+ Tesla referral code - https://ts.la/richard11209

 

[John Robertson]

On 2019/05/11 11:51 a.m., Rick C wrote:

On Saturday, May 11, 2019 at 1:48:52 PM UTC-4, piglet wrote:
On 11/05/2019 15:28, bitrex wrote:
On 5/11/19 3:23 AM, Piglet wrote:
On 10/05/2019 23:28, klaus.kragelund@gmail.com wrote:
Hi

So, trying to save money as always has led me down a treacherous path
of trying to do a design on 1 layer PCB

I have been told, by reasonable people, that single layer PCB will
have lower reliability due to the soldering giving up. Clinching of
pins, and gluing can mitigate this somewhat

Anyone studied 1 layer vs 2 layer reliability issues?

Cheers

Klaus

As others have noted the problem is not so much the number of layers
as the absence of a plated hole. Without the strength of the plated
hole you become reliant on the weak adhesive glueing the copper to the
substrate. So make the pads around holes as large as possible. Staking
and clinching also help. IME it is better to use right angle
connectors instead of vertical pin headers. With careful design and
build quality single sided reliability is good.

piglet


The reliability of the single-layer non-plated-thru PCBs of e.g.
synthesizer manufacturers like Roland and Oberheim products from the
1980s is not good. Well I mean it's probably OK until you actually want
to repair something else on the board. they're brittle and fragile and
the pads want to drop and lift at the slightest amount of heat from an iron

Plenty of hi-rel aerospace stuff in the 1950s and 1960s used single
sided before plated holes became mainstream.

You are right about the adhesive being the weak link. You see that
nowadays with smd rework too.

I'm not following. If the parts are mounted from one side and the copper is on the other side, how could there be any stress pushing the copper away from the substrate? The chips could only be pulled off the board, not into it.

Because not all parts were completely flush with the board, allowing for
pressure points which lead to cracks. You only need a tiny amount of
float above the PCB to allow enough movement to crack the solder.

Pin connections were the bigger weak link as they could move slightly
due to the wire connections tugging slightly (thermal
expansion/contraction, gravity, vibration,...) or pressure applied to
plug/unplug the connectors would lead to cracked solder mounds.

John:-#)#

--
(Please post followups or tech inquiries to the USENET newsgroup)
John's Jukes Ltd.
MOVED to #7 - 3979 Marine Way, Burnaby, BC, Canada V5J 5E3
(604)872-5757 (Pinballs, Jukes, Video Games)
www.flippers.com
"Old pinballers never die, they just flip out."

 

[Tom Del Rosso]

John Robertson wrote:

Because not all parts were completely flush with the board, allowing
for pressure points which lead to cracks. You only need a tiny amount
of float above the PCB to allow enough movement to crack the solder.

Did rivets (eyelets) help? I saw them in some radios and things.

 

[John Robertson]

On 2019/05/11 2:40 p.m., Tom Del Rosso wrote:

John Robertson wrote:

Because not all parts were completely flush with the board, allowing
for pressure points which lead to cracks. You only need a tiny amount
of float above the PCB to allow enough movement to crack the solder.

Did rivets (eyelets) help? I saw them in some radios and things.

I've seen parts held down with various hardware and those connections
were then usually fine.

If the device is free standing and has much mass (think 2 - 5 Watt (+)
resistor or TO220 package, then you get problems over time.

John :-#)#

--
(Please post followups or tech inquiries to the USENET newsgroup)
John's Jukes Ltd.
MOVED to #7 - 3979 Marine Way, Burnaby, BC, Canada V5J 5E3
(604)872-5757 (Pinballs, Jukes, Video Games)
www.flippers.com
"Old pinballers never die, they just flip out."

 

[John Robertson]

On 2019/05/11 5:54 p.m., Tom Del Rosso wrote:

John Robertson wrote:
On 2019/05/11 2:40 p.m., Tom Del Rosso wrote:
John Robertson wrote:

Because not all parts were completely flush with the board, allowing
for pressure points which lead to cracks. You only need a tiny
amount of float above the PCB to allow enough movement to crack the
solder.

Did rivets (eyelets) help? I saw them in some radios and things.



I've seen parts held down with various hardware and those connections
were then usually fine.

If the device is free standing and has much mass (think 2 - 5 Watt (+)
resistor or TO220 package, then you get problems over time.

Oh I didn't mean rivets holding down parts. I meant in vias on
single-sided boards, as a kind of solid thru-plating. They used them
mostly where wires, rather than parts, were soldered through a hole.

Also, modern power supplies are often single sided and seem to have more
problems from bad caps.

That was somewhat rare. Old small linear power supplies sometimes used
that idea and had trouble with bad solder connections to the hollow
eyelets - these were used mostly on pass transistors on the power supply
so one could take it apart for service. We quickly learned to check for
a good connection to the eyelet and the rest of the circuit...

Also one game - Computer Space (1st coin operated video game) - had an
early double-sided motherboard where there were no plated through vias!
So they had to put a bit of wire in each hole and solder it top and
bottom. Back in the day there were a few bad solder joints and
troubleshooting that was most annoying!

John :-#)#

--
(Please post followups or tech inquiries to the USENET newsgroup)
John's Jukes Ltd.
MOVED to #7 - 3979 Marine Way, Burnaby, BC, Canada V5J 5E3
(604)872-5757 (Pinballs, Jukes, Video Games)
www.flippers.com
"Old pinballers never die, they just flip out."

 

[Tom Del Rosso]

John Robertson wrote:

On 2019/05/11 2:40 p.m., Tom Del Rosso wrote:
John Robertson wrote:

Because not all parts were completely flush with the board, allowing
for pressure points which lead to cracks. You only need a tiny
amount of float above the PCB to allow enough movement to crack the
solder.

Did rivets (eyelets) help? I saw them in some radios and things.



I've seen parts held down with various hardware and those connections
were then usually fine.

If the device is free standing and has much mass (think 2 - 5 Watt (+)
resistor or TO220 package, then you get problems over time.

Oh I didn't mean rivets holding down parts. I meant in vias on
single-sided boards, as a kind of solid thru-plating. They used them
mostly where wires, rather than parts, were soldered through a hole.

Also, modern power supplies are often single sided and seem to have more
problems from bad caps.

 

John Robertson <spam@flippers.com> wrote in
news:sP-dncfOSfd9a0vBnZ2dnUU7-L_NnZ2d@giganews.com:

Like sex and most everything else, it just ain't the same as the
real thing!

You must have decided that on pinmame.

Because MAME is absofukenlutley 100% EXACTLY the same.... the same
ROMs... all of it.

I have over 155GB of MAME data, and I can even play the old laser
games that had laser disc players in them. over 35k games.

MAME is absolutely the best processor emulator ever made for this
class of processors and this era of the realm.

Perhaps you looked at it years ago.

I play pacman digdug asteroids space zap galaga galaxian Tempest
donkey kong popeye jack the giant killer defender burger time
centipede millipede joust bosconian battlezone crazy climber frogger
pole position robotron rally x qix omega race star castle ms pacman
space invaders

All exactly original, and as a bonus, I can even pump my LED
display with it and play on a bigger screen than ever! With a
proper set up, the emulation is perfect, so exact timing on exact
code = exact same gameplay experience.

So, you must be only really talking about the pinball emulator,
which cannot provide the same experience the way computer based
games do.

 

Piglet <erichpwagner@hotmail.com> wrote in news:qb71u0$sm$1@dont-
email.me:

You are right about the adhesive being the weak link. You see that
nowadays with smd rework too.

Not if you are good at it. ;-)

The trick to smd part removal is to be good at your job, which
means you must have the aptitude to know that you cannot move the
fucking part unless ALL of the solder is molten on that part's
'legs'.

Two pin parts are easy. Instant reflow and pull off with little
heat flex of the PCB pads. Use two tip method.

After that, multi-pin jobs usually come off best with a slow
stream of hot air locally applied (do not expect to 'save' or
'reuse' pulled parts)(unless they are REALLY expensive). Always
good in those cases to ramp the entire assembly up to a couple
hundred degrees F so the jump to reflow on the part is not as great.

Shocking a nearby part with a huge temerature differential in a
short period can (and does) damage mlcc cap terminations and other
parts too.

 

"Tom Del Rosso" <fizzbintuesday@that-google-mail-domain.com> wrote
in news:qb7ffi$h8a$1@dont-email.me:

John Robertson wrote:

Because not all parts were completely flush with the board,
allowing for pressure points which lead to cracks. You only need
a tiny amount of float above the PCB to allow enough movement to
crack the solder.

Did rivets (eyelets) help? I saw them in some radios and things.

Early PTH technology had elements called "annular rings". (they
still do). These are pre and post PTH. The through hole inner wall
plating connects between two annular rings on top of and on the
bottom side of the board. Early boards had connection integrity
issues and the number of VIAs was a critical thing to keep as low as
possible as early failure modes were many times associated with a
PTH continuity failure.

I am so glad that we have come so far since then.

Nobody counts vias any more. Multi-layer PCB circuit cards are of
very high integrity these days, even on the cheap quick turn jobs.

It really is amazing how quick a 2 or 4 layer board can show up
ten parts at yer door in a few days for a few hundred bucks! Wow!

 

John Robertson <spam@flippers.com> wrote in
news:sP-dncfOSfd9a0vBnZ2dnUU7-L_NnZ2d@giganews.com:

On 2019/05/10 11:04 p.m., DecadentLinuxUserNumeroUno@decadence.org
wrote:
John Robertson <spam@flippers.com> wrote in
news:y-OdnWiPHtpUwkvBnZ2dnUU7-a3NnZ2d@giganews.com:

On 2019/05/10 6:38 p.m., George Herold wrote:
On Friday, May 10, 2019 at 6:28:50 PM UTC-4,
klaus.k...@gmail.com wrote:
Hi

So, trying to save money as always has led me down a
treacherous path of trying to do a design on 1 layer PCB

I have been told, by reasonable people, that single layer PCB
will have lower reliability due to the soldering giving up.
Clinching of pins, and gluing can mitigate this somewhat

Anyone studied 1 layer vs 2 layer reliability issues?

Cheers

Klaus

No idea.. but I could imagine that the plated through holes
are going to be worse with only one side... with bigger copper
area you are probably OK.... I've seen single layer
pcb's. :^)

George H.


In our field (arcade games/coin operated amusements) Single
Sided PCBs (SS PCBs) are a serious pain! These boards are all
old school (1970s through 1990s) with through hole parts, not
SMD so take the following with a grain of salt.

When you have parts mounted on the non-copper side and using the
leads to hold said parts in place then the part in question must
be FIRMLY mounted down to the PCB or you will get cracks
developing in the solder connections over time due to vibration
or other stresses. The old Molex 0.100 or 0.125 pin connections
on single sided boards often developed cracks over time leading
to all sorts of intermittents. So experienced techs in the game
field always touch up solder joints on SS-PCBs when they run
into them.

If your board components are all SMD and the connections are
soldered just to one side and are properly anchored I see no
reason not to go with SS-PCB, but if you are using and holes to
mount connectors or leads through then you will have trouble
down stream.

John :-#)#

Any game you can name (other than most of the clones) has
likely
been emulated.

MAME, PinMAME...

Like sex and most everything else, it just ain't the same as the
real thing!


Here are photos of all the PCB assemblies used on many of the
arcade games we played.

http://www.progettosnaps.net/PCB/


You haven't been in my shop's storage area - I have a thousand or
so vintage PCBs...

John :-#)#

No need for any of the arcade variety. A micro atx pc running
windows or linux and an LED display in an old upright cabinet will
precisely emulate any of thousands of games.

So all I have to do to be legit is own an old dead mobo, such as
the mobos you have (not saying they are dead), and I then have a
right to use that ROM code in my machine (no need for the entire old
board).

MAME emulates everything from back then from an Altair to a Heathkit
to a pocket calculator.

I DL'd gigabytes of arcade game operator's manuals. I DL'd GB of
original sales posters for most of the games. Pics of the PCBs.
Pics of the cabinets and artwork.

I have all of it. You likely could not name an old arcade game
from then (the popular ones)that I could not start up and play.

Your game you mentioned was actually what ended up being bought by
Atari and released as "Pong". I lke sorting my list by year and
looking at all the old computers that came out year after year then
and how laughably slow they are by my jaded eyes now.

Like all of the Indy and Iris and Indeo from SGI boxes. A bunch
of HP machines... It is really fun to flip through the years
looking at how bad the vga displays were. And the polymers we used
then... wow.

 

[Joseph Gwinn]

On May 11, 2019, John Robertson wrote
(in article<16KdnVwPAJiH60rBnZ2dnUU7-XXNnZ2d@giganews.com&gt:

On 2019/05/11 5:54 p.m., Tom Del Rosso wrote:
John Robertson wrote:
On 2019/05/11 2:40 p.m., Tom Del Rosso wrote:
John Robertson wrote:

Because not all parts were completely flush with the board, allowing
for pressure points which lead to cracks. You only need a tiny
amount of float above the PCB to allow enough movement to crack the
solder.

Did rivets (eyelets) help? I saw them in some radios and things.


I've seen parts held down with various hardware and those connections
were then usually fine.

If the device is free standing and has much mass (think 2 - 5 Watt (+)
resistor or TO220 package, then you get problems over time.

Oh I didn't mean rivets holding down parts. I meant in vias on
single-sided boards, as a kind of solid thru-plating. They used them
mostly where wires, rather than parts, were soldered through a hole.

Also, modern power supplies are often single sided and seem to have more
problems from bad caps.

That was somewhat rare. Old small linear power supplies sometimes used
that idea and had trouble with bad solder connections to the hollow
eyelets - these were used mostly on pass transistors on the power supply
so one could take it apart for service. We quickly learned to check for
a good connection to the eyelet and the rest of the circuit...

I had that problem with the eyelets in Keystone battery holders - a little
ooze from the battery terminal would get between rivet and solder tab
terminal, and the corrosion would make the connection unreliable, The
solution was to pre-tin the rivet and tab before installation:

Using a toothpick, put a very small drop of tinner's (liquid acid used for
tinning steel) flux at the junction between rivet and tab. Solder rivet to
tab. Wash entire battery holder in hot water to remove the acid flux residue.
When dry, install and connect, using only radio solder.

I have not tried this, but I suspect that a dab of plumbers grease flux will
also work, and may not require the hot wash. (Back in the day, before I knew
better, I assembled a Heathkit VTVM using plumbers flux. Likewise, a signal
generator. I still have them, and they still work, decades later.)

War story: In the 1970s, I knew a geology postgrad who worked as a materials
engineer at a company that was building something for the US space program.
They were trying to solder a copper cooling tube coil wrapped around a round
cadmium-plated copper assembly of some kind, and the cadmium plating was not
taking the solder. They tried all manner of fancy kinds of flux, to no avail.
When I heard this story, I was amazed, because I had been soldering test
leads to cadmium-plated steel Mueller alligator clips since high school. The
key was to polish the cadmium-plated metal with a fiberglass brush, and use
plumbers flux to pre-tin the steel. The test lead was also pre-tinned with
radio solder. The lead was soldered to tinned alligator clip with radio
solder. She took this story to work, and they succeeded that same day using
plumbers flux from the hardware store across the street.

Joe Gwinn