> Subject: [EE]Through hole plating PCB
>
> What is the best way to plate through holes on a milled PCB.
>
> Is there any suggested sites or tutorials.
>
> BOB
> --
"Best" is a subjective term. LPKF sells the gear to do it but this may not be the "best" method if you are also looking for "inexpensive".
BOB wrote:
> What is the best way to plate through holes on a milled PCB.
>
> Is there any suggested sites or tutorials.
Why is plating even necessary, for a one-off type of application? Why not
thread a piece of wire through vias, and solder it on both sides? It would
seem that component leads can be used for the same purpose: to connect the
two sides of a pad.
> From: spam_OUTpiclistTakeThisOuTmaksimov.org
> To: .....piclistKILLspam@spam@mit.edu
> Subject: Re: [EE]Through hole plating PCB
> Date: Tue, 2 Mar 2010 09:10:52 -0700
>
> BOB wrote:
> > What is the best way to plate through holes on a milled PCB.
> >
> > Is there any suggested sites or tutorials.
>
> Why is plating even necessary, for a one-off type of application? Why not
> thread a piece of wire through vias, and solder it on both sides? It would
> seem that component leads can be used for the same purpose: to connect the
> two sides of a pad.
>
> Am I missing something?
>
> Vitaliy
I posted the link to the LPKF page about through-hole plating equipment. However with the boards coming off my machine, I do the above. If I need more than two of some board I'll just send it out for fab anyway.
During layout, if I know I'm going to be making just one on the milling machine, I tune the layout to take best advantage of any through-hole parts otherwise for moving between layers. Sometimes it makes the layout look a little odd but generally works just fine.
The most difficulty I have with non-plated holes is some connectors are difficult to route without being able to run traces on both sides. YMMV.
At 07:36 AM 3/2/2010, BOB wrote:
>What is the best way to plate through holes on a milled PCB.
We use something called "track pins" from Harwin. These are a small
tapered pin that fits tightly into a #66 drill hole. They come as a
stick of perhaps 50 pins - you simply insert the leading edge of the
first pin into the hole, push to lodge it into the hole, then bend
sideways to break the pin from its neighbor on the stick. Fast and
inexpensive.
Sorry - I don't know where to purchase them now - we get ours from a
local distributor.
Dwayne Reid wrote:
> We use something called "track pins" from Harwin.
[...]
> Sorry - I don't know where to purchase them now - we get ours from a
> local distributor.
£6.69 for 500x 0.84mm pins, £13.63 for 1000x.
£5.15 for 500x 1mm pins, £13.66 for 1000x.
The insertion tool is an eye-watering £156.70.... not that you need it.
I used to use them, but found that the pin heads necessitated a massive
increase in the "via restring" parameter in EAGLE. Think I ended up
setting it to 1.2mm for a 0.8mm drill or something like that. They also
tend to stand a bit off the board a bit...
Vitaliy ha scritto:
> Why is plating even necessary, for a one-off type of application? Why not
> thread a piece of wire through vias, and solder it on both sides? It would
> seem that component leads can be used for the same purpose: to connect the
> two sides of a pad.
> Why is plating even necessary, for a one-off type of
> application? Why not thread a piece of wire through
> vias, and solder it on both sides? It would seem that
> component leads can be used for the same purpose:
> to connect the two sides of a pad.
>
> Am I missing something?
Yes, you are missing that you are not allowed to solder perpendicular
surfaces. A piece of wire could be bent to make a surfaces parallel to
PCB; component leads can't (at least on the component's side).
Marechiare wrote:
>> Why is plating even necessary, for a one-off type of
>> application?
>>
>> Am I missing something?
>
> Yes, you are missing that you are not allowed to solder perpendicular
> surfaces. A piece of wire could be bent to make a surfaces parallel to
> PCB; component leads can't (at least on the component's side).
This is the first time I hear anything like this. "Not allowed", by whom?
There are of course good reasons why hole plating is useful -- e.g., someone
mentioned having to run traces under a high density connector -- but "you
need plated holes because you're not allowed to solder perpendicular
surfaces" sounds like sheer nonsense. I mean, we *are* talking about
hobbyist type applications, not intended to meet anyone's specifications but
one's own, are we not?
Em 3/3/2010 16:28, Marechiare escreveu:
>> Why is plating even necessary, for a one-off type of
>> application? Why not thread a piece of wire through
>> vias, and solder it on both sides? It would seem that
>> component leads can be used for the same purpose:
>> to connect the two sides of a pad.
>>
>> Am I missing something?
>>
> Yes, you are missing that you are not allowed to solder perpendicular
> surfaces. A piece of wire could be bent to make a surfaces parallel to
> PCB; component leads can't (at least on the component's side).
>
>
Why not? The solder creates a meniscus that connect the pad surface to
the pin surface, otherwise it would be impossible to solder boards
without plated holes.
Isaac
__________________________________________________
Faça ligações para outros computadores com o novo Yahoo! Messenger http://br.beta.messenger.yahoo.com/
>>> Why is plating even necessary, for a one-off type of
>>> application?
>>>
>>> Am I missing something?
>>
>> Yes, you are missing that you are not allowed to solder
>> perpendicular surfaces. A piece of wire could be bent to
>> make a surfaces parallel to PCB; component leads can't
>> (at least on the component's side).
>
> This is the first time I hear anything like this. "Not allowed",
> by whom?
Are you really in EE-related business? Ask those who repaired old
valve-based TV sets.
> There are of course good reasons why hole plating is
> useful -- e.g., someone mentioned having to run traces
> under a high density connector -- but "you need plated
> holes because you're not allowed to solder perpendicular
> surfaces" sounds like sheer nonsense. I mean, we *are*
> talking about hobbyist type applications,
You said "for a one-off type of application?", this is not a synonym
to "hobbyist type applications".
> not intended to meet anyone's specifications but one's
> own, are we not?
Isaac Marino Bavaresco escreveu:
> Why not? The solder creates a meniscus that connect
> the pad surface to the pin surface, otherwise it would
> be impossible to solder boards without plated holes.
> Marechiare wrote:
>>> Why is plating even necessary, for a one-off type of
>>> application?
>>>
>>> Am I missing something?
What you're missing is that regardless of "rules", it is frequently
not possible to solder the component-side of the PCB for MANY
components. The easiest examples are IC sockets, LED displays, and
header pins/sockets. But if you start making some PCBs without PTH,
you quickly find many more :-(
William "Chops" Westfield wrote:
>>>> Why is plating even necessary, for a one-off type of
>>>> application?
>>>>
>>>> Am I missing something?
>
> What you're missing is that regardless of "rules", it is frequently
> not possible to solder the component-side of the PCB for MANY
> components. The easiest examples are IC sockets, LED displays, and
> header pins/sockets. But if you start making some PCBs without PTH,
> you quickly find many more :-(
Bill, I realized and admitted my error. It's just been a while since I had
to think about design considerations for a single sided board without PTH.
Marechiare wrote:
>>> Yes, you are missing that you are not allowed to solder
>>> perpendicular surfaces. A piece of wire could be bent to
>>> make a surfaces parallel to PCB; component leads can't
>>> (at least on the component's side).
>>
>> This is the first time I hear anything like this. "Not allowed",
>> by whom?
>
> Are you really in EE-related business?
No, I use it only as a cover for my cannabis farming business. But that's
not important right now.
> Ask those who repaired old
> valve-based TV sets.
Can anyone confirm what Marechiare is saying? I am almost certain that I
have seen cheap, brown, production single sided boards that did not have
plated holes populated with components whose leads are impossible to bend as
MC suggests. May have been a VCR?
In the SnPb days, I helped built single sided non-PTH boards with DB25
connectors. Hundreds were built, and as far as I know there were no failures
due to the DB25 losing connection.
Em 3/3/2010 23:47, Marechiare escreveu:
> Isaac Marino Bavaresco escreveu:
>
>> Why not? The solder creates a meniscus that connect
>> the pad surface to the pin surface, otherwise it would
>> be impossible to solder boards without plated holes.
>>
> Did not you notice that pins and wires were bent?
>
Perhaps I missed something in the original post. Are you saying that the
pins and wires are bent on the alluded board or in all boards in general?
I have a friend that bends the tip of all the pins and jumpers on the
bottom side of the boards, but he does this because his products are
used in motorcycles that have too much vibration.
We have lots of products in single-sided boards and we never bend the
tip of the pins on the bottom. The solder creates a cone-like (more
likely a "volcano-like") meniscus that have its base anchored to the
PCB's pad and the pin or wire protrudes straight from the tip of the
"volcano", very well soldered also.
I would say that in a vehicle this technique may indeed lead to failure
due to vibration, but for stationary equipment it is less than a concern.
Somebody said in another post that some components are impossible to
solder on the components side. That's true, but not for machined IC
sockets. Machined sockets have several mm of exposed pins on the sides.
I have made some prototypes with this technique myself, and what I did
was to add a "via" near the pin that cannot be soldered on the top side
and solder a short wire across it.
A friend of mine also used this technique already and what he did was to
solder a very thin strand of wire across every hole that needed to be
connected on both sides, and then inserted the components and soldered
them only on the bottom side. The strands of wire are so thin that they
don't take virtually any cross-sectional area of the hole.
Best regards,
Isaac
__________________________________________________
Faça ligações para outros computadores com o novo Yahoo! Messenger http://br.beta.messenger.yahoo.com/
> I am almost certain that I have seen cheap, brown,
> production single sided boards that did not have
> plated holes populated with components whose
> leads are impossible to bend as MC suggests.
Made in China, most probably.
> May have been a VCR?
More likely cheap radios, toys etc.
> In the SnPb days, I helped built single sided non-PTH
> boards with DB25 connectors. Hundreds were built,
> and as far as I know there were no failures due to the
> DB25 losing connection.
The "DB25 losing connection" failure due to single sided soldering is
very hard to reveal because it is not 100% losing connection. It may
look and measure fine even if a pin is getting loose in the SnPb cone.
> I have a friend that bends the tip of all the pins and
> jumpers on the bottom side of the boards, but he
> does this because his products are used in
> motorcycles that have too much vibration.
Not only vibration, but also high temperature. SnPb may have thermal
expansion coefficient different to that of the pin's material.
> I would say that in a vehicle this technique may
> indeed lead to failure due to vibration, but for
> stationary equipment it is less than a concern.
Em 4/3/2010 10:10, Marechiare escreveu:
>> I have a friend that bends the tip of all the pins and
>> jumpers on the bottom side of the boards, but he
>> does this because his products are used in
>> motorcycles that have too much vibration.
>>
> Not only vibration, but also high temperature. SnPb may have thermal
> expansion coefficient different to that of the pin's material.
>
>
>
>> I would say that in a vehicle this technique may
>> indeed lead to failure due to vibration, but for
>> stationary equipment it is less than a concern.
>>
> Depends on the culture in my opinion.
>
> Best regards.
>
Then I must agree to Vitaliy, you are generalizing rules that you set to
yourself. It is not a matter of "it is not allowed" or "you can't" or
"it is not possible".
As I said already, we solder the pins straight because we never had any
problems of broken solder, and also the bent pins are much harder to
unsolder when repairing the boards and in my experience they make easier
to peel-off the pads when removing them.
Regards,
Isaac
__________________________________________________
Faça ligações para outros computadores com o novo Yahoo! Messenger http://br.beta.messenger.yahoo.com/
>Marechiare wrote:
> >>> Yes, you are missing that you are not allowed to solder
> >>> perpendicular surfaces. A piece of wire could be bent to
> >>> make a surfaces parallel to PCB; component leads can't
> >>> (at least on the component's side).
> >>
> >> This is the first time I hear anything like this. "Not allowed",
> >> by whom?
> >
> > Are you really in EE-related business?
>
>No, I use it only as a cover for my cannabis farming business. But that's
>not important right now.
>
>
> > Ask those who repaired old
> > valve-based TV sets.
>
>Can anyone confirm what Marechiare is saying? I am almost certain that I
>have seen cheap, brown, production single sided boards that did not have
>plated holes populated with components whose leads are impossible to bend as
>MC suggests. May have been a VCR?
Single sided boards in consumer stuff almost never have the leads cinched.
Often they're hacked off with a carbide circular blade.
Cinching is like NASA practice or something. ;-)
Back in the early days of electronics, pre-PCBs even, it used to be felt that
solder joints should NEVER be used to support the (then rather heavy)
component, but obviously that rule of thumb is no longer followed (SMT
parts, for example, seldom have anything but the solder joints supporting
them).
> Cinching is like NASA practice or something. ;-)
Can't they just afford double sided boards? :-)
> Back in the early days of electronics, pre-PCBs even,
> it used to be felt that solder joints should NEVER be
> used to support the (then rather heavy) component,
> but obviously that rule of thumb is no longer followed
> (SMT parts, for example, seldom have anything but
> the solder joints supporting them).
A bit irrelevant considerations. We are talking NOT about supporting
weight, we are talking about soldered surfaces being parallel, not
perpendicular. SMTs do have such pads parallel to PCB's copper
surface.
A long time ago (before plated through holes) we used shoe eyelets to
connect the top and bottom traces. Note that unlike a rivet the eyelet
has a hole all the way through and so component leads can then be
soldered into it. Keystone Electronics sells hand setting tools http://www.prc68.com/I/Fasteners.shtml#E
and Stimpson has a very nice compound leverage press for installing them: http://www.prc68.com/I/Fasteners.shtml#405
they are also a great way to attach 9 volt battery snaps to a PCB. http://www.prc68.com/P/68BA.html
Isaac Marino Bavaresco wrote:
"I have a friend that bends the tip of all the pins and jumpers on the
bottom side of the boards, but he does this because his products are
used in motorcycles that have too much vibration."
We used to bend the leads on our boards simply to hold the component in
place until it is soldered in.
>> Cinching is like NASA practice or something. ;-)
>
> Can't they just afford double sided boards? :-)
So you're saying that cinching is only "required" on single sided
boards? Hmm. I've never had "training" on SS assembly, and I don't
think I've ever been involved with professional assembly of anything
with less than two sides.
However, I have in front of me two SS PCBs from consumer equipment.
One is the controller from a PC keyboard, and one is the AC-side of a
non-switching power supply for some AV equipment of some kind. The
breakdown is interesting:
Cinched: jumpers, LEDs, small ceramic caps, resistors.
Not cinched: connectors (include the AC power connector), 40-pin DIP
IC, Electrolytic cap, large 60Hz power transformer, fuse holder.
(The power transformer is sorta interesting, because it has huge
"pads", a star-shaped solder mask, and quite substantial amounts of
solder.
Theory: the machine that does automatic insertion of axial components
automatically cinches the leads, so those get done. Heavy leads that
aren't cinchable don't get done. Other components are random.
OTOH, modern consumer electronics is not exactly known for extreme
reliability. On the third hand I don't often see solder joints blamed
for failures. (and on the forth hand, it isn't often that anyone does
failure analysis. I did have a TV that was "fixed" by re-heating some
of the joints in the power supply...)
Em 4/3/2010 15:23, William "Chops" Westfield escreveu:
> OTOH, modern consumer electronics is not exactly known for extreme
> reliability. On the third hand I don't often see solder joints blamed
> for failures. (and on the forth hand, it isn't often that anyone does
> failure analysis. I did have a TV that was "fixed" by re-heating some
> of the joints in the power supply...)
>
> BillW
>
I have caught already some equipments with unmistakable solder joint
failures, mostly CRT TVs and monitors, all with large, single-side
boards and heavy parts supported only by their pads.
Isaac
__________________________________________________
Faça ligações para outros computadores com o novo Yahoo! Messenger http://br.beta.messenger.yahoo.com/
Em 4/3/2010 15:04, Vitaliy escreveu:
> Isaac Marino Bavaresco wrote:
> "I have a friend that bends the tip of all the pins and jumpers on the
> bottom side of the boards, but he does this because his products are
> used in motorcycles that have too much vibration."
>
> We used to bend the leads on our boards simply to hold the component in
> place until it is soldered in.
>
> Vitaliy
>
Our boards are kept horizontally all the time. They are soldered in a
solder bath, so we don't need to turn them belly-up before they are
soldered.
We also preform the leads of most of our components, so they don't need
to be cut afterwards. The fact that the leads are short help to avoid
them being "pushed-up" by the superficial tension of the solder (or the
hydrostatic pressure maybe, remember that Sn-Pb is very heavy, iron
floats on it).
Regards,
Isaac
__________________________________________________
Faça ligações para outros computadores com o novo Yahoo! Messenger http://br.beta.messenger.yahoo.com/
Em 4/3/2010 13:40, Marechiare escreveu:
>> Cinching is like NASA practice or something. ;-)
>>
> Can't they just afford double sided boards? :-)
>
>
>
>> Back in the early days of electronics, pre-PCBs even,
>> it used to be felt that solder joints should NEVER be
>> used to support the (then rather heavy) component,
>> but obviously that rule of thumb is no longer followed
>> (SMT parts, for example, seldom have anything but
>> the solder joints supporting them).
>>
> A bit irrelevant considerations. We are talking NOT about supporting
> weight, we are talking about soldered surfaces being parallel, not
> perpendicular. SMTs do have such pads parallel to PCB's copper
> surface.
>
I don't understand your fixation to parallelism between surfaces to be
soldered!
Isaac
__________________________________________________
Faça ligações para outros computadores com o novo Yahoo! Messenger http://br.beta.messenger.yahoo.com/
Vitaliy ha scritto:
> Marechiare wrote:
>> Made in China, most probably.
>
> Comments like this will make your posts appear baseless and arrogant (some
> people may even consider them racist).
I agree that it's not fair, but in Italy this is "common talk on the
street".
It will change, I suppose, for the best of us all, Chinese and not.
>> Made in China, most probably.
>
> Comments like this will make your posts appear
> baseless and arrogant (some people may even
> consider them racist).
Not at all. It is just public knowledge that the most of world
production of single sided PCB-based product belong to China. And thus
it is only the natural way to suggest that the PCB you were talking
about was most probably made in China. It is you who is trying to
introduce "racist" theme in the discussion. And also there is
absolutely no need to call the opponent's post "baseless and arrogant"
just out of the lack of other arguments.
Thank you,
Best Regards.
>
> Can you reference any reputable source that supports your claim?
>
> Vitaliy
>
> -
William "Chops" Westfield wrote:
>>> Cinching is like NASA practice or something. ;-)
>>
>> Can't they just afford double sided boards? :-)
>
> So you're saying that cinching is only "required"
> on single sided boards?
Not at all. Sorry, I should have quoted more original post's lines:
Spehro Pefhany wrote:
> Single sided boards in consumer stuff almost
> never have the leads cinched. Often they're
> hacked off with a carbide circular blade.
>
> Cinching is like NASA practice or something. ;-)
I just thought (probably mistakenly) that the conclusion from the
above quotation could be that cinching single sided boards is like
NASA practice, that's why I made the joke.
> I have caught already some equipments with
> unmistakable solder joint failures, mostly CRT
> TVs and monitors,
And an important point, in my opinion, is that the mentioned "solder
joint failures", most probably, are not due to the vibration, even for
"heavy parts supported only by their pads".
Em 4/3/2010 18:53, Marechiare escreveu:
>> I don't understand your fixation to parallelism
>> between surfaces to be soldered!
>>
> Anyone EE on the list? Any considerations?
>
> ---
> PS
> Olin is out of the list, so the question is likely to remain unanswered.
>
It is as almost everything in product design, a compromise between
cost/benefit.
Bending pins is time-consuming and difficults rework. Why would I do it
if I only use 1/8W resistors, ceramic capacitors, some DIP ICs,
sometimes some SOIC or TQFP, on a board that will be bolted to a wall?
As I told, we *never* had any solder joint failure for such type of
board, so why would I change my procedure?
If I were designing a single-sided board for a motorcycle CDI, as my
friend does, then I certainly would bend the pins.
Regards,
Isaac
__________________________________________________
Faça ligações para outros computadores com o novo Yahoo! Messenger http://br.beta.messenger.yahoo.com/
Marechiare wrote:
> [..] there is
> absolutely no need to call the opponent's post "baseless and arrogant"
> just out of the lack of other arguments.
Clearsee, you have it backwards. You made the claim that "you are not
allowed to solder perpendicular surfaces". I asked a reasonable question:
- Can you reference any reputable source that supports your claim?
When your argument strategy consists of questioning people's EE-ness, it's
difficult to take you seriously. If I found myself in a similar situation, I
would either present the facts that support my assertion, or admit that I
don't know what the heck I'm talking about.
:: I have a friend that bends the tip of all the pins and
:::::: jumpers on the bottom side of the boards, but he
:::::: does this because his products are used in
:::::: motorcycles that have too much vibration.
IP610 states that this is a no no, especially in class 3 medical style
devices. Components that are intended for vibration environments,
then the the components are mounted above the board with the leads
curled as in military spec.
> - Can you reference any reputable source that
> supports your claim?
Elementary, my dear Watson, - any schoolbook on Physics, I almost
sure. But, please, take it easy, "claim" is not the best word to
describe my humble statement that the highest quality you company is
well known for "does not allow" you to implement sub-standard
technologies :-)
Imagine a connector's pin is soldered to the PCB's other side single
copper layer without the pin being bent.
You know that every material in the world expands or contracts as its
temperature changes. The expansion is expressed in ppm/°C or parts per
million per °C - coefficient of thermal expansion (CTE) . This
expansion has a second property, called Young's Modulus, defined as
how hard a material pushes when it expands.
The connector's pin material may have CTE different to that of SnPb.
When temperature excursions occur, due to changes in either ambient
conditions or the power dissipated by the circuit, these materials
will expand differently, leading to the creation of stresses.
The mismatch in thermal expansion will cause shear stress between the
solder and the connector's pin. After a sufficient number of thermal
cycles, the stress will eventually lead to work hardening of the
solder, which results in cracking of the solder joint itself. The
resulting intermittent electrical conductivity is unacceptable in
today's high reliability electronics applications.
The shear stress between the solder and the connector's pin is roughly
proportional to the thickness of the solder cone and the thickness of
connector's pin. In case of perpendicular surfaces the SnPb thickness
can be a few millimetres, in case of parallel surfaces it can be a few
microns, that is hundreds times less.
Of course, bent connector's pin (if the bending permitted) may not be
flat, but still the "effective thickness" of SnPb in this case is much
less compared to that in the case of non-bent pin.
> Elementary, my dear Watson, - any schoolbook on Physics, I almost
> sure. But, please, take it easy, "claim" is not the best word to
> describe my humble statement that the highest quality you company is
> well known for "does not allow" you to implement sub-standard
> technologies :-)
>
Highest quality what a joke, I had purchased a scantool from ebay for
$ 12.00 + $ 10 shipping, linking to a similar one here
The packaging looked identical so I told him that he got scammed, he
did not believe me so we unscrewed the 4 screws holding the plastic
casing down and surprise surprise
It was identical to the one I bought off ebay - except for the front label
So now this explains why all of a sudden chinese products have taken
on a very esteemed place in Mr Vitaliy's company. When you can
"engineer" a label slap it onto a chinese counterfeit product and
start charging a 500% markup wow that is a business model I would like
to be in....
> So now this explains why all of a sudden chinese
> products have taken
Chinese may be very decent product worth the money. Anyway, your post
is absolutely not in the course of this "[EE]Through hole plating PCB"
thread, in my opinion.
"Yes, you are missing that you are not allowed to
solder perpendicular surfaces. A piece of wire
could be bent to make a surfaces parallel to PCB;
component leads can't (at least on the component's side)."
I've been giving this significant thought and
have spent some time searching for relevant information via Google.
And I just don't get it.
There has to be billions of circuit boards that
have been manufactured that consist of wire leads
pushed through holes and soldered. Many of those
are double-sided boards with plated-through-holes
but I'd bet that the majority of those boards are
single-sided with NO plated-through-holes.
Now I'm being told by somebody (one person -
Marechiare) that practice is all wrong.
I can't find any relevant information anywhere that supports this claim.
So: I have to repeat what Vitaly asked: "Can you
reference any reputable source that supports your claim?"
dwayne
PS - I suppose that some of the modern LED
lighting assemblies that I am now seeing are
manufactured all wrong as well. Those consist of
4 small PCBs assembled into a square shape, with
another PCB soldered across the end of the box
thus formed by the 4 PCBs. All connections are
solder bridges from one PCB to another.
I'd be happy to look at some relevant standards
if anyone can point me in that
direction. Otherwise, I'm just going to have to
assume that Marechiare is making stuff up for the fun of it.
> > - Can you reference any reputable source that
> > supports your claim?
>
>Elementary, my dear Watson, - any schoolbook on Physics, I almost
>sure. But, please, take it easy, "claim" is not the best word to
>describe my humble statement that the highest quality you company is
>well known for "does not allow" you to implement sub-standard
>technologies :-)
>
>Imagine a connector's pin is soldered to the PCB's other side single
>copper layer without the pin being bent.
>
>You know that every material in the world expands or contracts as its
>temperature changes. The expansion is expressed in ppm/°C or parts per
>million per °C - coefficient of thermal expansion (CTE) . This
>expansion has a second property, called Young's Modulus, defined as
>how hard a material pushes when it expands.
>
>The connector's pin material may have CTE different to that of SnPb.
>When temperature excursions occur, due to changes in either ambient
>conditions or the power dissipated by the circuit, these materials
>will expand differently, leading to the creation of stresses.
>
>The mismatch in thermal expansion will cause shear stress between the
>solder and the connector's pin. After a sufficient number of thermal
>cycles, the stress will eventually lead to work hardening of the
>solder, which results in cracking of the solder joint itself. The
>resulting intermittent electrical conductivity is unacceptable in
>today's high reliability electronics applications.
>
>The shear stress between the solder and the connector's pin is roughly
>proportional to the thickness of the solder cone and the thickness of
>connector's pin. In case of perpendicular surfaces the SnPb thickness
>can be a few millimetres, in case of parallel surfaces it can be a few
>microns, that is hundreds times less.
>
>Of course, bent connector's pin (if the bending permitted) may not be
>flat, but still the "effective thickness" of SnPb in this case is much
>less compared to that in the case of non-bent pin.
>
>Best Regards.
>
OK, let's work on the assumption that you do know what you're talking about,
your digs notwithstanding. :)
So then the same principle applies to a plated hole, when subjected to
repeated
hot/cold cycles the pin loses contact with the pad, except where there is
little solder b/w the pin and the hole wall?
- IF you want to comment on this issue (Scantool "quality" and
pricing) please change the subject to an appropriate one and post
under the [OT] tag.
- If commenting, please make the nature of your claim or comment clear.
- IF your comments relate to the superior pricing of a product that
incudes stolen IP content please post them somewhere else.
See below for reasons, if you need them.
____________
I'm trying to make sense and merits of what is being claimed here.
I'm not certain of the relative status and/or legitimacy of the two
products being compared here, whether there are IP issues, and more.
Also it's not at all obvious that 'quality' is what is being referred
to disparigingly.
Whatever the answers to these points, they don't belong under this
subject heading (a new subject was very clearly started here), don't
belong in [EE] (the new subject is nothing to do with electrical
equipment that you can make yourself), and may not belong on the list
at all (eg if IP issues are involved and the ebay advertisers are
thieves and the comparison company owns the IP).
SO:
- IF you want to comment on this issue (Scantool "quality" and
pricing) please change the subject to an appropriate one and post
under the [OT] tag.
- If commenting, please make the nature of your claim or comment clear.
- IF your comments relate to the superior pricing of a product that
incudes stolen IP content please post them somewhere else.
(FWIW - markup from FOB China to point of retail sale typically has a
3:1 to 5:1 markup ratio. Larger or small ratios may well apply for
whatever reason.
>> Elementary, my dear Watson, - any schoolbook on Physics, I almost
>> sure. But, please, take it easy, "claim" is not the best word to
>> describe my humble statement that the highest quality you company is
>> well known for "does not allow" you to implement sub-standard
>> technologies :-)
>>
> Highest quality what a joke, I had purchased a scantool from ebay for
> $ 12.00 + $ 10 shipping, linking to a similar one here
>
> http://cgi.ebay.com/ebaymotors/ELM327-OBD2-Auto-Scanner_W0QQcmdZViewItemQQhashZitem19ba67c836QQitemZ110501546038QQptZMotorsQ5fAutomotiveQ5fTools
>
> Imagine my surprise when a buddy of mine showed me an interface that
> he bought off of scantool.net
>
> http://www.scantool.net/elmscan5-compact.html
>
> The packaging looked identical so I told him that he got scammed, he
> did not believe me so we unscrewed the 4 screws holding the plastic
> casing down and surprise surprise
>
> It was identical to the one I bought off ebay - except for the front label
>
> So now this explains why all of a sudden chinese products have taken
> on a very esteemed place in Mr Vitaliy's company. When you can
> "engineer" a label slap it onto a chinese counterfeit product and
> start charging a 500% markup wow that is a business model I would like
> to be in....
$22 sounds a bit too low, the eBay tool above is $30. A lot of times a good
chunk of the price is included in the shipping price (look at the envelope
your device came in, there's only a couple of bucks worth of stamps).
> Imagine my surprise when a buddy of mine showed me an interface that
> he bought off of scantool.net
>
> http://www.scantool.net/elmscan5-compact.html
>
> The packaging looked identical so I told him that he got scammed, he
> did not believe me so we unscrewed the 4 screws holding the plastic
> casing down and surprise surprise
>
> It was identical to the one I bought off ebay - except for the front label
We manufactured ElmScan 5 for four years (2005-2009):
During that time, a number of other companies introduced similar products
based on the ELM327 IC and created a situation that is commonly referred to
as a "red ocean": several companies competing in the same market, driving
down each other's prices.
Last summer, we launched OBDLink, which has a number of important
advantages, including firmware upgrades, faster refresh rates, and an
enhanced command set:
While it instantly outsold all of our other products by a wide margin, we
felt that there was a market for a lower end scan tool for people who care
more about price than they do about upgradeability, performance, or access
to non-OBD modules. So we did some research and located a company in Hong
Kong who was able to supply good quality scan tools at a fair price. We were
flattered to find that someone in China bothered to reverse engineer our
venerable ElmScan 5 and copy elements of its circuit design. :-)
Have you ever wondered why LeCroy's low end scopes look exactly like
Rigol's? It's called "rebranding". ElmScan 5 Compact (ES5C) is custom built
for us in China, but the final testing, programming, and packaging is done
in Phoenix, Arizona:
We buy samples from eBay and other places on a regular basis, and I assure
you that your friend got a better value for his money.
- He got a software CD with an automatic installer, software, documentation,
videos, and utilities. You got a mini CD with a picture of a smiling apple,
that had an old copy of ScanTool.net v1.13 software and a folder with FTDI
drivers.
- Your friend's scan tool came with a color illustrated Quick Start Guide.
Yours came with no documentation at all. You obviously have enough technical
knowledge for it not to matter, but it is a big deal to a lot of people.
- We guarantee that ES5C will connect to any OBDII compliant vehicle (we
have a custom built automated tester that verifies every protocol). OTOH the
ELM327 clones' performance is unpredictable. Sometimes they pass our tests
with flying colors, sometimes they connect only on one protocol, and
sometimes they refuse to work at all (usually, bad crystal or caps). I have
a sample here where the OBD transceivers are missing, and can give you a
link to a forum thread where the ELM327 quit after 15 minutes and almost
caught on fire.
- In fact, I am willing to bet that as-configured, your tool will not
connect to certain vehicles that ES5C will connect to without any problems.
- We give customers 90 days to return the scan tool for a full refund, for
any reason.
- All of our scan tools including the ES5C are covered by an unmatched
3-year warranty.
- ES5C is compatible with the latest version of ScanTool.net software, and
works with OEM (=deeply discounted) versions of commercial diagnostic
software.
Of course, none of these reasons matter to folks for whom price is the
overriding factor, and that's OK. Some of these good people end up asking
for support on our forums, because the seller either can't or won't help
them. If you're willing to take a chance, and are not easily embarrassed,
more power to you. ;-)
> So now this explains why all of a sudden chinese products have taken
> on a very esteemed place in Mr Vitaliy's company.
Mr Mike, I assure you that there is nothing "sudden" about it. We moved most
of our production to China in 2006, and never looked back. Communication can
sometimes pose a challenge, but with a little ingenuity and a stringent QC
we are able to get quality that is on par with, or better, than what we got
right here in the US.
Unfortunately, the false perception resulting from
xenophobia/racism/fear/whatever that products manufactured in China are
necessarily inferior to American or European made goods, still persists. The
good news is, "Assembled in China" on the back of the iPhone and many other
great products, is helping shatter this perception.
> When you can
> "engineer" a label slap it onto a chinese counterfeit product and
> start charging a 500% markup wow that is a business model I would like
> to be in....
I *wish* our margins were that good. :)
The beauty of a free market economy is that everybody is free to choose,
nobody forces anybody to buy anything at gunpoint. Our customers (including
your friend) have 90 days after placing the order, to change their mind. I
believe that they buy our scan tools because they understand that the
difference is more than just the label.
Russell McMahon wrote:
> - IF you want to comment on this issue (Scantool "quality" and
> pricing) please change the subject to an appropriate one and post
> under the [OT] tag.
Sorry Russell, I started replying at work and finished at home, and I didn't
see your message. Honest.
Dwayne Reid <dwayner <at> planet.eon.net> writes:
> I was astonished to see written:
>
> "Yes, you are missing that you are not allowed to
> solder perpendicular surfaces. A piece of wire
> could be bent to make a surfaces parallel to PCB;
> component leads can't (at least on the component's side)."
Some work standards specify that all leads must be bent against
the board before soldering. This prevents parts from falling out
in the wave soldering system and also prevents parts from falling
out later if some failure causes overheating under vibration. The
standards are likely military or high endurance related and no-one
uses them in consumer equipment, unless absolutely necessary.
Reworking such boards is a pain.
Preformed component leads can be made to touch the PCB horizontally
before making a 90 degree turn and entering the hole in the PCB.
Then the lead can be bent again to prevent falling out, or not.
Preformed leads of this type are obligatory for parts which must
stand off from the PCB, such as power resistors and LEDs, unless
spacers are used. Lead pre-formers are cheap and available, and
should be used if needed ...
"Yes, you are missing that you are not allowed to
solder perpendicular surfaces. A piece of wire
could be bent to make a surfaces parallel to PCB;
component leads can't (at least on the component's side)."
I've been giving this significant thought and
have spent some time searching for relevant information via Google.
And I just don't get it.
There has to be billions of circuit boards that
have been manufactured that consist of wire leads
pushed through holes and soldered. Many of those
are double-sided boards with plated-through-holes
but I'd bet that the majority of those boards are
single-sided with NO plated-through-holes.
Now I'm being told by somebody (one person -
Marechiare) that practice is all wrong.
I can't find any relevant information anywhere that supports this claim.
So: I have to repeat what Vitaly asked: "Can you
reference any reputable source that supports your claim?"
dwayne
PS - I suppose that some of the modern LED
lighting assemblies that I am now seeing are
manufactured all wrong as well. Those consist of
4 small PCBs assembled into a square shape, with
another PCB soldered across the end of the box
thus formed by the 4 PCBs. All connections are
solder bridges from one PCB to another.
I'd be happy to look at some relevant standards
if anyone can point me in that
direction. Otherwise, I'm just going to have to
assume that Marechiare is making stuff up for the fun of it.
It seems to support at least part of the original argument, especially as
relates to non-plated through holes.
For many reasons, not every device is built to NASA standards, but a
standard does exist, none the less.
> OK, let's work on the assumption that you do know
> what you're talking about...
School course on Physics might help you to get rid of the need in any
assumptions about other people knowledge :-)
> So then the same principle applies to a plated hole,
> when subjected to repeated hot/cold cycles the pin
> loses contact with the pad, except where there is
> little solder b/w the pin and the hole wall?
In the case of plated hole there is no solder "bulb" around the pin.
The width/depth of solder around the pin is only fraction of
millimetre. Such thin layer would produce very little shear stress
between the solder and the connector's pin. The solder joint would
crack not in, say, 10 months, but rather in let's assume, 100 years.
>> "Yes, you are missing that you are not allowed to
>> solder perpendicular surfaces. A piece of wire
>> could be bent to make a surfaces parallel to PCB;
>> component leads can't (at least on the component's
> side)."
>
> Some work standards specify that all leads must be
> bent against the board before soldering. This prevents
> parts from falling out in the wave soldering system and
> also prevents parts from falling out later if some failure
> causes overheating under vibration. The standards are
> likely military or high endurance related and no-one
> uses them in consumer equipment, unless absolutely
> necessary.
We are talking about professional, not consumer equipment. And we are
talking about soldering through hole only to single-layered PCB.
And also your phrase "no-one uses them in consumer equipment, unless
absolutely necessary." seems to be a tautology. Unless absolutely
necessary no one would use anything in equipment.
> Reworking such boards is a pain.
Single-layered PCB normally are NOT meant to be under re-working
regardless of the bent/non-bent pins. And even if they are, you may
wish to just cut the pins off the component before de-soldering them,
- heat and pull them from the plated side of the PCB; or just heat the
pin with the iron and insert the stainless steel strip between the
bent pin and PCB copper layer, or do whatever else using right tools,
- not a problem for a good dancer :-)
>
> Preformed component leads can be made to touch the PCB horizontally
> before making a 90 degree turn and entering the hole in the PCB.
> Then the lead can be bent again to prevent falling out, or not.
> Preformed leads of this type are obligatory for parts which must
> stand off from the PCB, such as power resistors and LEDs, unless
> spacers are used. Lead pre-formers are cheap and available, and
> should be used if needed ...
>
> -- Peter
>
>
> course on Physics might help you to get rid of the need in any
> assumptions about other people knowledge
My impression is that PCB materials, component lead materials, and
solder alloys are all chosen to minimize the differential thermal
expansion effects you allude to. Not to mention the relatively low
temperature extremes that most equipment is subject to.
I could look up numbers, I supposed, but I'm not sure I have any
knowledge about how to come up with a qualitative evaluation of what
sort of differences would be "dangerous."
The NASA document referenced seems to be more worried about stress
induced on the component leads by differential expansion of the
component itself and the PCB material. It looks like they want some
length of bendable lead in the path between component and solder
connection. For instance, there is no mention of clinching being
needed for axially mounted parts, and Section 8.4.5 states explicitly
"The lead may
be terminated by clinch, straight-through, or lap configuration." Lot
of diagrams with "SR" (strain relief) detail labels pointing at a
section of lead that is not soldered.
I dunno. The tradeoffs seem pretty ambiguous. Demonstrably, such
clinching is NOT done in most consumer products. I don't see how to
tell whether that's because the tradeoffs have been carefully
analyzed, or just due to the difficulty and expense of doing it...
:: he NASA document referenced seems to be more worried about stress
:: induced on the component leads by differential expansion of the
:: component itself and the PCB material. It looks like they want
:: some
:: length of bendable lead in the path between component and solder
:: connection. For instance, there is no mention of clinching being
:: needed for axially mounted parts, and Section 8.4.5 states
:: explicitly
I again point people in the direction of IP610 Class 3 and Mil Spec
documents - nice curly wires on PTH parts were/are required on Mil
boards. Not sure what is required for SMD parts.
On Sun, Mar 7, 2010 at 6:35 PM, William "Chops" Westfield
<TakeThisOuTwestfwEraseMEspam_OUTmac.com> wrote:
>
> On Mar 6, 2010, at 2:51 PM, cdb wrote:
>
>> IP610 Class 3 and Mil Spec
>
> Got a pointer to the relevant docs? An IP610 web search looks pretty
> hopeless, and "mil spec" is worse...
It should be IPC 610 Class 3. I think the standard is not free but
probably your employer will have a copy somewhere. http://www.ipc.org/default.aspx
On Fri, Mar 5, 2010 at 5:02 AM, Marechiare <RemoveMEmarechiareTakeThisOuTgmail.com> wrote:
>>> Made in China, most probably.
>>
>> Comments like this will make your posts appear
>> baseless and arrogant (some people may even
>> consider them racist).
>
> Not at all. It is just public knowledge that the most of world
> production of single sided PCB-based product belong to China. And thus
> it is only the natural way to suggest that the PCB you were talking
> about was most probably made in China.
Yes I agree this is common knowledge. There are many PCB
factories in China, some of them are very good, some of them
are so-so, some of them are bad. I read somewhere that
China still imported high end PCBs quite a bit and actually
the export of PCB was less than import of PCB. Maybe
the situration is still the same now.
On Fri, Mar 5, 2010 at 5:13 AM, Bob Blick <bobblickEraseME.....ftml.net> wrote:
> Retagged OT
>
>> >> Made in China, most probably.
>
> When I see brown single-sided boards, especially those with punched
> holes, I think "Radio Shack" :)
Indeed. When I was a teaching assistant in USA, I bought
these single-sided boards from Radio Shack. I tried to find
better (but cheap) boards but I failed because of time limit.
Still I think they are also made in China.
The next year, the new TA designed a two-layer board
and got them made in China with quite good quality.
Dwayne Reid wrote:
> There has to be billions of circuit boards
> that have been manufactured that consist
> of wire leads pushed through holes and
> soldered. Many of those are double-sided
> boards with plated-through-holes but I'd
> bet that the majority of those boards are
> single-sided with NO plated-through-holes.
Majority expressed in what?
- just number of items;
- total square inches;
- total sold in US$;
- total income generated;
I'd like to see the numbers (and we should keep in mind we are talking
about PCBs targeting professional market).
> PS - I suppose that some of the modern
> LED lighting assemblies that I am now
> seeing are manufactured all wrong as well.
> Those consist of 4 small PCBs assembled
> into a square shape, with another PCB
> soldered across the end of the box thus
> formed by the 4 PCBs. All connections are
> solder bridges from one PCB to another.
That's a good point.
- First, your example is a device to just display something. A short
set of, say, 1 ms signal bounces won't hurt user experience. But such
bounces may not be welcome on other logic, for instance on sync lines.
- Second, that PCB copper lines are not that thick as the component
pins, thus they would cause less shear stress to the solder layer.
(But in this case you, probably, would need a PCB with good resin
layer between the PCB itself and the copper layer)
At 03:30 AM 3/12/2010, Marechiare wrote:
>Dwayne Reid wrote:
> > There has to be billions of circuit boards
> > that have been manufactured that consist
> > of wire leads pushed through holes and
> > soldered. Many of those are double-sided
> > boards with plated-through-holes but I'd
> > bet that the majority of those boards are
> > single-sided with NO plated-through-holes.
>
>Majority expressed in what?
Total number of electronic products manufactured to date (many now
residing in land-fills or entering recycle centers).
Have a look at most electronic products manufactured for consumers
say: from 1965 to perhaps a couple of years ago.
Most of those products were manufactured with single-sided boards
with through-hole components. I'm talking about: TV sets
(CRT-based), VCR's, radios, portable electronic devices such as
walkman units, cassette players, radios, etc.
I've been inside of many of those units - I used to spend a LOT of
time fixing broken electronics items for friends and family back when
I had the time and interest.
Most of those products were still working just fine when they were
discarded - they simply went obsolete.
>- just number of items;
>- total square inches;
>- total sold in US$;
>- total income generated;
>I'd like to see the numbers (and we should keep in mind we are talking
>about PCBs targeting professional market).
I'm not sure what you consider to be 'professional market' but I'd
think that consumer items manufactured by the millions is pretty
darned professional.
dwayne
PS - I'm trying hard to NOT be confrontational - its just that the
opinion you expressed is something new to me and completely at odds
with all of my previous experience and with the experience of my
peers and mentors.
I don't consider myself to be an expert but I've certainly been
involved with the design and manufacture of 'professional' electronic
products for the past 30 years or so.
I guess that's why I'm asking: cite some references (preferably
on-line) that I can actually go and examine myself.
>>Majority expressed in what?
>
> Total number of electronic products manufactured to date
> (many now residing in land-fills or entering recycle centers).
> Have a look at most electronic products manufactured for
> consumers say: from 1965 to perhaps a couple of years ago.
My statement was about certain company which sells development tools,
not consumer electronics. Besides, let's take a typical family:
mobiles, mp3 players, computer cards and peripherals (HDD, DVD,
printers, scanners, PSUs etc, etc) modern TVs, digital cameras and
even battery chargers - many dozens PCBs - about all are double or
more sided PCBs. Single sided, I can suggest from the top of my head,
are only some single chip remote controls and some very cheap PSUs.
How am I expected to come to the conclusion that single sided PCBs are
majority?
> Most of those products were manufactured with single-sided
> boards with through-hole components. I'm talking about:
> TV sets (CRT-based), VCR's, radios, portable electronic
> devices such as walkman units, cassette players, radios,
> etc. I've been inside of many of those units - I used to spend
> a LOT of time fixing broken electronics items for friends and
> family back when I had the time and interest.
Yes, I did repairing 494 chip based computer PSUs and CRT-based TV
sets and monitors among other devices. It was not quite rare that
reheating some certain solder joints on that single sided PCBs did the
job.
> I guess that's why I'm asking: cite some references
> (preferably on-line) that I can actually go and examine
> myself.
I doubt there is an extensive info on that topic. The process of
soldering of single sided PCBs is more expensive than that of double
sided PCBs if to be done properly, I believe. And it just take more
single sided PCB room to host the components. The extra cost could
well eat all the "single sided" economy. Just simple concrete maths
and folks are moving to double sided technology.
Em 12/3/2010 16:09, Marechiare escreveu:
>>> Majority expressed in what?
>>>
>> Total number of electronic products manufactured to date
>> (many now residing in land-fills or entering recycle centers).
>> Have a look at most electronic products manufactured for
>> consumers say: from 1965 to perhaps a couple of years ago.
>>
> My statement was about certain company which sells development tools,
> not consumer electronics. Besides, let's take a typical family:
> mobiles, mp3 players, computer cards and peripherals (HDD, DVD,
> printers, scanners, PSUs etc, etc) modern TVs, digital cameras and
> even battery chargers - many dozens PCBs - about all are double or
> more sided PCBs. Single sided, I can suggest from the top of my head,
> are only some single chip remote controls and some very cheap PSUs.
> How am I expected to come to the conclusion that single sided PCBs are
> majority?
>
I owned two DVD players that went dead. Both because the disk spin motor
was brushed and the brushes worn out.
Both had a large single side board with the much smaller decoder board
(this one fiberglass with four layers and a lot of fine pitch SMDs) and
the optical pick-up attached to it.
Both equipment had through-hole component with straight pins and some
SMDs also, soldered to the main board.
The main board hosted the power supply, output connectors, keypad
circuit and the VFD display circuit, among other circuits.
>> Most of those products were manufactured with single-sided
>> boards with through-hole components. I'm talking about:
>> TV sets (CRT-based), VCR's, radios, portable electronic
>> devices such as walkman units, cassette players, radios,
>> etc. I've been inside of many of those units - I used to spend
>> a LOT of time fixing broken electronics items for friends and
>> family back when I had the time and interest.
>>
> Yes, I did repairing 494 chip based computer PSUs and CRT-based TV
> sets and monitors among other devices. It was not quite rare that
> reheating some certain solder joints on that single sided PCBs did the
> job.
>
Me too, but is was due to the very poor soldering quality from the
manufacturers.
Most cracked solder joints I observed were on boards that the solder was
no more than an aluminum foil thick, not a good solder cone.
Regards,
Isaac
__________________________________________________
Faça ligações para outros computadores com o novo Yahoo! Messenger http://br.beta.messenger.yahoo.com/
On Fri, Mar 12, 2010 at 2:45 PM, Olin Lathrop <RemoveMEolin_piclistspam_OUTKILLspamembedinc.com> wrote:
> Marechiare wrote:
>> How am I expected to come to the conclusion that single sided PCBs are
>> majority?
>
> Take a look under the dashboard of your car, for example.
And similarly, go to you big-box retailer, and open up all the mass
market electronic items. Not just DVD players and TVs (ie, home
electronics), but toys, calculators, timers, clocks, watches, CFL
lightbulbs, power tools, dimmers, hair dryers, GFCI devices, pathway
lights, keyboards, mice, etc, etc, etc.
All of the items that are created millions in a manufacturing run are,
if at all possible, done on a single sided PCB because it saves
significant money over the entire batch. Yes, it's only pennies extra
per board, but that adds up when you manufacture 10 million of them.
The difference is negligible when you are only manufacturing a few thousand.
But in terms of # of products produced, AND square feet of PCB
produced, single sided certainly trumps double sided. Imagine how
many $5 - $14 electronic children's toys or CFL bulbs are produced
each day.
I just repaired a Yamaha Click Track (fancy metronome) this AM; 2 ugly
bakelite pcbs. Looks so hi-tech on the outside.......so ugly on the
inside. Same thing with a Tascam stand alone digital 8 track hard drive
audio recorder a few months back.
> On Fri, Mar 12, 2010 at 2:45 PM, Olin Lathrop <RemoveMEolin_piclistTakeThisOuTspamembedinc.com> wrote:
>
>> Marechiare wrote:
>>
>>> How am I expected to come to the conclusion that single sided PCBs are
>>> majority?
>>>
>> Take a look under the dashboard of your car, for example.
>>
>
> And similarly, go to you big-box retailer, and open up all the mass
> market electronic items. Not just DVD players and TVs (ie, home
> electronics), but toys, calculators, timers, clocks, watches, CFL
> lightbulbs, power tools, dimmers, hair dryers, GFCI devices, pathway
> lights, keyboards, mice, etc, etc, etc.
>
> All of the items that are created millions in a manufacturing run are,
> if at all possible, done on a single sided PCB because it saves
> significant money over the entire batch. Yes, it's only pennies extra
> per board, but that adds up when you manufacture 10 million of them.
>
> The difference is negligible when you are only manufacturing a few thousand.
>
> But in terms of # of products produced, AND square feet of PCB
> produced, single sided certainly trumps double sided. Imagine how
> many $5 - $14 electronic children's toys or CFL bulbs are produced
> each day.
>
> -Adam
>
>
>The difference is negligible when you are only manufacturing a few thousand.
Actually, in smaller volumes (low hundreds to low thousands), going
to double-sided / plated-through-holes from single-sided adds about
30% to the cost. That is based upon both quotes as well as actual
PCB orders from both Canadian and Chinese PCB suppliers.
Most of our high-volume stuff (4K to 10K pieces per year) is
single-sided just because of that cost savings. I'll spend an extra
day or two to get the layout to fit into a single layer with no
jumpers (if possible) just because we and our customer recoup those
savings year after year after year.
>> Yes, I did repairing 494 chip based computer PSUs and CRT-based TV
>> sets and monitors among other devices. It was not quite rare that
>> reheating some certain solder joints on that single sided PCBs did the
>> job.
>>
>
>
> Me too, but is was due to the very poor soldering quality from the
> manufacturers.
> Most cracked solder joints I observed were on boards that the solder was
> no more than an aluminum foil thick, not a good solder cone.
Never trust a pretty cone (rephrasing Amanda Lear :-) Sometimes when I
de-soldered one leg of, say, 0.5W resistor, the other leg would just
drop out from what looked like a good cone. Preventive reheating of
almost all solder joints with flux after each 5 years (5000 work
hours) seems to be a good idea for a single sided PCB with more or
less massive and hot components (CRT monitors for example)
> All of the items that are created millions in a manufacturing
> run are, if at all possible, done on a single sided PCB
> because it saves significant money over the entire batch.
That's somewhat debatable point, for example modern low-budget HDDs
are created in tenth of millions and have a very small double (I
believe) sided PCB. I think they could span the components to bigger
one-sided PCB, but for some reason they don't.
M. Adam Davis wrote:
> But in terms of # of products produced, AND square feet
> of PCB produced, single sided certainly trumps double
> sided. Imagine how many $5 - $14 electronic children's
> toys or CFL bulbs are produced each day.
Does not work for me, for I got no $5 - $14 electronic children's
toys, currently no CFL bulbs, no pathway lights. My big PC
motherboards, big video-cards, peripheral and other double sided
boards definitely beat in square feet all my single sided boards.
On Fri, Mar 12, 2010 at 6:25 PM, Marechiare <RemoveMEmarechiareKILLspamgmail.com> wrote:
>> All of the items that are created millions in a manufacturing
>> run are, if at all possible, done on a single sided PCB
>> because it saves significant money over the entire batch.
>
> That's somewhat debatable point, for example modern low-budget HDDs
> are created in tenth of millions and have a very small double (I
> believe) sided PCB. I think they could span the components to bigger
> one-sided PCB, but for some reason they don't.
>
> Regards
They are almost certainly more than two layer given the fine pitch
controller and many signal traces. It's not possible to put a PQ208,
for example, on a crappy single sided phenolic board. An old DVD
player I just took apart (4 years old) has a large single sided board.
Like what one other person just mentioned, it also has a very small
multi-layer board with codec and controller chip on it (more or less.)
The last circuit board I designed (6 sq. inches) was 4 layers when 2
would have worked, because trying to route it for 2 layers was taking
too long. Since my time costs more money than the savings would have
been, it was pointless to spend more time on it. I could fit it on a
single layer board if I had about 36 inches^2, but I didn't. It's not
really debatable that engineers will try to save $0.10 if they can.
I'm reminded of the airline who removed a couple peanuts from their
snack bags and saved a couple million a year. If that airline only had
10 passengers a day then the peanuts don't matter.
At 04:25 PM 3/12/2010, Marechiare wrote:
> > All of the items that are created millions in a manufacturing
> > run are, if at all possible, done on a single sided PCB
> > because it saves significant money over the entire batch.
>
>That's somewhat debatable point, for example modern low-budget HDDs
>are created in tenth of millions and have a very small double (I
>believe) sided PCB.
Yep - horses for courses.
What I mean by that is: as components become smaller and more dense,
single-sided boards just aren't suitable and you see double-sided or
4-layer or 6-layer boards more and more often.
But you have to realize that is a RECENT trend - starting with, I
think, modern PC computers and their peripherals. The trend away
from single-sided boards will continue just because modern technology
demands more density.
But that trend is NOT caused by anything related to the statement
that you made: "Yes, you are missing that you are not allowed to
solder perpendicular surfaces."
It is that statement that I take exception to and it is that claim
that I'd like to you provide links to standards that specifically
make that statement.
I'm a geeky kind of a guy and I tend to disassemble anything new that
I purchase or am given - just to see what's inside. Mass-market
consumer electronics items still use a LOT of single-sided PC
boards. And, as Olin pointed out, the last automotive dashboard that
I disassembled to repair was also a large single-sided PC board.
Modern design engineers will use the least expensive technology
possible to manufacture their goods. They will use double-sided or
multi-layer boards if the design requires it. But they still use
single-sided boards with through-hole component leads stuck straight
through the board, then wave-soldered when appropriate. Those
through-hole components are all mixed up with SMT devices these days,
of course. And, often, lots and lots of wire jumpers just so that
they DON'T have to go to a double-sided board.
Dwayne Reid wrote:
> But you have to realize that is a RECENT trend - starting
> with, I think, modern PC computers and their peripherals.
> The trend away from single-sided boards will continue just
> because modern technology demands more density.
>
> But that trend is NOT caused by anything related to the
> statement that you made: "Yes, you are missing that you
> are not allowed to solder perpendicular surfaces."
>
> It is that statement that I take exception to and it is that
> claim that I'd like to you provide links to standards that
> specifically make that statement.
I agree with you about the trend. But my statement was about
particular company's particular product. Basically it means that for
the particular dev tools product in small quantities most probably it
would be better off using double sided boards if we target reliability
too.
And if for some reason one sided PCBs are to be used, then it would
make sense spending a few extra cents per device to try making Chinese
folks bend the leads if permitted. Bent leads solder joint are more
reliable from my experience as I said "Don't trust a pretty cone" (a
joke). No need to 100% comply with NASA standards, but why not doing
it to some reasonable extent?
> The process of soldering of single sided PCBs is more expensive
> than that of double sided PCBs if to be done properly, I believe.
I don't know how you come to that conclusion. My observations of mass
soldering of single sided PCBs is that it is quite probably considerably
cheaper than doing double sided. A solder bath machine would be easier to
run than any form of reflow solder machine that needs to do both sides of
the board, and would produce as good a job as any reflow solder machine,
from the results I have seen.
> All of the items that are created millions in a manufacturing run are,
> if at all possible, done on a single sided PCB because it saves
> significant money over the entire batch.
Not only that, but where possible the PCB will be phenolic based rather than
fiberglass, as the holes can be punched rather than drilled, which is much
quicker, along with the cheaper bare material cost.
maksimov.org
