Searching \ for '[EE]:: SMD switch mounting reliability - comments' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/index.htm?key=smd+switch+mounting
Search entire site for: ': SMD switch mounting reliability - comments'.

Exact match. Not showing close matches.
PICList Thread
'[EE]:: SMD switch mounting reliability - comments '
2008\11\28@012203 by apptech

face
flavicon
face
An existing push button N/O push to operate switch uses through hole
techology. Use rate is reasonably high - target is 100,000 reliable
operations in 5 years. It's behind a rubber boot and forces on it are liable
to be "reasonably vertical" most of the time. A user would have to try
reasonably hard to exert any noticeable sideforce on the switch. It's in
portable equipment which is liable to be in regular use every dau in many
cases and also liable to be subject to a reasonable amount of rough and
tumble The switch is not liable to usually be subject to extreme forces and
it's not a high vibration environment. [[No prizes for guestimating the
appliocation. Nor comment probably]].

SO

The manufacturer wants to change to a surface mount switch.
I'm a bit dubious about the probable longevity of the mounting if solder
alone is used. Adhesive could be added as a backup.

Any thoughts?

Ant experiences to relate?

Vaguely related data points:

- Solder is subject to creep tatigue.

- Make a key ring from an old SIM RAM card with SMD RAM chips. Notice a few
weeks/months on that all the ICs have invariably fallen off.



        Russell



2008\11\28@054326 by Gerhard

picon face
On Fri, 28 Nov 2008 19:21:41 +1300, apptech wrote:

{Quote hidden}

The older keyring remote controls in use here (car alarm, for example)
usually have SMD switches that are only soldered, and they seem to be
reasonably reliable over many years. There is also the method they use
today, with printed patters on the board and a conductive layer on the
inside of the rubber.

Gerhard

2008\11\28@054714 by Tony Smith

flavicon
face
> The manufacturer wants to change to a surface mount switch.
> I'm a bit dubious about the probable longevity of the mounting if solder
> alone is used. Adhesive could be added as a backup.
>
> Any thoughts?
>
> Ant experiences to relate?
>
> Vaguely related data points:
>
> - Solder is subject to creep tatigue.
>
> - Make a key ring from an old SIM RAM card with SMD RAM chips. Notice a
few
> weeks/months on that all the ICs have invariably fallen off.


On the RAM key rings I've made I ran a thin bead of superglue along the
edges of the chips (the ones without the pins) and I think they're all still
attached.  A good thump would probably dislodge them.  Epoxy would be
better, but is harder to handle.  You could glue the switch on first, so the
glue is under and not just around it.  Remember that superglue will fail
when heated, so the switches would need to be hand soldered.  Well, the
gluing would be manual too, so no big deal.  Check woodworking stores for
superglue, they sell high viscosity versions in largish quantities.  Often
used for coating pens, stabilising wood (eg knot holes) and so on.

The rubber boot could have a plastic insert, so any side load would mean it
would slide across the switch, and not transfer the movement.  Similar to
how lot of stuff (eg VCRs with the buttons in the faceplate) works.

Drown it in hot glue?

Design the housing so it fits around the switch?

Say 'sod it' and use a reed switch & magnet?

Tony

2008\11\28@061420 by apptech

face
flavicon
face
"Tony Smith"

said various good things ...

but

> Drown it in hot glue?

has problems. Hot melt glue is an easy and rapid way of retaining components
mechanically BUT the bond fails with age. Unless the glue is mechanically
keyed into holes or slots so it cannot mechanically be removed it in time
will form a smooth and non stick surface and just let go. Anything tyhat
depends mechanically on hot melt glue for support will become detached in
months to years.

The same applies to the hard ceramic-appearing glue beloved of Asian
manufacturers for "retaining" large capacitors and similar. It suffers the
same fate as hot-melt glue. I've never seen it applied so don't know how it
starts life.

One glue that does work very long term (but not a perfect solution in all
cases) is "silicon rubber". I use 'neutral cure' SR for various things and
it seems to kep on keeping on very well. Plus it's good for 20+ years in
outdoor conditions.


   Russell


2008\11\28@065248 by Tony Smith

flavicon
face
> >The manufacturer wants to change to a surface mount switch.
> >I'm a bit dubious about the probable longevity of the mounting if solder
> >alone is used. Adhesive could be added as a backup.
> >
> >Any thoughts?
> >
> >Ant experiences to relate?
>
> The older keyring remote controls in use here (car alarm, for example)
> usually have SMD switches that are only soldered, and they seem to be
> reasonably reliable over many years. There is also the method they use
> today, with printed patters on the board and a conductive layer on the
> inside of the rubber.
>
> Gerhard


Some of the rubbers (silicone) leach out some sort of oil that causes
problems.

Tony

2008\11\28@065721 by Tony Smith

flavicon
face
> said various good things ...
>
> but
>
> > Drown it in hot glue?
>
> has problems. Hot melt glue is an easy and rapid way of retaining
components
> mechanically BUT the bond fails with age. Unless the glue is mechanically
> keyed into holes or slots so it cannot mechanically be removed it in time
> will form a smooth and non stick surface and just let go. Anything tyhat
> depends mechanically on hot melt glue for support will become detached in
> months to years.
>
> The same applies to the hard ceramic-appearing glue beloved of Asian
> manufacturers for "retaining" large capacitors and similar. It suffers the
> same fate as hot-melt glue. I've never seen it applied so don't know how
it
> starts life.


Yeah, I'm not a big fan of hot glue, but if, as you say, you arrange things
so it's keyed it works well.  Thinking about it, it rarely works unless you
give it something to lock to.

You've reminded me I need to find something similar for another project.

Tony

2008\11\28@085128 by Gordon Williams

picon face
If you want to go the supper glue route then I would go for the low
viscosity type.  Go to your local hobby store and ask for THIN CA glue.
This stuff excells at being wicked into the smallest of spaces and works
best if only the thinest of films is used.

My suggestion is that you solder the componet down and then place ONE drop
of glue at the edge.  You should see it get wicked under the part.

CA (cyanoacrylate) glue is moisture activated (just the moisture in the air
is sufficient).  It does have a shelf life so try and get some that is
fresh.  If the bottle has dust on it in your local store, I would look
elsewhere.

Regards,

Gordon Williams


> On the RAM key rings I've made I ran a thin bead of superglue along the
> edges of the chips (the ones without the pins) and I think they're all
still
> attached.  A good thump would probably dislodge them.  Epoxy would be
> better, but is harder to handle.  You could glue the switch on first, so
the
> glue is under and not just around it.

2008\11\28@085631 by olin piclist

face picon face
apptech wrote:
> An existing push button N/O push to operate switch uses through hole
> techology. Use rate is reasonably high - target is 100,000 reliable
> operations in 5 years. It's behind a rubber boot and forces on it are
> liable to be "reasonably vertical" most of the time.
>
> ...
>
> The manufacturer wants to change to a surface mount switch.
> I'm a bit dubious about the probable longevity of the mounting if
> solder alone is used. Adhesive could be added as a backup.

I don't like surface mount for anything that will take external stress.  I
once used a SMD 1.3mm wall wart jack, and there were problems even though
the jack had a extra plastic protrusion that went thru a dedicated hole in
the PC board for that purpose.  However, this was a case where users could
easily excert sideways and even lifting force.  In some cases the trace came
off the PC board before the pin came off the trace.

In your case where the direction of the force is well controlled, it might
almost be appropriate to use surface mount.

I'm doing a similar thing on a product right now, but that's driven by space
constraints.  It's a totally crammed tiny board with parts on both sides.
It has to be small and light as it will be worn by the end user.  SMD
pushbuttons are smaller and disrupt less board area then thru hole.  In this
case the user pushes a small rod which pushes the switch, so forces are
vertical down only.  This is only in prototype stage so I have no lifetime
data over many units yet.  While I'd rather have thru hole with all being
equal, thing are far from equal and I'm reasonably OK with this one being
SMD.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000.

2008\11\28@161559 by apptech

face
flavicon
face
>> The older keyring remote controls in use here (car alarm, for example)
>> usually have SMD switches that are only soldered, and they seem to be
>> reasonably reliable over many years.

Useful data point - although their usage rates are liable to be much lower
than I specified.


       R

2008\11\28@225004 by Tony Smith

flavicon
face
> >> The older keyring remote controls in use here (car alarm, for example)
> >> usually have SMD switches that are only soldered, and they seem to be
> >> reasonably reliable over many years.
>
> Useful data point - although their usage rates are liable to be much lower
> than I specified.
>
>
>         R


I can't recall the last time I've had a SMD switch solder joint fail, I do
know the last one was where the circuit board flexed a little, and that
caused the track to crack.  Sockets, such as headphones, are more likely to
be ripped off the PCB in one way or another.

A test rig isn't too hard to set up, and motor/cam arrangement spinning at
60rpm will do 86400 presses in a day.  Those little microwave turntable
motors are great for this stuff, but they're usually only 5-6 RPM.  The cam
is nothing special, a round wooden disc with the hole offset from the centre
will work.  In this case the hole only needs to be offset a few mm as the
travel distance is short.

Side load can be simulated by adding a 'finger', i.e. a small piece of rod,
between the cam & the switch and setting it at an angle.  A small bit of
pipe around the finger to guide it would be useful.

Tony

2008\11\30@052437 by Peter

picon face
> The manufacturer wants to change to a surface mount switch.
> I'm a bit dubious about the probable longevity of the mounting if solder
> alone is used. Adhesive could be added as a backup.

Totally isolate and buffer the external force from the button using an elastic
lever (fulcrum) with the button and the booth acting on different parts of the
fulcrum.

Peter



'[EE]:: SMD switch mounting reliability - comments '
2008\12\01@054448 by apptech
face
flavicon
face
>> The manufacturer wants to change to a surface mount switch.
>> I'm a bit dubious about the probable longevity of the mounting if solder
>> alone is used. Adhesive could be added as a backup.

> Totally isolate and buffer the external force from the button using an
> elastic
> lever (fulcrum) with the button and the booth acting on different parts of
> the
> fulcrum.

Probably  agood idea, but not an easy option here. I largely have power of
veto or agreement. "Fancy" alternatives are fairly largely not an option.

If/when this gets re-engineered it will quite possibly get a magnet and Hall
switch solution. Even an optical interuption system could be 'interesting'
but the Hall sensor does give you full environmental isolation. Cost is the
standing current (smallish with some duty cycle multiplexed versions) and
cost of the sensor IC - which tends to be dearer for the multiplexed
versions. Reed switches have no standby power needs but reliability is
generally not stunning. I have seen some cheap consumer level Chinese
manufactured products with magnet and reed-switch switches.


         Russell


2008\12\01@080850 by Tony Smith

flavicon
face
> >> The manufacturer wants to change to a surface mount switch.
> >> I'm a bit dubious about the probable longevity of the mounting if
solder
> >> alone is used. Adhesive could be added as a backup.
>
> > Totally isolate and buffer the external force from the button using an
> > elastic
> > lever (fulcrum) with the button and the booth acting on different parts
of
> > the
> > fulcrum.
>
> Probably  agood idea, but not an easy option here. I largely have power of
> veto or agreement. "Fancy" alternatives are fairly largely not an option.
>
> If/when this gets re-engineered it will quite possibly get a magnet and
Hall
> switch solution. Even an optical interuption system could be 'interesting'
> but the Hall sensor does give you full environmental isolation. Cost is
the
> standing current (smallish with some duty cycle multiplexed versions) and
> cost of the sensor IC - which tends to be dearer for the multiplexed
> versions. Reed switches have no standby power needs but reliability is
> generally not stunning. I have seen some cheap consumer level Chinese
> manufactured products with magnet and reed-switch switches.


The 'cheap but perfectly adequate' LED bike lights use a magnet & reed
switch arrangement now.  Random eBay auction <
http://cgi.ebay.com.au/ws/eBayISAPI.dll?ViewItem&item=120341714872> or
<http://cgi.ebay.com.au/ws/eBayISAPI.dll?ViewItem&item=110320225950> if you
don't like black.

The previous design used a pushbutton.  The reed switch itself is about
12x2mm, and isn't treated in any special way.  It sits reasonably securely
in a groove, with the leads running up to the PCB.

Personally I prefer the pushbutton, and assume it would be cheaper.  Perhaps
not.  

I've got both versions, and both still work after a couple of years on bikes
and bouncing around in luggage.  Oddly enough, while the two versions appear
identical, the various parts aren't interchangeable (even the handlebar
mount), and the newer ones have a narrower beam.

Tony

2008\12\01@091602 by Peter

picon face
apptech <apptech <at> paradise.net.nz> writes:
> > Totally isolate and buffer the external force from the button using an
> Probably  agood idea, but not an easy option here. I largely have power of
> veto or agreement. "Fancy" alternatives are fairly largely not an option.

There is nothing fancy about a simple straight plastic lever or mild steel leaf
that can be mounted with the boot and pokes one end through a hole in the pcb or
wraps around its edge at the other end. The way some products are made now, the
only thing that matters seems to be that they get out of the factory door in one
piece.

Cheezy ascii drawing of cutaway view follows:

Boot  =-T-=
Leaf   ----------    
Button--> [ ^ ]  \__ <-- leaf end soldered on board
Board ================

Peter


More... (looser matching)
- Last day of these posts
- In 2008 , 2009 only
- Today
- New search...