Soldering SMDs

[Paul Burridge]

I can't avoid 'em any longer. I've got no choice but to solder this
itty-bitty spec of sh*t to a PCB and would appreciate a steer on the
best way. Three pins per side with 1mm between pin centres. I've got
the finest soldering iron tip I can find and a lab-grade stereoscope
so I can see what I'm doing. Now all I need is to *know* what I'm
doing. Any constructive advice gratefully received...
--

"I believe history will be kind to me, since I intend
to write it." - Winston Churchill

 

[A E]

Paul Burridge wrote:

I can't avoid 'em any longer. I've got no choice but to solder this
itty-bitty spec of sh*t to a PCB and would appreciate a steer on the
best way. Three pins per side with 1mm between pin centres. I've got
the finest soldering iron tip I can find and a lab-grade stereoscope
so I can see what I'm doing. Now all I need is to *know* what I'm
doing. Any constructive advice gratefully received...
--

"I believe history will be kind to me, since I intend
to write it." - Winston Churchill

What I end up doing sometimes is to just blob the solder on, and then remove the
excess with really fine braid. Of course, continuity testing afterwards is
mandatory. I've managed to solder 0.5mm Samtec connectors like this.

 

[Andrew Paule]

What you need is FLUX - I solder these kind of things with a moderate
sized iron, regular (0.028) solder, and flux.

Get a corner pin located, tack solder, locate other corner, tack solder
(just put a little solder on the tip of you iron) -flux the pins on a
side you didn't tack, and then drag the tip and solder along, repeat,
and clean with flux remover -I learned this from a lady who did
commercial rework on SMD's for a lving - I've put 0.5mm parts on with
not problem pretty quick. Now a 201 resitor or cap is another story -
I'm a tombstoner. Anyone know how to do these things well?

Andrew

Paul Burridge wrote:

I can't avoid 'em any longer. I've got no choice but to solder this
itty-bitty spec of sh*t to a PCB and would appreciate a steer on the
best way. Three pins per side with 1mm between pin centres. I've got
the finest soldering iron tip I can find and a lab-grade stereoscope
so I can see what I'm doing. Now all I need is to *know* what I'm
doing. Any constructive advice gratefully received...
--

"I believe history will be kind to me, since I intend
to write it." - Winston Churchill

 

[Spehro Pefhany]

On Thu, 07 Aug 2003 11:29:48 +0100, the renowned Paul Burridge
<redscar@waitrose.notformail.com> wrote:

I can't avoid 'em any longer. I've got no choice but to solder this
itty-bitty spec of sh*t to a PCB and would appreciate a steer on the
best way. Three pins per side with 1mm between pin centres.

1mm? That's HUGE. And only 3 pins/side? No sweat.

I've got
the finest soldering iron tip I can find and a lab-grade stereoscope
so I can see what I'm doing. Now all I need is to *know* what I'm
doing. Any constructive advice gratefully received...

Get good fresh rosin core solder (eg. Kester 44 63/37). Doesn't have
to be especially fine (eg. 0.031"). Place the part in position (that's
the fumbly bit). Tack one side down with a bit of solder on the iron.
Blob *fresh* solder on the other side, shorting a couple of pins
together at a time. If you do it right the blob should travel from one
end to the other as you move the iron with the excess ending up on the
iron. Shake (sponge) it off and repeat on the other side. If you get
gummy shorts, fix it by adding fresh solder.

The part should self-center on the pads while you're at it.

Best regards,
Spehro Pefhany
--
"it's the network..." "The Journey is the reward"
speff@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com

 

[John Larkin]

On Thu, 07 Aug 2003 14:36:45 GMT, Spehro Pefhany <speff@interlog.com>
wrote:

On Thu, 07 Aug 2003 11:29:48 +0100, the renowned Paul Burridge
redscar@waitrose.notformail.com> wrote:


I can't avoid 'em any longer. I've got no choice but to solder this
itty-bitty spec of sh*t to a PCB and would appreciate a steer on the
best way. Three pins per side with 1mm between pin centres.

1mm? That's HUGE. And only 3 pins/side? No sweat.

Yeah. Practise on that, then move on to the small stuff.

Another technique: get some liquid RMA flux. Position chip on pads and
put some flux on/under each pin with a toothpick. Now put a small blob
of solder on the tip of your iron and just touch each pin. The solder
will slurp off the tip onto each pad sort of automagically.

John

 

[A E]

Spehro Pefhany wrote:

On Thu, 07 Aug 2003 11:29:48 +0100, the renowned Paul Burridge
redscar@waitrose.notformail.com> wrote:


I can't avoid 'em any longer. I've got no choice but to solder this
itty-bitty spec of sh*t to a PCB and would appreciate a steer on the
best way. Three pins per side with 1mm between pin centres.

1mm? That's HUGE. And only 3 pins/side? No sweat.

I agree... I wish stuff had 1mm pitch these days. Did you see the SOT-553
package?

I've got
the finest soldering iron tip I can find and a lab-grade stereoscope
so I can see what I'm doing. Now all I need is to *know* what I'm
doing. Any constructive advice gratefully received...

Get good fresh rosin core solder (eg. Kester 44 63/37). Doesn't have
to be especially fine (eg. 0.031"). Place the part in position (that's
the fumbly bit). Tack one side down with a bit of solder on the iron.
Blob *fresh* solder on the other side, shorting a couple of pins
together at a time. If you do it right the blob should travel from one
end to the other as you move the iron with the excess ending up on the
iron. Shake (sponge) it off and repeat on the other side. If you get
gummy shorts, fix it by adding fresh solder.

I wonder why I haven't seen a hot solder dispensing pencil for that purpose. If
we could balance the diameter of the pencil vs. the viscosity of the solder and
the amount of solder that will be wicked up by a typical pad, and have a flux
dispenser right next to the pencil's nozzle, we could have a product here.

The part should self-center on the pads while you're at it.

If it's small enough, maybe. But you do get things like tombstoning on
resistors, and you can get parts to totally misalign themselves with hand
assembly, the best bet is to use some kind of adhesive to hold the part in
place, or at least some sort of mechanical contrivance, I use a dentist pick on
a 'third arm' thing.

 

[Richard]

Paul Burridge wrote:

I can't avoid 'em any longer. I've got no choice but
to solder this itty-bitty spec of sh*t to a PCB and
would appreciate a steer on the best way. Three
pins per side with 1mm between pin centres.

You'll find there are several techniques. From trial & error (and later
a little advice from a rework tech at Intel), here's mine for working
with 0.5mm pitch, 128-pin SMDs:

* Flux the pads
* Position the piece
* *Very light* downward pressure to prevent shifting
* Place a small bit of solder on the iron tip
* Tack one corner pin
* Ensure device is positioned properly
* Tack opposite corner

* Flux the pins
* Place a small amount of solder on the iron
* Stroke each pin away from the body toward the tip
- Flux will help the solder wick around/under the pin
- The stroke guides solder balls to the tip of the foot
- You can do several pins at once if the tip is wide
* Cleanup with solder braid

Some key tricks:

- I didn't have good success with the "blob it on, run to the end of the
row" technique at lower temperatures. The pin-by-pin technique is a
little more tedious, but more consistent with fine-pitch devices.

- *Don't* use a very fine tip. I started with a narrow-tip 1/32"
"screwdriver" tip, but discovered that it had problems transferring
enough heat on larger packages. A conical 1/32" screwdriver tip worked
much better, as did a common 1/8" conical screwdriver tip.

- Good tinning on the tip is key, since you want the solder to flow off
smoothly, not ball up and come off in globs. If you do many, you may
find yourself replacing the tip regularly. There are good tricks to
prolonging the life of the tip's tinning - search google for 'soldering
iron tip tinning'. Basically, cover the tip in solder when it's idle in
the stand to prevent oxidation; wipe it off before you start a joint,
not after.

- You don't need thin-gauge solder, since you're applying it to the tip,
not the pad. Some folks prefer paste solder, but I've had better
success with wire solder.

- Use a flux pen, like MG Chemicals' (Fry's Electronics). It goes on
thin, applies easily, makes less mess.

- Flux is your best friend here. You're bound to get bridges or solder
globs behind pins. They can be coaxed out with flux, an outward
stroking motion, and braid to remove excess.

- Rather than using a stereo microscope, I use a jeweller's magnifying
visor in the strongest strength to do the soldering, with a jeweller's
loupe for inspection. Micromark.com sells both, and the loupe attaches
to the visor.

- Bright lighting is good. I use a small 20W halogen desk lamp on an
arm to position it close to the work and move it around to minimize
shadows on the pin edges.

- Use a temp controlled iron, set to ~220 C. Check the specs on your
parts for maximum soldering temperature and time tolerances.

- If you'd like some practice, check out the dummy parts at
http://www.practicalcomponents.com and practice first on a spare PCB.
They supply exact mechanical parts for about $1 each - worthwhile if
your live parts are $10-$30.

Cheers,
Richard

 

[Spehro Pefhany]

On Thu, 07 Aug 2003 07:55:49 -0700, the renowned John Larkin
<jjlarkin@highlandSNIPtechTHISnologyPLEASE.com> wrote:

Another technique: get some liquid RMA flux. Position chip on pads and
put some flux on/under each pin with a toothpick. Now put a small blob
of solder on the tip of your iron and just touch each pin. The solder
will slurp off the tip onto each pad sort of automagically.

What type do you use? Does it clean up well for high-Z circuits?

Best regards,
Spehro Pefhany
--
"it's the network..." "The Journey is the reward"
speff@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com

 

[John Larkin]

On Thu, 07 Aug 2003 09:27:37 -0700, Richard <rh86@azglobal.com> wrote:

Paul Burridge wrote:
I can't avoid 'em any longer. I've got no choice but
to solder this itty-bitty spec of sh*t to a PCB and
would appreciate a steer on the best way. Three
pins per side with 1mm between pin centres.

You'll find there are several techniques. From trial & error (and later
a little advice from a rework tech at Intel), here's mine for working
with 0.5mm pitch, 128-pin SMDs:

* Flux the pads
* Position the piece
* *Very light* downward pressure to prevent shifting
* Place a small bit of solder on the iron tip
* Tack one corner pin
* Ensure device is positioned properly
* Tack opposite corner

* Flux the pins
* Place a small amount of solder on the iron
* Stroke each pin away from the body toward the tip
- Flux will help the solder wick around/under the pin
- The stroke guides solder balls to the tip of the foot
- You can do several pins at once if the tip is wide
* Cleanup with solder braid

Some key tricks:

- I didn't have good success with the "blob it on, run to the end of the
row" technique at lower temperatures. The pin-by-pin technique is a
little more tedious, but more consistent with fine-pitch devices.

- *Don't* use a very fine tip. I started with a narrow-tip 1/32"
"screwdriver" tip, but discovered that it had problems transferring
enough heat on larger packages. A conical 1/32" screwdriver tip worked
much better, as did a common 1/8" conical screwdriver tip.

Metcal makes a "hoof tip" just for this. It looks like a medium-small
conical tip that has a flat on one side. When you put solder on it, a
nice juicy blob parks on the flat, and as you run the pointy tip down
the row of fluxed pins, the solder slurps off onto each pin in just
the right amount. My people solder 240-pin FPGAs in just a minute or
two this way.

If anybody is doing any serious amount of soldering/desoldering,
Metcal gear is well worth it.

John

 

Andrew,

I totally agree with what you've said below. I used to do i486 and
Pentium mainboard rework by the palet before and this is exactly what we
do from 8 pin SOIC to 208 pin QFPs. Flux, right heat, medium soldering
iron tip (small one don't deliver that much heat) and 63/37 solder will
do the job. The result is almost like that of the production ones after
cleaning up the excess flux.

As for the braid, I use it for stubborn solder bridges and I don't
recommend the shotgun approach of filling everything with solder then
clean up with the braid as what some has suggested. You'll end up with
misaligned pins and damaged ugly board.

For SMT resistors and caps, if you have two irons then just put one tip
on each end and lift or you can buy those fancy tips.

PJ

Andrew Paule wrote:

What you need is FLUX - I solder these kind of things with a moderate
sized iron, regular (0.028) solder, and flux.

Get a corner pin located, tack solder, locate other corner, tack solder
(just put a little solder on the tip of you iron) -flux the pins on a
side you didn't tack, and then drag the tip and solder along, repeat,
and clean with flux remover -I learned this from a lady who did
commercial rework on SMD's for a lving - I've put 0.5mm parts on with
not problem pretty quick. Now a 201 resitor or cap is another story -
I'm a tombstoner. Anyone know how to do these things well?

Andrew

Paul Burridge wrote:

I can't avoid 'em any longer. I've got no choice but to solder this
itty-bitty spec of sh*t to a PCB and would appreciate a steer on the
best way. Three pins per side with 1mm between pin centres. I've got
the finest soldering iron tip I can find and a lab-grade stereoscope
so I can see what I'm doing. Now all I need is to *know* what I'm
doing. Any constructive advice gratefully received...
--

"I believe history will be kind to me, since I intend
to write it." - Winston Churchill

 

[Siol]

"John Larkin" <jjlarkin@highSNIPlandTHIStechPLEASEnology.com> wrote in message

Metcal makes a "hoof tip" just for this. It looks like a medium-small
conical tip that has a flat on one side. When you put solder on it, a
nice juicy blob parks on the flat, and as you run the pointy tip down
the row of fluxed pins, the solder slurps off onto each pin in just
the right amount. My people solder 240-pin FPGAs in just a minute or
two this way.

I'm using this thing (its equivalent from Weller) to solder 100 pin FPGAs
and it does work. But you need flux to prevent solder from spilling over
multiple pins and shorting them together. I also found that this works better
with higher temperature, I do it at around 400C.

Soldered in about 600 over the years this way.

Siol

 

[Siol]

<pj1108_no_spam_@telus.net> wrote in message news:3F32A0E8.2F877BC1@telus.net...

Andrew,

I totally agree with what you've said below. I used to do i486 and
Pentium mainboard rework by the palet before and this is exactly what we
do from 8 pin SOIC to 208 pin QFPs. Flux, right heat, medium soldering
iron tip (small one don't deliver that much heat) and 63/37 solder will
do the job. The result is almost like that of the production ones after
cleaning up the excess flux.

As for the braid, I use it for stubborn solder bridges and I don't
recommend the shotgun approach of filling everything with solder then
clean up with the braid as what some has suggested. You'll end up with
misaligned pins and damaged ugly board.

I use flux and that flat tip from John Larkin's post. Just run it over the pins
and it will suck off all excessive solder between joints (surface tension).

Siol

 

[John Larkin]

On Thu, 07 Aug 2003 16:57:47 GMT, Spehro Pefhany <speff@interlog.com>
wrote:

On Thu, 07 Aug 2003 07:55:49 -0700, the renowned John Larkin
jjlarkin@highlandSNIPtechTHISnologyPLEASE.com> wrote:

Another technique: get some liquid RMA flux. Position chip on pads and
put some flux on/under each pin with a toothpick. Now put a small blob
of solder on the tip of your iron and just touch each pin. The solder
will slurp off the tip onto each pad sort of automagically.

What type do you use? Does it clean up well for high-Z circuits?

Best regards,
Spehro Pefhany

Kester 1585. Seems clean. Immediately after soldering (well, within an
hour or two) we soak it in the deflux basket in our vapor degreaser,
then spray it down good with clean (distilled) solvent. Most of our
stuff wouldn't mind a few na leakage, but we do make a cryo
temperature signal conditioner that needs pa-level leakages, and that
seems fine; we bake and coat that one.

John

 

[Richard]

Siol wrote:

I'm using this thing (its equivalent from Weller)
to solder 100 pin FPGAs and it does work.

Do you happen to have the tip number handy? I'm using a conical tip
with a flat grind on it ("screwdriver" tip), but I haven't seen a hoof
tip from Weller like John describes.

But you need flux to prevent solder from spilling
over multiple pins and shorting them together.

Absolutely.

I also found that this works better with higher
temperature, I do it at around 400C.

Yes, I've seen the same, but what about the temp tolerances for the
parts you're soldering? I haven't picked over the JEDEC specs (if they
apply to temps), but at least one of the TQFP128's I'm using tolerates
10 seconds at 220C max.

Soldered in about 600 over the years this way.

Obviously you're not frying chips (at least not obviously .

 

[Siol]

"Richard" <rh86@azglobal.com> wrote in message news:3F32CE11.F7F084E2@azglobal.com...

Siol wrote:
I'm using this thing (its equivalent from Weller)
to solder 100 pin FPGAs and it does work.

Do you happen to have the tip number handy? I'm using a conical tip
with a flat grind on it ("screwdriver" tip), but I haven't seen a hoof
tip from Weller like John describes.

No, a salesperson from Weller recommended this at some fair and I picked it up
at the local electronics shop. It looks like the Ersa 832 PW as much as I see from
the catalog. The tip is not just flat but slightly concave.
I can make a dig. closeup for you.

I also found that this works better with higher
temperature, I do it at around 400C.

Yes, I've seen the same, but what about the temp tolerances for the
parts you're soldering? I haven't picked over the JEDEC specs (if they
apply to temps), but at least one of the TQFP128's I'm using tolerates
10 seconds at 220C max.

Well, the leads are tiny and don't transfer much heat, I guess. And with this tip
you're heating mainly the leads and not for long at it.

Soldered in about 600 over the years this way.

Obviously you're not frying chips (at least not obviously .

Never fried a single one that way.
I did destroy a few before I learned to do it properly. The problem was that I
occasionally destroyed the PCB or broke the pin while fixing the bridges.
Maybe 5 PCB's in total, which is still good. 5/600...

Siol

 

[Richard]

The tip is not just flat but slightly concave.
I can make a dig. closeup for you.

Looks like they don't offer this tip in the ET tip line, which is what
my iron uses. Too bad. Ah well, at least my new one is temp
controlled, which is an improvement on my old general-purpose stick
iron.

I would be interested in a part number, URL, or photo for reference,
though. Sounds like a handy part. Is it in the same price range as
common tips?

Cheers,
Richard

 

[John Larkin]

On Thu, 07 Aug 2003 15:09:21 -0700, Richard <rh86@azglobal.com> wrote:

Siol wrote:
I'm using this thing (its equivalent from Weller)
to solder 100 pin FPGAs and it does work.

Do you happen to have the tip number handy? I'm using a conical tip
with a flat grind on it ("screwdriver" tip), but I haven't seen a hoof
tip from Weller like John describes.

Oh, one other tip: Metcal makes two hoofers, big and small, and my
people prefer the big one. It holds more solder and works great on the
fine-pitch parts.

All this stuff is counter-intuitive.

John

 

[Richard]

John Larkin wrote:

Oh, one other tip: Metcal makes two hoofers, big and small, and my
people prefer the big one. It holds more solder and works great on the
fine-pitch parts.

All this stuff is counter-intuitive.

I gather your team does a fair amount of hand SMT work - what have you
seen that makes the Metcal's so superior?

Their approach to tip cartridges is new to me, so that no calibration is
needed when changing tips, but with so much integrated into the tip I'd
be concerned about the replacement costs. How often are you finding
that the tip cartridges need to be replaced?

And when you get into the "serious" price range, why wouldn't your folks
prefer to use a hot air pencil instead of an iron?

 

[Alex Gibson]

"Richard" <rh86@azglobal.com> wrote in message news:3F32CE11.F7F084E2@azglobal.com...

Siol wrote:
I'm using this thing (its equivalent from Weller)
to solder 100 pin FPGAs and it does work.

Do you happen to have the tip number handy? I'm using a conical tip
with a flat grind on it ("screwdriver" tip), but I haven't seen a hoof
tip from Weller like John describes.

xytronics also make a "wave" tip
see http://www.howardelectronics.com/xytronic/tips.html

 

[John Larkin]

On Fri, 08 Aug 2003 07:46:50 -0700, Richard <rh86@azglobal.com> wrote:

John Larkin wrote:
Oh, one other tip: Metcal makes two hoofers, big and small, and my
people prefer the big one. It holds more solder and works great on the
fine-pitch parts.

All this stuff is counter-intuitive.


I gather your team does a fair amount of hand SMT work - what have you
seen that makes the Metcal's so superior?

Temperature regulation at the tip is extreme. It pours in as much
power as it needs to. Great for soldering test circuits onto
copperclad.

Their approach to tip cartridges is new to me, so that no calibration is
needed when changing tips, but with so much integrated into the tip I'd
be concerned about the replacement costs. How often are you finding
that the tip cartridges need to be replaced?

Tips are expensive, $20 up, but they last forever. You can turn off a
metcal when it's not in use, because it warms up in a few seconds.
There's not much in the tip, actually; just a curie-point ceramic
thingie, pumped with RF from the main box. Sounds stupid at first, but
it really works.

And when you get into the "serious" price range, why wouldn't your folks
prefer to use a hot air pencil instead of an iron?

Hot air just blows the parts around, and doesn't transfer much heat.
We have a hot-air pencil, and sometimes use it for desoldering, but
the production people don't like it... infrared works better. For most
regular assembly, we solder-paste stencil, pick-and-place, solder in a
conveyer oven, and wash in a solvent degreaser thing. Hand soldering
is used for small stuff, rework, and sometimes for fine-pitch FPGAs.

John