[EE]:: SMD soldering and desoldering
RussellMc
desoldering tool" thread. As it deals as much with soldering as
desoldering and as it gives a fairly detailed commentary on a variety
of methods it seemed worth positing in its own right.
____________________________
Ken says:
We have a Hakko 936 temperature-controlled soldering station, a Hakko 850
hot-air SMT rework station, and a Hakko 474 vacuum desoldering station.
All three tools work very well and I'm confident to use them to solder and
desolder pretty much any type of component having any lead pitch on boards of
any number of layers (with the exception of BGA and similar high-density large
pin-count leadless packages).
Soldering fine-pitch SMT devices using the 936 requires a wedge type tip, liquid
flux, and sound technique - but once you get used to it and have practiced a
bit you can for instance lay down and solder a 240-pin 0.5mm pitch QFP in less
than 30 seconds. I believe that you can get special SMT soldering tips that are
in the form of a sort of "spoon" that can be charged with solder - but I've
never bothered as I find that the standard Hakko SMD wedge tip (Hakko 900M-T-K)
works well enough for me.
The basic technique is to first place the IC on the board, accurately align all
leads with the pads, apply a small amount of liquid flux to the corner pins,
press down on the package gently (I used the tips of a pair of fine tweezers)
and tack a couple of the corner pins. Next apply liquid flux liberally to the
leads along one side, charge the wedge-type tip with enough solder to do all the
leads on one edge of the IC, and then just drag the sharp edge of the along
(pulling the iron towards you) in contact with the pads using the lead-ends as a
guide.
Sometimes you will get solder bridges between leads but these are easily dealt
with by either wiping the tip (with as little solder on it as possible) along
the offending leads in an inwards-to-outwards direction, or if necessary by
using desoldering braid. Extra liquid flux is also often helpful when dealing
with solder bridges. For IC's having leads on four sides solder the opposite
side next and then the other two. If some leads fail to solder on the first
pass just apply a bit more flux (and perhaps put a bit more solder on the tip)
and try again.
Soldering passives with two leads using the 936 is also very easy. You place
the component (say an 0805 resistor) so it overlaps the pads by about 25-30% (no
need to be too fussy with alignment) and apply liquid flux. Charge the wedge
type tip with an appropriate amount of solder (it takes a while to get used to
judging that), and bring the tip into contact with the two pads (as near to
simultaneously as possible). The solder will flow across the pads and when it
reaches the device the surface tension of the solder will pull the device into
proper alignment relative to the pads, and the tip can then be withdrawn. Done
right the soldered joint should be every bit as good as one done in a reflow
oven.
I generally set up a number of passives and then solder them all one after the
other. Once you have practised the technique you can lay down and solder
passives and other 2-leaded SMT devices surprisingly quickly.
I almost never use the 850 for soldering as solder paste is a pain. Not only
does it take a comparatively long time to apply to the pads, it always dries out
eventually in the syringe and you are for ever cleaning the hypodermic. You
also have to be much more careful in placing the IC after applying the paste.
It's much better to use the 936.
Desoldering IC's etc. from SMD boards using the 850 requires an appropriate
"nozzle" for the target package - and they are relatively expensive. We have
built up a reasonable selection of nozzles and have on occasions temporarily
adapted a nozzle to suit a different package using adhesive metal foil, or thin
tinplate or aluminium. Just heat all leads of the device evenly, while
maintaining a gentle sideways force on it using the tips of fine stainless steel
tweezers. Alternatively use thin piece of stainless steel wire slid under one
corner of the device. As soon as the device moves sideways withdraw the nozzle
and flick the device off the pads. With a bit of practice the device will come
away cleanly on the first attempt. Don't apply too much force and don't apply
it too early or you run the risk of lifting pads. Clean up the pads with
desoldering braid and the 936 - taking care to always wipe along the longer
axis of the pad from outside to inside, and clean any flux using an appropriate
flux cleaner on a rag. If done right the board should look as if an IC has
never been in place.
Smaller SMD devices with relatively few pins can best be desoldered using a
round nozzle on the 850. We have a 3mm nozzle (good for discrete semis in
SOT23's and small IC packages) and a 5mm nozzle (good for larger discretes like
rectifiers and somewhat larger low pin count IC's).
Note that if a high pin-count fine-pitch QFP is KNOWN to be faulty, the safest
approach to its removal is to use a scalpel to shear off all leads right at the
point where they enter the package moulding, let the package body fall away, and
then use the 850 with a 3mm nozzle and tweezers to remove the leads a few at a
time. Clean up the pads as described above.
The 474 excells are removing leaded components and has enough heat capacity to
desolder large component leads (as on large electrolytic capacitors) from holes
embedded in large copper planes. The tricks are to use the right temperature
setting, keep all the filters clear (replace regularly), empty the solder
reservoir regularly, and to use a nozzle with the right hole size. I have
adapted a number of smaller nozzles by drilling them out so I can better deal
with larger leads. I also split the ceramic fibre filters in half by peeling
them apart. Not only does this improve air flow but you only have to buy them
half as often.
One trick with the 474 (essential on boards having minimal solder) is to first
apply solder to all leads of the target component so there is a generous amount
of solder on each joint. It is a bit counter-intuitive but it's necessary to
get a good thermal path from the tip of the iron to the joint (and the flux from
the solder may help also). The other trick is to apply the tip to the joint and
wait for the joint to heat up (and the for the solder to melt completely all the
way through the hole and on the component side of the board) before pulling the
trigger to apply vacuum.
The technique is to pull the trigger and then sort of rotate the tip so it
orbits around the lead a couple of times. This causes the lead to orbit in the
hole and helps encourage the solder to be sucked out. Don’t apply too much
sideways force to the lead so there is a danger of bending or breaking it, and
apply only the minimum pressure to the pad (otherwise there is a danger of
damage while the substrate-copper adhesive is hot and soft.
Regards,
Ken Mardle
_______________________
_soldering
--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
alan.b...@stfc.ac.uk
> capacity to desolder large component leads (as on large electrolytic
> capacitors) from holes embedded in large copper planes. The tricks
> are to use the right temperature setting, keep all the filters clear
> (replace regularly), empty the solder reservoir regularly, and to use
> a nozzle with the right hole size. I have adapted a number of smaller
> nozzles by drilling them out so I can better deal with larger leads
This advice is as relevant to the Weller units as well, that I mentioned
previously.
--
Scanned by iCritical.
Jonathan Hallameyer
I'll second all of this, with just a bit extra caution pointed out to using
desoldering braid to clean up pads. Take care to not get lazy and have too
long of a "tail" of desoldering braid, because if you do, its quite likely
that it could get too cold, solidify, and end up yanking a pad off the
board, and reapplying pads certainly isnt basic rework anymore.
With enough flux, and a tip that works well for you, pretty much no leaded
package should be a problem. Some high thermal mass LCC packages can be
more difficult, as you can get solder on the metalization on the side, but
its not clear if its flowed underneath yet or not sometimes. On the
accelerometers I've put down (square analog devices ones, like sparkfun
uses) I like to add a bunch of flux and reflow it with the hot air pencil so
the surface tension can align it perfectly to the pads.
--
Jonathan Hallameyer
Roger Weichert
Jonathan Hallameyer wrote
>
> On the accelerometers I've put down (square analog devices ones, like
> sparkfun
> uses) I like to add a bunch of flux and reflow it with the hot air pencil
> so
> the surface tension can align it perfectly to the pads.
> --
> Jonathan Hallameyer
So you're saying that once you've soldered the device in, you then apply
more flux, and reflow it.
What sort of hot air pencil are you using?
What flux do you use?
How long does it take to get hot enough to reflow the device so that it
aligns? I would be nervous about overheatin the ic.
How big (how many pins) are the devices you use this technique with.
Thanks again for the education.
Regards, Roger
Jonathan Hallameyer
> Hi guys, this is a great thread and very helpful. Thanks.
>
>
>
> So you're saying that once you've soldered the device in, you then apply
> more flux, and reflow it.
>
> What sort of hot air pencil are you using?
>
> What flux do you use?
>
> How long does it take to get hot enough to reflow the device so that it
> aligns? I would be nervous about overheatin the ic.
>
> How big (how many pins) are the devices you use this technique with.
>
> Thanks again for the education.
>
> Regards, Roger
>
Roger,
Yes, I'm soldering the package, allowing it to cool then reflowing with hot
air. If the ceramic under the metalization of the "leads" on the LCC
packages isn't up to temp, wetting seems to be poor at times
I'm using the Weller HAP200 that came with the WR3000M I use at work, with
the airflow usually around ~15-30% no real way to quantify that, just low
enough to not blow the chip around :o) I use a bigger tip in it ( ~4 or 5
mm) and it looks as if this accelerometer is in the same package
http://www.sparkfun.com/commerce/product_info.php?products_id=843
I'm using some kester RMA flux, since I'm using 63/37 Sn/Pb solder, It may
not be the best for high production stuff (easier to clean, or no clean may
work in some applications) but lets not get side tracked onto a flux debate,
whatever works for someone works.
On just a bare board, without heavy ground planes or heat sinking, I'd say
16 SOIC is about as big as you want to go without specialized tips, and just
a big nozzle and a lot of hot air, as for overheating and how long, your IC
manufacturer should have at least some profiles for reflow, to get at least
a rough idea on how long you can heat it and you can actually heat too
fast, you don't want to just max the hot air pencil out and burn the crap
out of it.
Some of the fancy pants hot air rework stations have a thermocouple to put
on and it makes the chip follow the programmed temp profile.
If you're fighting with big ground planes or thermal pad on the bottom of an
IC, preheating can work wonders. If you're populating a bare board, then a
simple hot plate can get the chips and board up to a warm but safe temp
(I.e. the chip is fine at that temp for a while, it wont be damaged) and
just use the hot air to get the chip to reflow. IR is a lot better for
preheating populated boards, but its $$$$
With a WR3000M, a microscope, and some good tweezers, I've been able to
rework anything from 0402 passives, small chip-scale stuff, and some smaller
(10-12 mm ) BGA chips. Weller makes a shroud/pickup kit for the WR3000M
that has a few different metal shrouds, and a spring loaded plunger with a
silicone suction cup on the end, hook it up to the vacuum pickup on the
base station, and add flux, and heat until all the leads flow, and the
plunger gently lifts the chip off the board. With the shrouds, you can work
larger chips, but if you're doing a lot, I'd suggest a proper hot-air
station that accepts tips for different packages to direct heat at the
leads.
Time and practice :) Plus if you're using rosin flux, when smoke starts
pouring off the board, you're probably a little on the hot side :o)
-Jon H.