Surface mount soldering

Message Date: Mon, 30 Nov 1998 22:11:18 -0800
Author: "Mark G. Forbes"
                                                                                             
[Harold commented...]

Years ago I recall seeing a vapor phase soldering system for SMT
that I thought was pretty neat. I haven't seen it since. It consisted
of a Pyrex container with condensor tubes around the top. You'd put some
liquid in the bottom and run water through the condensor tube. the
liquid would be heated to boiling, but the condensor kept it from
escaping. You'd lower the SMT board into the vapor, and everything would
be instantly soldered (the parts were held in place with solder paste).
It's probably been 5 or 10 years since I saw that machine. No
one else seems to have seen it. Anyone know if it's still produced?

[and Herb added...]

I don't how easy this would be to implement but a medical electronics
manufacturer I worked for used the "Vapor Phase Process" for surface mount
circuits. This involved using two chemicals, one had a lower boiling point
but a higher density than the other. These chemicals would be heated in a
vat. The chemicals would vaporize but one would stay above the other like
a thermal blanket retaining the heat of the other chemical vapor in the
vat. The circuit board with solder paste silk screened on it would have
the components placed on it and lowered into the vat of hot vapor which
would melt the paste. The circuit boards would be of ceramic cladded metal
which could stand up to the high vapor heat. The solder paste was finely
ground tin/lead solder mixed with rosin. Variations would include using
hot air instead of the chemical vapors.

[And my point is....]

I think vapor phase reflow has been pretty much abandoned these days; it
used flurocarbons, which we've since discovered are not-so-nice for the
ozone layer. (Or so it's said by some, anyhow.) A board would be populated
with parts after solder paste screening, then lowered into the vat of hot
vapor. The vapor would condense on the board, giving up its heat and dripping
to the bottom of the tank where it was re-boiled. After a dwell time, the
board emerged from the tank and went through a cooldown cycle. It was not
necessary to use a special board; plain old FR-4 works just fine. The metal-
core boards are usually used for performance or RF reasons, not because of
manufacturing method.

These days, most everybody I know is doing hot-air reflow. It's a simple
conveyor oven, with several temperature zones for preheat, soldering,
ramp-down and cooling. It's important not to thermally 'shock' the board,
since you can crack ceramic parts or even a trace if they're small. One
variant on hot-air reflow is hot-nitrogen reflow, where a bottle of
nitrogen gas provides a low-level continuous purge of the oven. This cuts
down on oxidation and improves flux wetting. Some people are still doing
radiant oven reflow, but the hot-air process produces more repeatable
results, I'm told.

If you can lay your hands on an old convection oven (check garage sales or
Goodwill) they work great. It's just an oven with a high-temperature fan
blowing the air around inside. Put it in, set it for 30 degrees or so above
the melting point of your solder, and let it sit until the solder melts.

I just hand-solder SMT's with an iron. It's not really that hard; get
a head-mounted magnifier and a Metcal iron with the 600 degree 1/64" tip.
You also want to use an RMA flux solder (Kester "44" rosin works best,
Multicore is awful) in the finest gauge you can get. I use the 63/37 alloy,
in a .010 diameter. A dab of RMA rosin flux in an insulin syringe also
helps a lot. Using this setup I've soldered my prototype runs (a dozen
or so boards sometimes) with parts as dense as 208-pin QFP's on 20 mil
centers. It's not really that tough, once you get the hang of it.

If you go with the oven approach, you'll still need the hand-work equipment,
since you'll be chasing down bridges on your QFPs. If it's just SO and
PLCC packages, you'll be fine, but watch out for bad joints on PLCCs; they'll
tend to flow out and fail to wick up the leg. I try to avoid them for
that reason when I'm designing a board and have a choice.

If you want to try vapor phase soldering, my CRC handbook lists
Dodecane [CH3-(CH2)10-CH3] with a boiling point of 216C. 63/37 solder melts
at 183C, so that would work. Just be aware that it's flammable (it's
a long-chain hydrocarbon) so any ignition sources would probably put
an early end to your experiment. I'd suggest you don't do this, but
I include it for reference. If you die trying it, don't blame me. :-)

Mark G. Forbes, R & D Engineer | Acres Gaming, Inc. (541) 766-2515
KC7LZD | 815 NW 9th Street (541) 753-7524 fax
forbesm@peak.org | Corvallis, OR 97330
http://www.peak.org/~forbesm
mforbes@acresgaming.com

"There has been an alarming increase in the number of things I know nothing
about."
---Anomalous