Are hole-through components in a SMD always hand soldered?

One of those questions I never found an answer to. Consider this design for a B&O amplifier. Most the polarised caps are hole-through but there are a few that aren't. Other than the larger farad caps it is purely a cost cutting exercise? Perhaps it made more sense when labour was cheap?

No, it won't be hand soldered - through-hole components have been flow-soldered for many decades.

Ok googled that.

So the top which is mainly smd would be baked or would the components all be glued and then wave soldered? I can't visualise a bath of solder leaving no residue. Then the through-hole are inserted by machine onto a spot of glue and the cutting of the legs is done prior to running the (now reverse) board back through the wave soldering machine?

As far as I'm aware SM components are always glued in place.

Surface mount components are usually only glued when a wave soldering process would disturb them. Single and double sided SM boards can be reflow soldered with solder paste without glue.

The surface mount components on the top of the board won't need to be glued in place for the wave soldering process.

Also, there is now a process called selective soldering which often means that surface mount components on the bottom don't need to be glued down either.

The traditional wave is solder flowing over a weir that is wider than the board, and the bottom of the board touches that wave while the whole length of the board is moved across the wave.

In selective soldering there is a small nozzle with solder flowing out of it and down the sides. It's a weir that is circular, maybe 6 - 10 mm diameter. That is moved around (or the board is moved to it) and only brought in contact where the through-hole components have to be soldered. If there are no SM components on the underside that are close to the pins of the through-hole components, they'll stay in place fine with no glue.

Interesting, so on that board above, they'd have; glued those smd components, flipped and wave soldered, flipped and fitted the through-hole and then flipped and wave soldered the rear?

Nigel Goodwin said:

As far as I'm aware SM components are always glued in place.

Click to expand...

Disagree. We only glued the components on the bottom side of the board. The tackiness of the solder paste is more than enough to hold the top side parts in place until they go through the oven.

Screen solder paste on bottom side
Robot applies glue dots on bottom side
Robot places bottom side components
Wait
Screen solder paste on top side
Robot places top side components
Run board through oven

Another variation is not to screen the bottom side, glue on the SMT parts, install any PTH parts, and wave the bottom side to solder everything. Preventing solder bridges among the SMT pins was tricky sometimes; you have to orient some parts with respect to the direction of the wave.

fun fact: SMT component placement by a robot does not have to be absolutely perfect. If a small part such as an 0805 resistor is slightly crooked on its pads, it will move into place when the solder becomes liquid in the oven. Surface tension and gravity equalizing fluid pressures.

ak

ACharnley said:

So the top which is mainly smd would be baked or would the components all be glued and then wave soldered? I can't visualise a bath of solder leaving no residue. Then the through-hole are inserted by machine onto a spot of glue and the cutting of the legs is done prior to running the (now reverse) board back through the wave soldering machine?

Click to expand...

Actually, there is another option called "pin-in-paste". The screen printer that deposits solder paste on all of the SMT pads also puts paste on the through-hole pads. Some of that paste is squirted down into the plated hole. Either before or after the SMT parts are placed by the robot (depending on the design), the PTH parts are hand-inserted. Then the board is run through the oven and everything is solderedWe did this mainly when adding connectors around the edges of an all-SMT board. The solder paste that was in the hole gets pushed out and hangs on the end of the pin; when it liquefies in the oven, it wicks up the pin to the pad and into the PTH, where it meets some of the top side solder wicking down. Some of our MIL customers were concerned with SMT connectors getting ripped off the boards in a "high-energy event". This technique, developed, tested, and endorsed by AMP, Molex, Burndy, Cannon, etc., made them happy. It was especially good at soldering the pins of high-density MIL circular (38999) connectors without the solder bridges common with a wave.

The connectors used high-temperature plastic rated to withstand the oven. At some point in the mid-2000's (working from faded memory), Samtec or Molex or 3M or someone like that went to using the high-temp plastic on everything, eliminating a part number hassle.

ak

Just to throw a wrench into the mix .... We currently have a board with through hole on both sides. In this case the side with the most through hole components are populated and we are populating the through hole components on the other side manually.

I've got a component which is smt and through-hole (USB-C port). The smt part is done first then on reverse paste is squirted onto the legs. I'm not entirely sure if this is what was intended but it does work.