Selective soldering is often thought of as a “dark art”. In this series of posts we hope to pull back the veil and reveal the interior nature of this process as a whole with an emphasis on the strategy behind the programming of our new CRICKIT. Let’s dive right in!
Initial considerations: designing the panel
From the outset it’s important to consider how many units of a given item you project to make. In this instance we knew from the beginning that we would be making a LOT of CRICKITS. This is of crucial importance to the selective soldering machine programmers, operators and maintenance personnel. The lead-free selective soldering process is one that demands fastidious upkeep of the machine and its various components. The cleanliness of the solder pot, pumps, and nozzles are of crucial importance to maintaining process integrity and efficiency. Due to a variety of factors the cleanliness of these components will erode over time. It’s essential to consider the accumulated wear on mechanical components in the macro and micro sense.
In the aggregate, mechanical components and metals experience wear. Macro wear in the selective soldering realm is manifested in the calcified deposits at the base of the solder pot, filled channels within the impeller housing or even the relative loss of germanium within a lead free alloy such as SN100C. On the micro level, we see how wear can manifest during the soldering of a SINGLE PANEL! Large panels with many solder “sites” can take longer to solder. The longer a lead-free nozzle is exposed to air the faster it will oxidize. When a nozzle begins to oxidize, the solder “wave” breaks down and has greater trouble “snapping off” of leads (little legs or pins that protrude from underneath the board which must be soldered).
So, why am I telling you all of this?
I’m telling you all of this because it’s imperative when laying out a panel that you consider how many boards are physically ON that panel. A panel with too many boards won’t be good because the panel will begin bowing in the middle due to it’s own weight and the heat that it’s being subjected to. Conversely a panel that’s too small might not be great because you have that many more physical panels lying around you! How will you house and prep that many panels? Do you have enough space? One must consider this.
It’s all about striking a balance between these criteria. In the case of the Crickit we were thinking initially of going with a 4-board panel-in either a 1×4 or 2×2 configuration. We were initially concerned that larger panels would become too heavy once populated with SMT (surface mount) components from pick-and-placing as well as TH (thru-hole) components from selective soldering.
Aside from a lack of space, smaller panels inevitably run faster through the machine. Sure, that sounds wonderful but how much rework do you project is going to be needed on those panels? You don’t want to overwhelm the person (or people) who may have to rework them and you don’t want to simply shift your bottleneck to a different segment of your process. The name of the game is equilibrium.
Long enough to be short
We opted to go with a 6-board panel in a 2×3 configuration. A panel of 6 boards would do something I like to call “long enough to be short”-which just means that the panel’s runtime won’t be too long to cause excess oxidation of the solder nozzle but also not too short that we can’t rework the previous panel (if need be) while we have the next one in the machine.
We specified that there also be 1/2″ rails on the top and bottom of each row. The amount of board “real estate” you have when selective soldering is key to optimizing your soldering paths. Extra space gives you more room to snap off of solder sites and create excellent joints. If the rails are too narrow and you have a site near the edge of the board the nozzle may snap off into the machine’s rail and get knocked over. We gave ourselves an additional 1/2″ rail separating the rows in case we needed to design a jig to support the board if it indeed proved too heavy.
In our experience a glossy finish on a PCB can lead to higher frequency of solder balling. Glossy finishes seem to get marred more easily by higher heat (which we need when running lead-free alloy). For these reasons we specified that the Crickit PCB have a matte finish.
So now that we have our panel, let’s continue our programming journey! Tune in next time as we explore the strategies and tactics employed within our initial foray into solder and flux path creation!
AND! Here are some of the things we’re build WITH CRICKIT right now!
deep dive part 2:
We usually put side rails on our arrays, as well (completing the outside rectangle). That helps stiffen up the array even more. Not sure if there was some assembly reason for leaving those off?