NEW APPROACH - BORROW CONNECTORS FROM A SACRIFICIAL IBM SLT CARD
I came to the realization that if I could take the specially shaped connectors from an IBM Solid Logic Technology (SLT) card and install them on my board, it would give me a card that will snap into the backplane socket and lock in place. This absolved me from the need to design and build a hold-down method for the card.
I grabbed a card from my stock, took the plastic connector covers off, and then unsoldered the 24 connectors. I hadn't shaped the copper pads on my printed circuit board to the exact shape of the IBM connectors, but I experimented soldering some onto the card and found that they would work with the existing PCB design.
I used my heat table and hot air rework tool with solder paste to attach the connectors. Many of them slid into decent alignment, but others were skewed or shifted relative to the card edge. The plastic connectors that fit over the card connectors have dividers that fit between each contact, thus alignment is more critical than simply to ensure pin contact without shorting to adjacent connectors.
GOING TO NEED SOME KIND OF A JIG OR NEW INSTALLATION PROCESS
The next and more serious problem arose when I flipped the card over to install the connectors on that side. The heat that melted the solder paste and allowed the contact to pull into position was conducted to the underside and also remelted the solder under the previously installed contacts.
The contacts on the bottom side would either fall off or shift out of position. Some turned 180 degrees from their intended positioning. While I could finesse the contacts enough to get the plastic covers over them, unless the contacts were all at the proper position at the edge of the card, the locking action of the backplane pins wouldn't be reliably applied to all 48 connectors. That might be solvable with enough fussing, but the desoldering effect on the other side of the card is the show stopper.
I had my hot air tool set to 180 C but the solder paste has a melting point of 138 C thus I should drop the temperature to maybe 145 C. My table was set to 130 C which is too high given the low melting point of the paste, so I should drop that to perhaps 115C. I still believe the heat will conduct through the PCB and release the contacts on the other side from where I am working.
One idea I had was to use two solder types, with different enough melting temperatures. I could install on side using the higher melting point solder, then lower the heat and use the other solder for the back side. If done properly, I would never soften the solder joints from the side I first installed. I just ordered some solder paste with a temperature of 183 C which should arrive tomorrow.
That wouldn't perfectly solve the alignment issues, so I also am investigating some kind of alignment jig that could hold all 48 connectors in the proper spacing on the pads of the printed circuit card, so that I could somehow get hot air in to melt the solder paste for all the pads. Once they cooled I could remove the jig and have all the connectors in their intended locations.
The jig would need to be able to withstand temperatures of about 150 C (around 300 F) since I was using 138C solder paste. It also must not stick to solder. 3D printed nylon parts generally use filament that melts at about this level, making them unsuitable. Quick manufacturing with laser cutting or 3D printing typically involves materials that won't stand up to the heat.
No comments:
Post a Comment