HP JetDirect 615n Reflow and Snow damage

Couple good ones this time...

I regularly help out at my son's school with computer issues, and the teacher there gave me an interesting one Friday.  An old HP LaserJet 2300 lost network connection, and she said there were consistent failures with these HP JetDirect 615n networking cards in the printer.  She wasn't willing to give up on the printer since they are bullet-proof and very cheap to operate in a school setting... so I dug in.  Apparently there is a know issue with the 615n cards (and many other current-day BGA-attached processors for that matter) the cards suffer from either a poor solder job on the main ARM processor or were poorly designed thermally that repeated cycling of the processor package results in solder contacts breaking from the ball-grid array under the main chip.

I took the cards home and  removed the plastic and metal shield around the network connector.
I prepped a lab ring stand with clamps to hold the boards 4" above a 1000F Master heat gun, the temperature control (air-throttle) was set to about halfway closed.  I mixed a little plain paste-flux into 91% isopropyl alcohol and dropped several drops around the perimeter of the main chip.  I placed the boards in a 130C (265F) oven for 4 minutes to pre-heat the board and drive off the IPA and water.  Turn on the heat gun, pull the board out and clamp it with the heat gun centered on the back side of the PCB below the main chip.  I tracked the temperature of the chip from the top with a cheap IR point-and-shoot temperature gun.  In about 90 seconds it ramped up to 200C.  I pulled the heat gun and aimed it carefully at the main chip for about 30 seconds to reach ~220-230C.  Then put the heat gun back below the board and switch it to fan only for about 2 minutes which cooled it back to 70C or so.  I ended up doing this on two cards, and both began operating in the printer again.  Personally, I was amazed.

 This is a test board where I successfully desoldered an 80 pin QFP chip to test the method before hitting the boards.

My other project was on my own roof.  The recent east coast snow storm dumped 38" at our house.  This time the snow catchers saved the gutters (which were trashed in the 30"+15" incher in 2010), but apparently the weight of all the snow melt sliding down my 10/12 pitch roof and solar panels was too much for the transition boxes to take the wiring to conduit.  Both boxes were ripped free of the roof and rolled over, snapping a 1/4" fiberglass tab on all four corners.  I was able to re-drill new holes and reattach, but even some of the MC2 connectors had pulled free of the module wiring.  This too some 3/4" poplar foot pegs to allow me to Spider-man/rock climb up the 10/12 pitch and reconnect wires that were two-thirds of the way up the slope...  this transition box system is going to get a re-design in the coming months...

Fiberglass box ripped off the S-5! clamps & conduit

Start Spider-Man!  My wife was kind enough to photo
for the insurance claim/hospital visit.

Where did that wire go!?!?!