
Profile temperature measurements that cycle up and down
Why are some profiles more wavy than others?
Wavy looking profiles may be caused by something not always expected. In this case we were reflow soldering a very large and thick board; a back plane board that was 14 inches wide and 30 inches long and over 0.25 inches thick. One would expect that a board of this mass would heat very slowly and with a somewhat smoother temperature profile graph.
The cycling up and down of the profile temperature, especially in the area between oven zones, was in this case caused by the boards large thickness. The surface of the board was quickly heated by the oven’s convection. Where that convection rate is lower, such as between air jet holes and between zones, the inner core of the board, which is not yet as hot as the surface due to the boards low thermal conductivity, pulls the surface temperature down toward the core temperature. Then as the boards moves into a new zone or under a convection air hole, the surface temperature is again pushed back up to a higher temperature, creating the wavy up and down look to the profile. It’s important to remember that the thermal profile of very thick boards, depending on their internal thermal conductivity, may behave like this as they are reflow soldered, even though common sense may say they should have a much smoother profile look.

#1 by Joe Turnbull - March 27th, 2009 at 13:23
I have found that wavy plots can be caused by excessive
ventilation. This will cause the heaters to cycle more frequently. This will show up in your profile. To much air flow cools the oven tunnel, then the heaters respond. This will also decrease the life of your heating elements. Use a anemometer to check velocity. Adjust to oven manufacturer suggestions.
#2 by TylerShay - May 6th, 2009 at 09:31
We’ve been reflow profiling for a lead-free transition and as such we’ve been putting thermocouples on the undersides of BGAs to make sure they’re hitting a high-enough TAL to melt the solder balls. Anyway, I didn’t secure one of the thermocouples well enough and it ended up kind of coming loose but not all the way off the BGA, it was kind of hanging on the edge of it, and it caused the graph to get really wavy because sometimes it would collect air temperature and others it would collect BGA temperature. Live and learn I guess.