You can’t shrink heat
Brad Thompson, Contributing Technical Editor brad@tmworld.com -- Test & Measurement World, 12/1/2007 2:00:00 AM
|
|
|
Regular readers may recall that I volunteer at a nonprofit computer recycling facility. Recently, the incoming e-waste stream yielded a batch of approximately 100 castoff HP Model 1530 15-in. LCD monitors. While many still worked, approximately one third showed only black screens.
Shining a flashlight on a typical screen showed faint black characters, indicating that the display’s LCD worked but not the backlight. Investigating further, I found a blown fuse that supplied power to the inverter that drives the backlight’s fluorescent tubes. Replacing the fuse restored operation, but then the fuse failed again.
| Read Brad's other articles about working at the computer recycling facility: Tossed any PCs lately?, So, where’s the mercury? |
I measured a current of 1 A through the 2-A fuse, well within its rating. A current probe showed normal-looking waveforms in both tubes’ return leads. In frustration, I jumpered the fuse and waited to see what would happen.
An hour later, the screen went black, and an inverter transistor (a TO-126-packaged Sanyo 2SC5706) released a wisp of smoke, obligingly unsoldered itself from the circuit board, and dropped onto the workbench. I incautiously poked it and received a small burn on my fingertip. Can you say “thermal runaway”? I could and did, along with a few unpublishable words.
I examined several failed displays and noted that the inverter circuit boards, manufactured by BenQ, had discolored areas around the inverter transistors. Mounted close to the board, the 2SC5706s’ 7.5-mm2 collector tabs hardly qualify as convection heat exchangers. Heat produced by the transistors dissipates primarily via the board’s copper traces.
The single-sided, low-density printed-circuit board layout crams the transistors next to the inverter’s step-up transformers. Finding suitable substitutes for the fast-switching and high-current 2SC5706 proved surprisingly difficult. I wanted to use TO-220-packaged switching transistors in place of the 2SC5706s, but there’s barely room for TO-18 transistors and small heat sinks.
Driving fluorescent tubes more closely resembles an art than a science (see sidebar), but in our industry’s crazed pursuit of “smaller, cheaper, and faster” electronics, we’ve evidently forgotten “hotter”—and heat doesn’t take kindly to miniaturization.
Jim Williams gets a shout-out in Forbes
08/26/2008Don't monkey around with analog circuits
08/01/2008Terms of Use
12/11/2002Pico tech opens accessories Web site
09/03/2003AF Web site aids collaboration
02/29/2004
-
FLIR offers IR camera for under $3000
-
Don't let the economy compromise quality (Guest commentary)
-
Danaher speeds up restructuring, acquires life-sciences businesses
-
Agilent’s Cover-Extend technology eliminates need for physical test points for in-circuit test
-
So many combinations: Testing a switch-matrix board
View All Webcasts
