| ??? 03/01/05 08:21 Read: times |
#88748 - Practical findings Responding to: ???'s previous message |
We are using 8x8 bi-color matrices ("dice") for (indoor) displays.
While running them on near the maximum rated current in a rather closed cabinet (together with the power supply and control board = sources of heat), in the long run (hours to days) when running an all-LEDs-on test, some of the LEDs remained dark. So we investigated further, trying to deliberately heat up the "dice", and found a surprising (at least for me) thing - depending on the manufacturer, the LEDs that stopped working started again - some manufacturers' when cooled down, and some when heated up... Our (unofficial) conlusions are: - LEDs in the "dice" suffer from bad heat transfer (bad cooling), as there is no direct metal lead going from the LED chip (and lot of quite well thermally insulating epoxy around) - different materials used in "dice" probably have different thermal expansion and they probably tear off the chip's bond or cause cracks in the PCB Jan Waclawek |
| Topic | Author | Date |
| Further thoughts on large LED displays | 01/01/70 00:00 | |
| Practical findings | 01/01/70 00:00 | |
| I'd agree | 01/01/70 00:00 | |
| cost... | 01/01/70 00:00 | |
| Except | 01/01/70 00:00 | |
| Heatsinking | 01/01/70 00:00 | |
| Luxeon heatsinking | 01/01/70 00:00 | |
| Much brighter discrete LEDs | 01/01/70 00:00 | |
| 5 x 8 matrices | 01/01/70 00:00 | |
| size | 01/01/70 00:00 | |
| Contradictory requirement! | 01/01/70 00:00 | |
| This works | 01/01/70 00:00 | |
| "spider" LEDs are cool | 01/01/70 00:00 | |
| but that requires | 01/01/70 00:00 | |
| ...and also... | 01/01/70 00:00 | |
| track width | 01/01/70 00:00 | |
| 1 mm track | 01/01/70 00:00 | |
| Display of LED problems | 01/01/70 00:00 | |
| color code | 01/01/70 00:00 | |
Big sign | 01/01/70 00:00 |