Posted for Jeff Peterson <jeffreyb.peterson@gmail.com>:

Hamid Wasti wrote (referring to this
http://www.barkeraircraft.com/files/AOA_rDisplay.pdf )

study the circuit on page 6, you will see that the LEDs are connected to the
"+14V In" input voltage and they are set up to run at 24mA not the "typical
10mA" that you pulled out of thin air. At 13.8V input and a drop of 2.2V on
the LEDs, the chip has to dissipate 11.6V at 24mA or 278mW. At a Theta-J-A
of 55C/W, this leads to a 15C increase per LED that is actually on.

I can verify that this is real error in the circuit, since I burned up an
LM3914 using a similar schematic.

It should solve the problem to add a series resistor for each LED.
This will reduce the voltage across the LM3914 and keep its temp within
limits.
Or use multiple LEDs in series and make light rather than heat.

By the way, we all could benefit from everyone trying to make more light
than heat...please?!

Cheers,

--
Jeff Peterson