The problem is the transistor isn't letting enough current through. When I connect the LEDs directly to ground they use 360mA, but when I use the transistor they only use 270mA (the LED cathodes to the emitter, ground to the collector, and base to either Arduino pin or directly to ground to test).
The datasheet says the transistor should be able to handle 800mA, but...obviously I'm missing something. Is there a better PNP transistor for this?
If you are going to use a transistor between the load (LEDs) and ground you should use an NPN transistor. Your PNP transistor would be placed between the power supply and the load (LEDs).
Thanks for the reply but I don't think I can use NPN transistors. They work when you place current on the base, correct? I need the transistors to work when you connect the bases to ground.
Eventually the transistor will be driven by the output of a 74hc138 (3:8 decoder), on this chip the nonselected outputs are high, and the selected output is low.
I suppose I could use inverters on the chip's outputs so I could use NPN transistors, is this the only way to make this work?
Can you move the transistors so that they're connected to the Anodes of the LEDs? As shown here:
If not, I think you'll need to get some NPN transistors.
Note than an HC138 can drive LEDs directly, about as well as an Arduio can. (OK for single, high-brightness LEDs, but probably not enough for the common cathode of a segment display.)
Thanks but I can't connect the transistors to the anodes; I need to connect those to LED drivers. Basically I'm building a smaller (16x16) version of EMS's Peggy.
Now I think the problem is just transistor selection, not circuit type.
The spec sheet says maximum current of the PN2907a is 800mA, well above the 480mA (16 LEDs @ 30mA ea) I need. But after spending the evening googling around, now I think the important spec is power dissipation, and 480mA * 5v = 2.4W, well above the 625mW that the PN2907a can do.
So I think I just need to find a different PNP transistor, one that can handle that much power?
Well, no. The relevant voltage for power dissipation is the voltage dropped across the transistor (which ought to be about .2V for a "turned-on" switching transistor), not the supply voltage that you're trying to switch.
74HC238 is a non-inverting equivalent to the 74HC138...