Transistor not fully turning off.

Hi, I have a slight problem I have traced down to a transistor not turning off fully. It is a 2n3906.

I control two 'common cathode RGB led' cathodes with a 2n3904 and 2n3906, the base resistors go to the same digital pin. Effectively when HIGH, the 3904 side is active, LOW the 3906.

I have tried pulling various things HIGH & LOW but I still can't get the 3906 to fully turn off. Is there some sort of de-coupling needed? Or maybe diodes to prevent the 3904 interacting?

I connected another 3904 to the common ground of the two transistors with its base LOW. With this separate transistor off the LED still had a slight illumination which confused me more.

I'm in the middle of spreading my design over a larger breadboard to debug easier, so I will try do up some schematics if I cannot explain clearly enough.

Any insight is appreciated. Cheers.

The output of the digital pin may not be high enough to turn off the PNP transistor. I use a second resistor from the transitor base to the power supply.

V+ -- R1 -- base -- R2 -- digital pin

I just gave that another whirl, equal resistors for R1 & R2 leaves the led illuminated the tiniest amount, but never turns on fully. With R2 = R1 * 1.3 the PNP is clearly open.

I have a spare buffer chip somewhere, I'll wire it up to see if it can turn off the PNP fully.

With R2 = R1 * 1.3 the PNP is clearly open.

Open meaning not conducting and LED is off?

edit: If you have a buffer that is open collector output, that would be best.

No I meant the PNP is allowing current through/closed circuit.
The buffer is a CD4050BE.

With and without a base resistor from buffer to PNP base the LED still lights up. I'll do up a circuit to post soon if more tests fail.

Are you certain you do not have the PNP connected reverse polarity? Emitter to power, collector to load.
http://www.fairchildsemi.com/ds/2N/2N3906.pdf

You always need a base resistor for what you are doing. A schematic will allow us to see what you are actually doing, your description is very ambiguous.

Absolutely we need to see the schematic.

Yeah I'm doing a schematic up as we speak, its taking a while so I'm going to finish it tomorrow.

One thing that always puzzled me is how the hell you can make anything without a schematic. I know I can't.

@Grumpy_Mike: I normally agree with your statement on more complex circuits, but I would not need a schematic for this.

edit: A picture might help in this case. My concern is the PNP transistor is connected "backwards" (collector to power, emitter to load).

My concern is the PNP transistor is connected "backwards" (collector to power, emitter to load).

That will work too, especially for turning the circuit off.

It just has low beta if wired backwards.

@dhenry: You are probably correct. I just can't remember what happens to the base-emitter current in reverse polarity.

While we all twiddle our thumbs waiting for this maybe mythical schematic:-

I just can't remember what happens to the base-emitter current in reverse polarity.

If you swap the collector and emitter round on a transistor it will still work but with a much reduced gain.

but I would not need a schematic for this.

And how many times do you screw it up? I always draw out something that involves three components or more.

...and what happens to the base-emitter current? The base and collector are at +5v, and the emitter is at ground. Is there no current passing from the base to the emitter in this configuration? I can't remember.

At this level of complexity, not in a long time! This should be three components. Two resistors and a transistor.

And I make as many of these kind of mistakes if I am using a schematic or not. A schematic does not prevent me from installing a transistor incorrectly. I have blown up my share of transistors, and most using a schematic.

If you think about it a transistor p = collector n = base p = emitter is still the same configuration of junctions when the emitter and collector are swapped.

What is different is only the geometry of the collector and emitter, with them swapped round the emitter is less efficient at emitting and the collector less efficient at collecting. The actual currents are in the same direction only the gain is very much lower.

@Grumpy_Mike: I have been thinking about it. I like reading your stuff, and value your technical expertise. That is why I asked you.

The collector and emitter were not my concern. I will presume zero current from collector to emitter. Does the base-emitter junction pass any current in this setup, or is it blocked? Normally (and you say this is kinda "normal") the base-emitter current could be enough to light a LED dimly.

Yes the base emitter is a diode like any other junction so if you mangle the connections you could get it either conducting current like a normal diode or if it is reverse biased there would be the reverse bias leakage, like a normal diode.
However, it is likely that the diode in this case would go into reverse biased breakdown and allow more current to flow thus lighting any in line LED.

I had this problem as well. Increased amperage same voltage and it worked fine after.