Drive PNP Darlington transistor with microcontroller

[brodin]

I am trying to drive a PNP darlington power transistor (BD680) with a microrontroller (ATMEGA32) output with the following circuit. Note that he circuit for the microcontroller itself is not complete, i just put it there to show what device i am using. I know that the microcontroller is running ok. The supply voltage for the uC is 5V but i want to control a 12V load with the transistor.

My software toggles the direction register of PORTC once per second. When the pin becomes a high impedance input I expect the load to "turn off", and when i set the pin an output in low state i expect the load to "turn on".

This works good in simulations in Proteus ISIS. But when i set up this circut with real components the transistor is always "turned on".

When the uC pin is an high impedance input i get about 10 volt on the base on the transistor instead of 12v (which would make (Vemitter-Vbase) == 0).

Is it not possible to design the circuit like this? What am i doing wrong?

Some tips would be very much appreciated!

 

[k7elp60]

If you change the L1 to between the emitter and the +12V line, when the ucontroller output is low the lamp will turn on.

 

[brodin]

The problem is that i am building a small circuit board that needs to have 12V outputs, which i want to control with a uC. So the load is not always a lamp, it could be anything that is driven by 12V more or less. Therefore I do not want a "grounding" output, but a "driving" output.

 

[MikeMl]

You forgot one very important detail: When the port pin turns off, it cannot be forced more than 0.5V higher than the Vdd of the ucontroller without forward biasing the parasitic diode structure inside the I/O pad. The base resistor will try to force current into the port pin, but the diode will clamp at about 5.5V, meaning that the PNP darlington is still turned on.

Here is how I would do it:

or another version:

 

[brodin]

Thanks a lot for the answers!

I will go with using a small NPN transistor (eg. BC547) to sink the base voltage of the PNP power transistor, just like you suggested Mike.