current limiting with transistor

[perts]

i am making a high side driver with a pnp transistor with a gain of 100 to supply about 30 ma. the switch on the base of the pnp controls if 12 volts should go to the load

ckt: 12V---pnp-----load--gnd
|
|
/ switch
|
R
|
gnd

I also want to protect against short ckt, eg if there is a short in the load, i want to limit the current going through the pnp. Can I just size my R on the base of the pnp so that it only allows only 1 ma on the base which then it turns only allows for maximum of 100 ma (x100 gain) to the load.

Is that a suitable short ckt protection scheme. Or do I need to put a sense resister in series with the load and turn off the pnp using a npn across the Rsense.

 

[IRstuff]

If you recall EE101, transistor beta is not a usable design parameter in this fashion. You can only depend on beta being larger than some value, but not being less than some other value.

A sense resistor is one approach and may work, depending on the amount of current you're willing to tolerate and the degree to which you need to maintain 12V across the load.

You can also come up with something to monitor if the output voltage is more than 2 V below the input voltage for shutting off the drive.

TTFN

 

[OperaHouse]

If you have some extra voltage to play with, you could use a three terminal regulator in series. Just size the resistor from the "output" to the "common" for the current. A 5V regulator limiting at about 30ma would be 170 ohms. The "common" is now the output. Don't forget the caps or it may oscillate. These devices make great current regulators. It seems you don't care about a stable voltage.

A small lamp is also useull as a current limiter since its resistance increases dramatically when the current increases. Look into a two lead device from polyswitch that acts the same way.

 

[felixc]

As IRstuff says, a current limiter has a side effect of reducing the voltage at the load, and when in current limit condition the voltage that appears at your trannsistor will generate heat. Can you stand it or do you prefer to see the load being turned "off" upon an overload?
Polyswitches are nice devices to work with, but they are not perfect resettable fuses. The trip current varies with temperature and they feature a "maximum current" spec which you also must take into account when you evaluate the overload or short circuit conditions.
Defining the working conditions of your design will tell which way to go.

 

[melone]

You can use a second transitor, and a series resistor to perform over current protection, however, you will incur a drop across the resistor. This is a VERY rough solution. Basically, all you need to do is wire up a second PNP BEFORE your pass transistor, and place a series resistor of ~2.5ohms between the Base-Emitter junction. For example:

12V -> Emitter #2 AND 2.5ohm
2.5ohm -> Base #2 AND Emitter #1 (Existing PNP pass element)
Collector #2 -> Base #1

Size R so that once your second / sense PNP turns on, then the voltage at the base of the pass PNP biases the junction "off". Actually, what will happen is that the circuit will find some intermediate stage that will keep both transistors in the active region, but will limit the current to the load.


ckt: 12V--2.5O--pnp-----load--gnd
| | |
-----pnp----|
/ switch
|
R
|
gnd

 

[melone]

Well? What was your final solution?

 

[perts]

thanks.

mine turned out to be more complicated.

pnp1 with a 10ohm sense resistor in series to the load. the pnp1 was controlled by another pnp2, where the collector of pnp2 was tied to base of pnp1. when pnp2 was on, pnp1 was off and vice versa. across the 10 ohm sense resistor there was npn1 where the collector was tied to pnp2 base. when more than 60 ma went across the sense resistor, it turned on npn1, which then turned on pnp2 which caused pnp1 to turn off. this push pull limited the current to a maximum of 60 ma.