Basic transistor theory ?

[tron87]

Whats more important to consider when switching loads using a transistor the max power dissipation or max collector current ratings ?

If its possible to exceed the max collector current rating without exceeding the max mw rating of the transistor then can the max Ic be exceeded without damage being done to the transistor ?

Why are pnp transistors better at sourcing a load then an npn is for example ?

 

[audioguru]

Max power dissipation is the max amount of heat that the transistor can produce without damage.
Max current is the max amount of current without damage.

If you exceed the max power dissipation but cool the transistor better than normal (with a fan or with liquid nitogen) then it is fine.

But if you exceed the max allowed current then it will probably burnout the connecting wires inside.

A PNP transistor sources current (its collector output goes high).
An NPN transistor sinks current (its collector output goes low).
But an NPN transistor used as an emitter-follower can source current but has a Vbe voltage loss.
A PNP transistor can also be used as an emitter-follower to sink current but has a Vbe voltage loss.

 

[tron87]

I don't fully understand if transistor a is dissipating heat well within its max limits then how can this destroy the internals of a transistor if the current on the collector is exceeding the limits if its the heat that does the damage ?


When using a transistor as a switch for a load what is generally a better solution to sink a load or source in respect to the power dissipation limits of the transistor ?

 

[ericgibbs]

I don't fully understand if transistor a is dissipating heat well within its max limits then how can this destroy the internals of a transistor if the current on the collector is exceeding the limits if its the heat that does the damage ?


When using a transistor as a switch for a load what is generally a better solution to sink a load or source in respect to the power dissipation limits of the transistor ?

Its as stated in agu's post.
But if you exceed the max allowed current then it will probably burnout the connecting wires inside.

 

[tron87]

So are power dissipation and damage though too much current flow two different things ? Doesnt the connected wires inside a transistor require heat energy to burn them out and if the power dissipated by the transistor is well within the limits but current isaint then how can this burn the wires out ?

 

[audioguru]

When a transistor is dissipating power then its entire chip heats up. It takes time to heat the mass.
But if the current is exceeded then only its connecting wires are heating up and they burn out in a very short time.

Why stress a transistor anyway? You gain nothing and lose everything:

1) A 2N3055 transistor has a max allowed collector current of 15A. But with a collector current of only 10A the base current needs to be 3.3A for it to pass the 10A poorly.

2) A 2N3904 transistor has a max allowed collector current of 200mA. But the datasheet doesn't show you its poor performance above only 100mA.

 

[Sig239]

Power dissipation takes voltage into account and refers to the "whole package", and the ability of the whole package to get rid of heat. The current spec is more specific to the wires inside the package and has nothing to do with voltage. You can think of the little wires connecting the die to the package leads like little fuses that will break if too much current runs through them.

Edit: Sorry, AG beat me to the punch and did a much better job of it to boot.

 

[tron87]

If i were using a transistor or a fet to switch (sink) a load that varied in resistance between basically a dead short or to a figure at the other of the spectrum how could i protect the tran or fet from passing too much current if there was a dead short across v+ and the collector or drain if i had no fuses ?

 

[crutschow]

You could add a current limiter circuit. If you use an NPN bipolar transistor with emitter to ground as the switch, then a resistor in series with the emitter to ground and two diodes in series from base to ground (after the base resistor with cathodes toward ground) will start to limit the current when the emitter current through the resistor generates a resistor voltage drop of about 0.65V.

 

[tron87]

Could you please add a picture of that circuit ?
What resistor values will i need to enable the current limiting circuit to start restricting the sinking current on a n channel fet when it reaches 4 amps at 12v ?

 

[crutschow]

Could you please add a picture of that circuit ?
What resistor values will i need to enable the current limiting circuit to start restricting the sinking current on a n channel fet when it reaches 4 amps at 12v ?

Below is a limiter that should work.

The 2N2222a bipolar transistor will start to turn on at about 0.6V and removing the FET gate drive, so the source sense resistor should be about 0.6V/4A = 0.15Ω. The simulation gave a 4A limit with Rs = 0.16Ω.