lm317 with a pass transistor?

[orangepeel]

hi everyone,
i want to build an adjustable power supply like the one shown here:
**broken link removed**
https://www.bristolwatch.com/ele/lm317.htm
i want to know how much current i can safely draw with one transistor (the MJE3055 is rated for 10 amps) and if this will effect the stability of the regulator.
thanks:smile:

 

[srizbf]

safe drawing current of that reg+pass transistor is based on :
the input-output differential voltage and max power dissipation of MJE3055.

 

[KlausST]

Hi,

the MJ3055 has low hfe, therefore i doubt you can draw 10A.
Power dissipation depends on input voltage.

The circuit is simple and has poor output voltage regulation.


Klaus

 

[Audioguru]

The excellent voltage regulation from the LM317 is ruined by the low gain transistor simply hanging on its output. If a transistor is found to deliver 10A from that horrible circuit then its output voltage will drop as much as 2.6V from the output voltage of the LM317.

 

[betwixt]

Use a PNP transistor (2N2955 or similar) and wire it as a bypass for the regulator. It only needs one extra resistor, keeps the regulation of the LM317 and also, if you are careful, extends the overload protection of the LM317 to the transistor as well.

Brian.

 

[orangepeel]

The excellent voltage regulation from the LM317 is ruined by the low gain transistor simply hanging on its output. If a transistor is found to deliver 10A from that horrible circuit then its output voltage will drop as much as 2.6V from the output voltage of the LM317.

so is there any way to fix this? should i use a different transistor?
thanks for your help.:thumbsup:

Hi,

the MJ3055 has low hfe, therefore i doubt you can draw 10A.
Power dissipation depends on input voltage.

The circuit is simple and has poor output voltage regulation.

Klaus

thanks for you help,
would a 2n3055 or a tip3055 (they are the same transistor) work better?

Use a PNP transistor (2N2955 or similar) and wire it as a bypass for the regulator. It only needs one extra resistor, keeps the regulation of the LM317 and also, if you are careful, extends the overload protection of the LM317 to the transistor as well.

Brian.

thanks for your help,
what is the value for the resistor if i want to have a 14.5-15v input and will a PNP transistor work better than an NPN?

 

[betwixt]

It isn't a case of whether NPN or PNP is better, what matters is the path the current flows through and where the heat is dissipated. In a linear regulator you can not avoid heating, that's physics I'm afraid.

In the present design, the current from the LM317 goes to the base of the NPN transistor which works as a current amplifier but at the same time isolates the output voltage from the mechanism that provides voltage regulation. That is hard wired inside the LM317 and anything after it is subject to variation as the load changes.

The better design requires the transistor to be wired with input to it's emitter, that's why it should be PNP so the polarity is kept correct. All you do is add a resistor in series with the 'IN' pin of the LM317 and wire the transistor emitter pin to the voltage source, base after the resistor (LM317 input) and collector to the LM317 output. So the resistor is across the B and E of the transistor and in line with the input to the LM317. The way it works is at low currents, the LM317 works as a 'stand alone' regulator, doing all the work itself. As you draw more current from it, more voltage is dropped across the input resistor, when that voltage reaches ~0.7V, the transistor starts to conduct and works as a bypass to the regulator. The great thing is that if you choose the resistor value carefully, the overload and thermal shutdown in the LM317 will also stop the transistor conducting so you don't compromise safety like the design you posted. Choose the resistor to drop 0.7V at about 50% of your anticipated maximum load current, that ensures around 75% of the current still goes through the LM317 and the remainder through the transistor. I suggest bolting both to the same heat sink so the themal shutdown work on heat generated by both devices.

If you really want to use an NPN bypass transistor, you can still do it but you need an additional stage of voltage inversion, look at the data sheets for the LM117/217/317 and you will see how it is done. Where you see the bypass transistor being an LM195, you can use an ordinary NPN power transistor instead.

Brian.