LM317 current boost circuit variations

Not open for further replies.

Thread Starter


Joined Aug 1, 2013
Here are two variations of a current booster circuit for an LM317 regulator. Note that all filter/compensation/transient capacitors are intentionally deleted for clarity.

A: In another thread I proposed a current booster circuit for an LM317 that is different from the standard circuit on the datasheet. Mine uses an NPN power transistor outside the 317 feedback loop. Advantages are that NPN power transistors are cheaper and better performing than their PNP "counterparts", there is no PNP driver needed as with a Sziklai pair, etc. Disadvantages are increased headroom, and the NPN power emitter follower is outside the 317 control loop so the output will sag a little bit as the output current increases due to increasing Vbe. For the application in that thread it is quick and dirty and meets all of the requirements. Thinking about the output sag led to B.

B: This circuit has the emitter follower inside the 317 control loop. In theory this circuit should compensate for the existence of and changes in Vbe. I have not searched the innergoogle to see if anyone has thought of this already.

So have at it: Is there any reason circuit B will not work, or will be only marginally stable, or or or.




Joined Mar 14, 2008
At first glance it looks like it should work but it won't for any 3-terminal regulator.
The regulator adjusts Vout to be about 1.2V above Vadj.
Thus any output resistance or other voltage drop, such as from the transistor will actually make the output regulation worse, not better.
For example, if the transistor Vbe equaled 1.2V then the emitter output voltage would be essentially 0V to maintain Vout-Vadj at 1.2V.
That's why you always put R4 as close to the regulator output as possible, even if that seems counter-intuitive.

See simulation below:



Joined Feb 17, 2009
Version B is worse then version A. In version B LM317 want to keep Vout - VADJ = 1.25V constants. So any change in Vbe will now change the voltage drop across R4 and this will change the R4 and R5 current. And because of this change in current and in Vbe the Vout will also change. And that is why version B is worse than A. In version A Vout changes only due to Vbe is changing.
Not open for further replies.