Very Low voltage Current Source.

Thread Starter


Joined Nov 22, 2014
I am building an adjustable bench P.S.U. for home use (I could buy one, but this way is more fun).
The final regulator in my P.S.U. is an LTC3080.
The regulator requires 0.5mA minimum load to work correctly.
So I thought of using a current source to provide that minimum load.

The obvious choice is a shunt current source connected across the output.
But I have hit a limitation in the fact that these current sources need a minimum supply voltage to operate.
The best I could find is a LM334Z and it requires a minimum voltage of 0.8V to operate correctly.

I want my P.S.U. to be able to regulate correctly below that 0.8V, down-to few mA, so the output voltage does not float, when there is no load.
I have not been able to come-up with a solution to this problem.
I may have to live with it! But from an engineering point of view I would like to know how to solve this problem.




Joined Mar 2, 2015
It's an LT3080.

I had the same problem with that chip and its minimum load current requirement. I solved it by making an auxiliary negative supply to operate the current sink from; as I recall, I used an LTC1044 to create the negative voltage. Problem solved.


Joined Mar 14, 2008
Just connect an NPN transistor collector (2N3904 or just about any small transistor with a voltage rating higher than you maximum output voltage) to the regulator output with the emitter to ground.
Connect the base through a large resistor to the unregulated input voltage.
That will give a reasonably good current source down to less than a 100mV.
You will have to experimentally determine the size of the resistor to give you a mA or so through the transistor (or adjust it to give 100mV or so at the output with the regulator set at 0V with no load).

Depending upon the gain of the transistor it will likely need a base resistor value to give a few μA of base current.
The more base current, the lower the minimum output voltage, but that will give a higher constant current at the higher output voltages.
The limit would be when the dissipation of the transistor exceeds its rating at the highest output voltage.
Last edited:


Joined Mar 14, 2008
Here's an LTspice simulation of my suggested circuit using a BJT as a constant current sink.
The minimum no load output voltage is ≈50mV and the maximum transistor current is about 4.3mA for the base resistor shown and a transistor with nominal gain, and a 25V input voltage.



Last edited: