Inrush current limiter
I designed this circuit in order to limit the
current draw by vacuum tube heaters. When the heater is cold it has a
lower resistance and the inrush current can be three to four times
bigger. This is stated to be a major cause of tubes failure.
The idea is simple, an n-channel mosfet
is used as variable resistor. The C1 capacitor is slowly charged
through R2 resistor. So the transistor will start to turn on. There is
no reason to wait until the voltage across C1 reach the gate threshold.
To avoid this delay the capacitor is charged very fast just below the
threshold voltage. This voltage is set by R3. A fixed 2.6V reference is
created with a blue led, D1. This is more than enough since the
transistor is a logic level mosfet that turns on at around 2v
R1
= 1kΩ, R2 = 1MΩ, R3
= 10kΩ, D1
= blue LED, D2 = SB160, D3 = 1N4007, C1 = 150μF(tantalum), Q = IRLR024NPBF
If the input voltage is
turned off then C1 will remain charged so the limiting function will be
lost for a while. This is the reason
why D3 was added, to create a discharge path through the load.
Another thing I noticed is that an
aluminum electrolytic capacitor will not work for C1. The leakage
current is large enough to prevent it to fully charge. Because R2 is
also very large, since I want to have a few seconds of limiting action.
So C1 is a tantalum capacitor.
Sometimes it might be desirable to put the mosfet
across the positive rail, in this case it should be a p-channel type.
Bellow there are two examples of a 500mA heater
from a 6AC7 with and without current limiter. Notice the large initial
current spike in the second picture:
A more precise current limiting could
be obtained with a LM317 and a single resistor, but it will add a
voltage drop. The circuit above has the advantage that after the
initial limiting phase it cause only a very small loss of voltage
thanks to the low Rds of the mosfet. Not to mention that this low Rds
can be made even lower by adding in parallel as may
transistors you like :).Extra gate capacity is not an issue
here.
Also it should be considered the
voltage threshold dependance on temperature in case significant power
is disipated during current limiting.