Adding a current limiting feature to a LM317HVT based regulated power supply

[vr6]

Hi,
some time ago I've built a power supply based on 3x LM317HVT IC's in parallel (circuit in the attachment) and now I'd like to add current limiting to it. I was looking for a circuit that would limit the output current to 3A (max 3.5A) but I couldn't find anything that would do that at any given voltage (1.25V - 52V). I'm using a transformer with 44V AC secondary. Can anyone help me with building such circuit?
Thanks in advance

[Zero999]

Connecting regulators in parallel like that, can cause instability. It will also need a minimum load current of between 5mA to 30mA. The data sheet shows the correct way to boost the current, by using additional pass transistors or multiple devices in parallel with current sharing resistors and an op-amp.
http://www.ti.com/lit/ds/symlink/lm117hv.pdf

Using two LM317HVs, rather than three, would make the current limit closer to 3A.

[iMo]

You have to add small resistors (like 0.1-0.3ohm) into output of each LM317. Otherwise the regulators will fight each other (each regulator has got a little bit different output voltage between out and adj pins).

The problem with 3xLM317s is that even you regulate the output voltage down to 1.25V upon the short, the output current limit will still be 3x2.2A (datasheet).

It seems the only way to limit the current will be an external circuit placed before the 3xLM317s input.

Below an example (simulation only). Use at your own risk.

[vr6]

Thanks for suggestions I guess I'll add those resistors and test that circuit.

[iMo]

With TIP127.
The current limit is set by the R1 (0.22ohm for aprox 3A) resistor.
Use at your own risk.

[David Hess]

The LM317 is not suited to parallel operation without external help to enforce current sharing and adding ballast resistors to the output compromises load regulation.

Imo has the right idea for current limiting in this case; sense the current at the input and if it is above a threshold, pull the adjust pin low.  But without a negative bias supply, this can only current limit down to 1.25 volts output at which point the regulators will provide full current.

Update, there is a way to get it all with a single transistor:

If the input is floating (do not ground the input side components), then implement the current sensing on the negative side after the input filtering and before the adjustment pin connection with a single NPN transistor 2 or 3 x Vbe current limiter.  Now when the current limit activates, the emitter is 2 or 3 x Vbe negative with respect to ground so the transistor can pull the adjustment pin below ground reducing the output to around zero volts or even negative if 3 x Vbe is used.  The extra Vbe drops are provided by diodes.

A series RC network between ground and the adjustment pin may be needed for stability during current limiting.

Do not forget to put a low value resistor in series with the base to prevent the surge current from blowing the transistor's base-emitter junction as shown in Ian's example.

[iMo]

I think you cannot limit the current of a LM317 by manipulating the adj pin (pulling it to gnd or to negative). That would be too easy

The LM317 always keeps 1.25V between out and adj, and its current limit is always 2.2A regardless how you mess with adj.

Therefore my above idea with current limiting externally _before_ you enter the 3xLM317 regulator.

[Zero999]

I think you cannot limit the current of a LM317 by manipulating the adj pin (pulling it to gnd or to negative). That would be too easy

The LM317 always keeps 1.25V between out and adj, and its current limit is always 2.2A regardless how you mess with adj.

If the input voltage is always higher than 15V it would, as the safe operating area protection will kick-in, causing the current limit to drop.

You could either monitor the negative rail using a transistor, or put an op-amp on one of the current sharing resistors.

[iMo]

Here is an NPN version of the Current limiter, combined with a Capacitance Multiplier.

[David Hess]

I think you cannot limit the current of a LM317 by manipulating the adj pin (pulling it to gnd or to negative). That would be too easy

The LM317 always keeps 1.25V between out and adj, and its current limit is always 2.2A regardless how you mess with adj.

If the adjust pin is -1.25 volts so the output is 0.00 volts, very few loads are going to draw much current.  In practice the circuit I described will pull the output low enough to regulate the output current.  This might be difficult if the output impedance was close to zero but the current shunt resistance adds to the output resistance when current limiting.

Emitter follower and class-ab outputs have no trouble regulating current with suitable control circuits.  If the LM317 had better tolerance, then like the LT1083 it could be used in parallel with an acceptably low ballast resistor.

[Zero999]

But the LM317 will have a much lower current limit, when the input-output differential is high. Pull the adjust pin down to near 0V and the current will be limited to a very low level, unless the input  voltage is also high. In this case the original poster is using a transformer with a 44VAC secondary, which would give an output voltage of 62V peak and is a little over the maximum voltage rating of the LM317HV, although it will drop under load and is unlikely to damage it.

[David Hess]

But the LM317 will have a much lower current limit, when the input-output differential is high. Pull the adjust pin down to near 0V and the current will be limited to a very low level, unless the input  voltage is also high. In this case the original poster is using a transformer with a 44VAC secondary, which would give an output voltage of 62V peak and is a little over the maximum voltage rating of the LM317HV, although it will drop under load and is unlikely to damage it.

That is a different problem that I have not addressed; there are several ways to add high voltage capability to a 3 terminal regulator.  I would not rely on high differential voltage for current limiting.

[David Hess]

I was looking for something else and found two applicable application circuits in the old LM317 datasheets.

The first one shows exactly what I described with a 1 x Vbe NPN current limit for a floating input.  I would add two diodes in series with the emitter to make it 3 x Vbe.

The second and third show two ways to safely parallel regulators.

[not1xor1]

Hi,
some time ago I've built a power supply based on 3x LM317HVT IC's in parallel (circuit in the attachment) and now I'd like to add current limiting to it. I was looking for a circuit that would limit the output current to 3A (max 3.5A) but I couldn't find anything that would do that at any given voltage (1.25V - 52V). I'm using a transformer with 44V AC secondary. Can anyone help me with building such circuit?
Thanks in advance

Apart the other valid advices, I think nobody spotted the 1kΩ limited LED in the circuit... 

That's about 35mA in worst case (no load and a bit of AC overvoltage) (*) through a LED that I guess is just as an ON indicator.
Make that 10-22kΩ. Modern LEDs are bright and in any case (vintage or modern) would last much longer when fed with moderate current.

And if you want to get 5A out you'd better double the input capacitor. Make that 2 x 4700µF.
OK... now that I've read more carefully that is 3-3.5A... so make that 2x3300µF 

(*) the circuit diagram specifies 24V, but in the text you write 44V (so it would be even worse)...