[SOLVED] Linear voltage regulator - current protection

[Eugen_E]

Hello,

I'm using a regulator with the schematic posted in the image attached, powered from a 30V source with 20000uF filter caps that can probably deliver more than 20A for short periods.

Unfortunately, the pass transistor blows instantly on short circuit. Q2 should limit the output current to 6-7A, but it seems it acts too slow, or the MOSFET needs a special type of protection like current foldback.
What kind of protection circuit would you recommend, so that the regulator can withstand accidental shortcircuit for 1-2 seconds?

Thanks

 

[andre_luis]

Search for Soft-start protection circuits.
At you case, maybe a simple diode+capacitor at R1 is enought, to limit current at time domain.

+++

 

[FvM]

Q2 should limit the output current to 6-7A, but it seems it acts too slow, or the MOSFET needs a special type of protection like current foldback.

The current limit of 6-7 A would most likely "blow" the output transistor even at nominal output voltage, with a short, it clearly violates the safe operation area, how perfect the heatsink may ever be.

It's simply a problem of much too low transistor power handling. As you said, a fold back current limit would increase the margin, but I think you should either reduce the current limit or increase the transistor power capability. Depending on the heatsink size, an overtemperature protection has to supplement it.

 

[Eugen_E]

Thank you for the replies.

In normal operation the output current is below 2 A.
The problem is the short circuit protection... I want it to survive a 1 second accidental short circuit, rarely, not continous. I used a 0.1 ohm resistor for current sensing because I had it, I will get a 0.22 ohm that should limit at 3 A.
I should mention that after the short, the heatsink is cold, it's just a problem of too high instantaneous current, and I don't understand why it's not limited.
Can you recommend a more suitable MOSFET in TO220 case? I tried IRFZ44, with the same result.

Thanks again

 

[FvM]

With 3A current limiting, the transistor should survive. Did you check, if the BC547 is still alive? It should better have a current limiting base resistor.

 

[svhb]

With 3A current limiting, the transistor should survive

30V * 3A = 90Watt during a short circuit! The transistor will survive this only if you can keep the case on 25 degrees Celcius (actually 107 Watt according to the datasheet (Fairchild IRF540A).

I think one-two seconds at this power level --> no way with a single TO220.

 

[FvM]

I think one-two seconds at this power level --> no way with a single TO220.

The transistor can survive, if the heatsink is still cold at the moment of short. Of course, this shouldn't be expected in regular operation. My point was, the transistor will be "reliably" blown when applying 7A@30V.

I agree, that the circuit should be designed to provide short circuit protection out of nominal loaded operation, with hot heatsink. Generally, a 3A three-terminal IC voltage regulator with thermal shutdown would be a much easier way to achieve unconditional safe operation.

 

[svhb]

FvM, indeed, it can survive, but I'm not putting any money on it.

Also to consider is how to define a short of about 1-2 seconds. This means you need a supervising system anyway that shuts off the system. Suppose it goes short, just at the moment you go for a coffee??

Stefaan

 

[FvM]

I completely agree. I was referring to the question, under which conditions the transistor will be blown immediately, even with a cold heatsink. In this comparision, 3 A versus 7 A current limit makes a difference. But it's far from safe operation.

 

[Eugen_E]

This is part of a programmable source controlled by a microcontroller, which measures the output current every 2ms.
A current limit is preset, and if overloaded it reduces the reference voltage for the regulator - which is obtained by filtering a PWM signal with an RC filter having a rise time of about 2ms.
So the uC current limitting should act after maximum 4 ms. The uC software is OK, because it maintains the preset current for normal loads, only the short is problematic.
The uC limiting was present all the time, but I thought it was too slow, so I wanted a "hardware protection".
It seems 4ms was too slow and the "hardware limit" of 7A too high - or the BC547 without base resistor was blown instantly... hope with 3A limit it will be OK...