DC short circuit: maximum allowed disconnection delay

[earckens]

What is the maximum allowed delay for disconnection of a DC power source (15V, up to 20A) in case of a short circuit?
The danger in this case would mainly be the risk of fire.

I read data suggesting ground fault protection for household appliances disconnection times are in the 300 ~400ms range.
But this is mains AC, my post concerns low voltage high amperage.

The reason for my question is the concept requirement for a circuit controlling 5 DC power sources.
My design uses a microcontroller reading current sense outputs from these sources. Reaction times when measured on a scope range in the 10 ~75ms ; if faster reactions are needed then I would have to change the concept to hardware detection: analog comparators directly driving mosfet's.

 

[Nigel Goodwin]

I wouldn't imagine there's any 'maximum allowed' time, and your reaction times are more than adequate.

More of a concern would be the actual disconnection of the load (if you use a relay DC arcs across the contacts considerable - hence the wide difference between AC and DC ratings) - if that's what you're wanting to do?. Is there some reason for not wanting a simple fuse?.

If you look at the blowing speed of fuses, you'll see it's not of much concern.

 

[Diver300]

There are a lot of factors that alter how fast a source has to disconnect. It depends what you are trying to protect, and how large the fault current could be.

Many DC power supplies are current limited, so there is no need for short circuit protection. If the supply is from a battery, then there may be a very large fault current.

If you are trying to stop wires from overheating, you can calculate how fast they will heat up for any fault current and that will allow you to work out the cut-off time needed.

 

[earckens]

....

More of a concern would be the actual disconnection of the load (if you use a relay DC arcs across the contacts considerable - hence the wide difference between AC and DC ratings) - if that's what you're wanting to do?. Is there some reason for not wanting a simple fuse?.

If you look at the blowing speed of fuses, you'll see it's not of much concern.

The load is resistive.
The reason for using a controller and not fuses is because
1. short circuits may be quite frequent
2. the controller also performs other tasks (displaying status etc..)
3. the use of a controller allows for auto resets and eventually manual reset (push button)

 

[Diver300]

What is the power supply? I think that more information is needed about what this thing does.

If your controller is cutting off in 75 ms, that may well be fine.

A short circuit may not be a complete short that reduces the voltage to zero, so the short itself may get hot. That can be very difficult to protect against as the current may not be increased by much.

 

[earckens]

The supplies deliver current sense. It is hence not output voltage but output current that is being measured.

 

[Diver300]

The supplies deliver current sense. It is hence not output voltage but output current that is being measured.

A short circuit is usually taken to be a fault that reduces the voltage to near zero volts. How much that will increase the current will depend on what the power supply is.

Are you going to tell us?

Some power supplies have "fold back current limiting", where they produce less current when shorted than into their normal load. I had problems with that were there was a 2.5 A power supply, and the circuits were fused at 2 A. A short circuit resulting in the power supply only producing around 0.5 A and the fuse didn't blow.

If you have a power supply like that, current measurement won't help, so please say what the power supply is, and what it is that you are trying to protect.

 

[Nigel Goodwin]

A short circuit is usually taken to be a fault that reduces the voltage to near zero volts. How much that will increase the current will depend on what the power supply is.

Are you going to tell us?

Some power supplies have "fold back current limiting", where they produce less current when shorted than into their normal load. I had problems with that were there was a 2.5 A power supply, and the circuits were fused at 2 A. A short circuit resulting in the power supply only producing around 0.5 A and the fuse didn't blow.

If you have a power supply like that, current measurement won't help, so please say what the power supply is, and what it is that you are trying to protect.

There was a similar issue with microwave ovens, the slightly strange half wave configuration meant that if the rectifier went S/C it tended not to blow the mains fuse - but the transformer still passed enough current to overheat and kill it. They overcame this in later years by fitting a 'protection diode' (like a super big zener) which also went S/C if the rectifier failed - this then blew the mains fuse.

Only problem was the protection diodes used to blow for no reason

 

[eTech]

What is the maximum allowed delay for disconnection of a DC power source (15V, up to 20A) in case of a short circuit?
The danger in this case would mainly be the risk of fire.

I read data suggesting ground fault protection for household appliances disconnection times are in the 300 ~400ms range.
But this is mains AC, my post concerns low voltage high amperage.

The reason for my question is the concept requirement for a circuit controlling 5 DC power sources.
My design uses a microcontroller reading current sense outputs from these sources. Reaction times when measured on a scope range in the 10 ~75ms ; if faster reactions are needed then I would have to change the concept to hardware detection: analog comparators directly driving mosfet's.

There is probably IEC standards for short circuit testing.

Google "IEC short circuit test"

 

[earckens]

There is probably IEC standards for short circuit testing.

Google "IEC short circuit test"

I Googled "IEC short circuit reaction time"; mostly very theoretical explanations, while I need one concrete figure.
One ABB site mentions 0.1s for fast breakers.
When searching for "ground fault reaction time", figures as low as 1/30 of a second appear (for residential application), yet no mention of legal maximum reaction time?

So be all means and measures, if I get reaction times for my design between 10 ~75ms (which by my latest software iteration yesterday I managed to get down to 5 ~25ms) I believe my design will not cause fires if overcurrent or short circuit occurs .. and yes, this is not for a microwave app nor for any power source that may give dubious output when stressed.

And if that is still not fast enough I can revert to a design with analog comparators and mosfet load switching. And still have microcontroller features.

 

[Diver300]

With mains power supplies, disconnection times have to be kept to a short time as there is a risk that someone could be electrocuted if the power is left connected for too long after a short, and times around 30 - 100 ms seem reasonable for that.

That limitation does not apply with a 15 V supply. It's simply a good idea to disconnect before things get burned.

Bench power supplies usually have adjustable current limits for that reason. If the current limit is set to a low value, it's very common for the wires to able to sustain a short circuit indefinitely.