Electronic circuit breaker design

[TheRedDevil18]

I'm building an electronic circuit breaker as follows,

I have some questions about the short circuit protection.

I tested it in simulation and it trips in about 11 ms. How would I work out the interrupt rating ?, which I read is the maximum short circuit current the breaker can interrupt, And also, How would I know if the transformer wiring won't burn ?, is their a way to calculate the wire temperature ?. I'm assuming 11 ms is a very short time for the temperature to rise and start burning the wires ?, do I need to work out the short circuit current of the transformer ?

And, in my design currently I have a peak detector circuit in which I am calculating the rms current using the peak voltage through the ADC. So far it works well in simulation, but I read in reality the AC sine wave is not perfect and that the true rms would be more accurate, though if I did a true rms then I would have to sample the entire waveform and then compare, so that would add 1/50 Hz or 20 ms to the trip time ?, making it slower

That's all for the questions for now

Any suggestions on improvements or what I'm doing wrong would be much appreciated, thanks

 

[anorlunda]

The interrupting capability depends on the ability to quench arcs between the contacts. That's not a thing readily calculated.

To visualize that, consider the time when the contacts are 1 micron apart. Then only one volt voltage difference means 1 million volts per meter gradient across the gap. A spark starts, ionizes the air and that creates a conductive path between the contacts.

If you are interrupting AC, the usual strategy is to trip at the moment of current zero crossing.p

The principles are illustrated in this video.

 

[Rive]

You would do better with a solid state relay which has its own zero cross switch feature.

Regarding the question about the current and heat-up time of wires: if the wire thicknesses are OK for the continuous maximal load and the tripping point is well set (so there is no chance of overcurrent for long time) then usually if the it's not a concern. The common way to solve such problems is to use a non-resettable thermal fuse close to the winding.

 

[Svein]

I tested it in simulation and it trips in about 11 ms.

Yes. You have used an electro-mechanical relay. Those are slow! I shouldn't wonder if the main part of the 11ms is the reaction time of the relay.

On another note. As @anorlunda pointed out, breaking a current is much harder than you suspect. Here is a link to some commercial ones: http://new.abb.com/low-voltage/products/circuit-breakers. BTW - do not be confused by the "low-voltage" label. Low voltage means "below 5000V"!

 

[TheRedDevil18]

The interrupting capability depends on the ability to quench arcs between the contacts. That's not a thing readily calculated.

To visualize that, consider the time when the contacts are 1 micron apart. Then only one volt voltage difference means 1 million volts per meter gradient across the gap. A spark starts, ionizes the air and that creates a conductive path between the contacts.

If you are interrupting AC, the usual strategy is to trip at the moment of current zero crossing.p

The principles are illustrated in this video.

Ok, so basically that depends on my relays current breaking capacity. I checked the datasheet and it says nothing about the contact breaking capacity,
Datasheet
http://www.megasan.com/service/pdfhandler.ashx?fileid=3385

How well would a snubber work in suppressing the arcs across the relay ?

 

[TheRedDevil18]

You would do better with a solid state relay which has its own zero cross switch feature.

Regarding the question about the current and heat-up time of wires: if the wire thicknesses are OK for the continuous maximal load and the tripping point is well set (so there is no chance of overcurrent for long time) then usually if the it's not a concern. The common way to solve such problems is to use a non-resettable thermal fuse close to the winding.

I wanted to use a solid state relay but they are really expensive. I will look into a termal fuse, thanks

 

[TheRedDevil18]

Yes. You have used an electro-mechanical relay. Those are slow! I shouldn't wonder if the main part of the 11ms is the reaction time of the relay.

On another note. As @anorlunda pointed out, breaking a current is much harder than you suspect. Here is a link to some commercial ones: http://new.abb.com/low-voltage/products/circuit-breakers. BTW - do not be confused by the "low-voltage" label. Low voltage means "below 5000V"!

I'm trying to use a triac and optocoupler design like this,

Though, it ends up being slower than the relay. Maybe because it waits for the zero crossing before tripping ?

 

[TheRedDevil18]

Also, would it be better to find the true rms instead of the peak ?

 

[anorlunda]

Also, would it be better to find the true rms instead of the peak ?

Can you back up a bit. Requirements first, design second. What are you trying to accomplish? Can you list your requirements?

 

[Svein]

Maybe because it waits for the zero crossing before tripping ?

Yes. Triacs functions like a latching relay.

 

[TheRedDevil18]

Can you back up a bit. Requirements first, design second. What are you trying to accomplish? Can you list your requirements?

I am looking to build a 10A circuit breaker using a microcontroller. It will have programmable current and trip time settings. It doesn't have much requirements besides what I just said.

From what I know, if I use a relay then their is going to be arcing across the gaps, so I tried using a triac as shown in the previous post but the trip times are slower than with the relay.

I also read that true rms calculation would be a better way to find the current. Although, I think it would make the circuit breaker slower since a whole wave needs to be sampled before it compares.

 

[Rive]

I also read that true rms calculation would be a better way to find the current.

You don't want to measure: you want to trip.
All you have to know is that/if in an application it'll be a fixed function between peak/limit current and (peak/limit) power.

BTW, I think you will need a small load in parallel with C4 (or to move R4 to the other side of the diode).

 

[anorlunda]

I am looking to build a 10A circuit breaker using a microcontroller. It will have programmable current and trip time settings. It doesn't have much requirements besides what I just said.

Here's the killer question. Does your breaker need to interrupt short circuit current? That is the more difficult requirement to meet. Without short circuit, we tend to call them just on/off switches, not circuit breakers. Compare to the light switch on the wall, backed up by a circuit breaker back in the service box.

So, is the 10A the max load or the short circuit current?

 

[TheRedDevil18]

All you have to know is that/if in an application it'll be a fixed function between peak/limit current and (peak/limit) power

I don't quite understand, can you maybe explain a little more please ?

So, is the 10A the max load or the short circuit current?

Its the max load current, and yes it must interrupt short circuit currents also

 

[anorlunda]

Its the max load current, and yes it must interrupt short circuit currents also

Then you need to know how much that short circuit current is. That depends on the wiring of the AC supply. But if the AC is supplied by a circuit in a building, then there is already a circuit breaker. Typical specs for a 15 Amp household breaker is 10000 AIC (Amps Interrupting Current). That's why the question is important, short circuit can be as much as 1000 times more than max load current.

The consequence of a short circuit which the breaker fails to quench is fire or explosion, so it is a safety related question. A commercial breaker rated 10000 AIC costs less than $4.

 

[TheRedDevil18]

Then you need to know how much that short circuit current is. That depends on the wiring of the AC supply. But if the AC is supplied by a circuit in a building, then there is already a circuit breaker. Typical specs for a 15 Amp household breaker is 10000 AIC (Amps Interrupting Current). That's why the question is important, short circuit can be as much as 1000 times more than max load current.

The consequence of a short circuit which the breaker fails to quench is fire or explosion, so it is a safety related question. A commercial breaker rated 10000 AIC costs less than $4.

View attachment 212158

So that would mean calculating the short circuit current of the transformer primary ?

Would a snubber work to quench the arcs across the relay contacts ?, I was also thinking of using an optocoupler and triac instead of the relay (since there will be no arcs) though it seemed to be a bit slower when tripping

 

[anorlunda]

So that would mean calculating the short circuit current of the transformer primary ?

Would a snubber work to quench the arcs across the relay contacts ?, I was also thinking of using an optocoupler and triac instead of the relay (since there will be no arcs) though it seemed to be a bit slower when tripping

What transformer? Not the CT in your circuit, but rather the short circuit current of V1 in your diagram.

A snubber might help, but it needs to be tested to be sure.

You're skating on thin ice here on a safety related question. What are your experience and skills related to electrical safety?

 

[TheRedDevil18]

What transformer? Not the CT in your circuit, but rather the short circuit current of V1 in your diagram.

A snubber might help, but it needs to be tested to be sure.

You're skating on thin ice here on a safety related question. What are your experience and skills related to electrical safety?

I am just a student studying engineering :)

This is one my projects that I have to do

 

[anorlunda]

I am just a student studying engineering :)

This is one my projects that I have to do

One of the things to learn is when to innovate and when to use existing codes and standards. For a circuit breaker, you should start with the codes and standards. Do you know how to look them up?

 

[TheRedDevil18]

One of the things to learn is when to innovate and when to use existing codes and standards. For a circuit breaker, you should start with the codes and standards. Do you know how to look them up?

Saw this circuit breaker which uses a microprocessor similar to my design
https://stevenengineering.com/Tech_Support/PDFs/45CBEIIC.pdf

On page 9 of the breaker above, they take a sample of the full cycle and calculate the true rms. That would mean for a 50 Hz wave, a time of at least 20 ms before it trips. If a short circuit occurs, the breaker should trip instantly with little time delay.

Unless, I think I'm misinterpreting how it works

 

[anorlunda]

So, what does your own research on capacitor energy density say?

Saw this circuit breaker which uses a microprocessor similar to my design
https://stevenengineering.com/Tech_Support/PDFs/45CBEIIC.pdf

On page 9 of the breaker above, they take a sample of the full cycle and calculate the true rms. That would mean for a 50 Hz wave, a time of at least 20 ms before it trips. If a short circuit occurs, the breaker should trip instantly with little time delay.

Unless, I think I'm misinterpreting how it works

You can try to trip earlier, but you might fail to quench the arc before the zero point anyhow. Remember, for a sin wave, the difference between peak and RMS is a fixed factor of $\sqrt 2$. If you see any instantaneous more than 15A on a 10A circuit, wny wait for true RMS, go ahead and trip.

But before posting again, explain how your breaker would avoid fire and explosion in case of a short circuit. It you can't make it safely, we won't help you.

 

[TheRedDevil18]

You can try to trip earlier, but you might fail to quench the arc before the zero point anyhow. Remember, for a sin wave, the difference between peak and RMS is a fixed factor of $\sqrt 2$. If you see any instantaneous more than 15A on a 10A circuit, wny wait for true RMS, go ahead and trip.

But before posting again, explain how your breaker would avoid fire and explosion in case of a short circuit. It you can't make it safely, we won't help you.

Ok, I saw that video you posted, and they say one of the solutions is to trip on a current zero. Since I have my current transformer, I can make a zero cross detector going to my MCU so I know when to trip ?, and maybe also a snubber across the terminals for added protection ?

They also say submerge the contacts in an oil solution, but do they actually have relays like that ?

 

[anorlunda]

but do they actually have relays like that ?

I think not, because relays are not intended to be circuit breakers. I was hoping that you would take the hint from post #19 to stop designing and start reading.

 

[TheRedDevil18]

I think not, because relays are not intended to be circuit breakers. I was hoping that you would take the hint from post #19 to stop designing and start reading.

I was reading about TVS diodes, would be easier to use one of those instead of a snubber circuit. How well would a tvs diode work ?