Protection from Thunder Lightning

[alok1982]

Hi All!

I want to protect my power supply circuit from thunder lighting.
This power supply is based on flyback topology.
I had installed it near to main gate and mostly fails because of lightning.

One more power supply is there where I am using step-down transformer which converts 240V/50Hz to 28V. Same problem happens with this circuit.
This power I am using for emergency bell.

Can I use MOV's or Fuses at input? Please guide with some schematic.


Thanks!

 

[Cicero]

I'd probably put 3 MOVs in, one on each line to Earth, and one between the lines. Then add in some sort of fusing in line to protect from sustained overvoltage/overcurrent conditions, like a PTC or regular slow blow fuse.

This type of setup https://diyaudioheaven.wordpress.com/tutorials/power-supplies/mains-filters/

On the DC side, I'd also take extra precautions and add in Tranzorbs and polyfuses.

 

[dknguyen]

I'd probably put 3 MOVs in, one on each line to Earth, and one between the lines. Then add in some sort of fusing in line to protect from sustained overvoltage/overcurrent conditions, like a PTC or regular slow blow fuse.

This type of setup https://diyaudioheaven.wordpress.com/tutorials/power-supplies/mains-filters/

On the DC side, I'd also take extra precautions and add in Tranzorbs and polyfuses.

On the DC side you can use unidirectional TVS diodes. On the AC side you can use bidirectional TVS diodes. TVS diodes different from MOVs in that they don't have an inherent wear mechanism (limited number of cycles) and can clamp to lower voltages, however TVS diodes also generally have lower power capability.

You just places these (TVS diodes and MOVs) between the line you want protected and ground. Obviously for the unidirectional TVS diode you place it in the direction that doesn't cause a permanent short-circuit.

 

[Cicero]

On the DC side you can use unidirectional TVS diodes. On the AC side you can use bidirectional TVS diodes. TVS diodes different from MOVs in that they don't have an inherent wear mechanism (limited number of cycles) and can clamp to lower voltages, however TVS diodes also generally have lower power capability.

You just places these (TVS diodes and MOVs) between the line you want protected and ground. Obviously for the unidirectional TVS diode you place it in the direction that doesn't cause a permanent short-circuit.

Hmm, I've heard of people using TVS diodes on the hot AC side, but I've always been a bit sceptical.

I'm not sure how they'd hold up under official Immunity testing scenarios? Surges etc

Also, are those specific TVS diodes UL approved, whats their failure state (S/C or O/C)?

 

[dknguyen]

As semiconductors, they usually fail short.
https://www.vishay.com/docs/88440/failurem.pdf

Some are UL rated and some are not.

I think TranzOrb is just a brand name for a TVS diode.

What's the basis of your skepticism about bidirectional TVS on the hot AC?

Hmmmm...apparently TVS diodes turn on faster than MOVs too.

 

[Cicero]

As semiconductors, they usually fail short.
https://www.vishay.com/docs/88440/failurem.pdf

Some are UL rated and some are not.

I think TranzOrb is just a brand name for a TVS diode.

What's the basis of your skepticism about bidirectional TVS on the hot AC?

Hmmmm...apparently TVS diodes turn on faster than MOVs too.

Yes correct, for some reason over the years I've got into the bad habit of referring to all TVS diodes as Tranzorbs.

My scepticism just stems from my experience with EMC testing, TVS diodes are much quicker but can't handle the power that MOV's can, so I'd worry when testing surges (IEC_61000-4-5) that they'd blow and fail my test. For safety with a commercial product you cannot mess around with the AC side of things, UL/CSA specifications are stringent (look at X and Y caps for example, you cannot just use any cap with a high voltage rating due to their failure states) and a TVS diode that fails S/C poses a shock risk. I also haven't seen many TVS diodes rated for mains AC use available, with UL approval.

I could be completely wrong, wouldn't be the first time.

I've only ever used TVS diodes on the DC side, or low voltage AC side (after a step down transformer) for example, often paired with resettable fuses which limit both voltage and current, and make sure the TVS power dissipation is always below its max allowance.

 

[dknguyen]

Yes correct, for some reason over the years I've got into the bad habit of referring to all TVS diodes as Tranzorbs.

My skepticism just stems from my experience with EMC testing, TVS diodes are much quicker but can't handle the power that MOV's can, so I'd worry when testing surges (IEC_61000-4-5) that they'd blow and fail my test. For safety with a commercial product you cannot mess around with the AC side of things, UL/CSA specifications are stringent (look at X and Y caps for example, you cannot just use any cap with a high voltage rating due to their failure states) and a TVS diode that fails S/C poses a shock risk. I also haven't seen many TVS diodes rated for mains AC use available, with UL approval.

I could be completely wrong, wouldn't be the first time.

I've only ever used TVS diodes on the DC side, or low voltage AC side (after a step down transformer) for example, often paired with resettable fuses which limit both voltage and current, and make sure the TVS power dissipation is always below its max allowance.

That's fair. I mean, their main weakness is power handling capability. They hold advantages in pretty much all other areas except that one.

I thought failing short was desirable though? It protects the device, lets the user know it's no longer working, and trips the breaker which should prevent any shock hazard, no? Whereas failing open just removes the protection and no one is the wiser.

 

[DerStrom8]

A typical lightning protection circuit includes the following parts:

• Tripolar gas discharge tubes
  These should be placed across the supply lines closest to where the "outside world" comes in, with the middle terminal connected to earth ground
• Inductor or other surge protection
  This should be placed on each line just following the gas discharge tube. Inductors (~10uH) are one option. An alternative would be the TBU series of surge suppressors from Bourns, Inc.
• TVS (transient voltage suppression) diodes
  These will be located as close to your circuit as possible. Make sure their clamping voltage is below the rating of your circuit's components, but make sure its breakdown voltage is high enough that it will not start conducting during normal operation.

Below is a rough MSPaint sketch:



That's about as simple as you can get while maintaining reliability.

 

[dknguyen]

A typical lightning protection circuit includes the following parts:

• Tripolar gas discharge tubes
  These should be placed across the supply lines closest to where the "outside world" comes in, with the middle terminal connected to earth ground
• Inductor or other surge protection
  This should be placed on each line just following the gas discharge tube. Inductors (~10uH) are one option. An alternative would be the TBU series of surge suppressors from Bourns, Inc.
• TVS (transient voltage suppression) diodes
  These will be located as close to your circuit as possible. Make sure their clamping voltage is below the rating of your circuit's components, but make sure its breakdown voltage is high enough that it will not start conducting during normal operation.

Below is a rough MSPaint sketch:

View attachment 111631

That's about as simple as you can get while maintaining reliability.

My god DGT are small. I never they were available in such small sizes. I always envisioned something the size of a vacuum tube.

 

[AnalogKid]

Hmm, I've heard of people using TVS diodes on the hot AC side, but I've always been a bit sceptical.

Oh count rare ...

Tranzorbs are superior to MOV's in just about every way except cost. An AC tranzorb is two high voltage zener diodes in series back-to-back. They have a way faster response time. AND (the biggie), they do not degrade nearly as much with each transient hit. In round numbers, an MOV's tranisent performance decreases about 10% with each big hit. The clamping voltage increases, the response time increases, the clamped impedance increases, and the "knee" between high and low impedance gets softer. 4 strikes. At work (small and medium-sized MIL electronic systems, air, land, sea, AC, DC, whatever) and at home, I use nothing but.

Cons: stiffer leads, more careful handling when bending the leads to prevent over-stressing the internal diode junctions, cost.

ak

 

[dknguyen]

Oh count rare ...

Took me a few reads out loud to figure out what you were trying to say. "Au contraire", errrr I think. Whatever "on the contrary" is in French

 

[alok1982]

Thanks for your replies!

Can you further help me what value of MOV is suitable for my application? Same with TVS and Fuse?
Input voltage is 240Vac/50Hz.
For another circuit I am using Step down transformer. Input is 240Vac/50Hz and output is 28V, which I use for battery charging circuit.

Thanks!

 

[Cicero]

Oh count rare ...

Tranzorbs are superior to MOV's in just about every way except cost. An AC tranzorb is two high voltage zener diodes in series back-to-back. They have a way faster response time. AND (the biggie), they do not degrade nearly as much with each transient hit. In round numbers, an MOV's tranisent performance decreases about 10% with each big hit. The clamping voltage increases, the response time increases, the clamped impedance increases, and the "knee" between high and low impedance gets softer. 4 strikes. At work (small and medium-sized MIL electronic systems, air, land, sea, AC, DC, whatever) and at home, I use nothing but.

Cons: stiffer leads, more careful handling when bending the leads to prevent over-stressing the internal diode junctions, cost.

ak

Interesting ak, I'll look into them a bit more then. Sometimes you just get into a habit of doing something one way over the years.

How do they fair during your IEC_61000-4-5 surge tests?

What topology? Both lines to earth, and line-to-line? What other components do you have around there as well?

 

[AnalogKid]

Standard triangle: L-N, L-G, N-G. MIL 704, 1275, and 1399, no problems.

MIL-704 and MIL 461, 28 Vdc, input capacitor ripple current was a much bigger headache.

ak

 

[alok1982]

Hi,
Can you please suggest any part number which suits best for my application?

Thanks!