Surges and electronics

[paulengr]

Doesn't good ground have most effectiveness for a lightning surge?

Transients from switching aren't necessarily seeking ground as much as another system conductor I would think, though one system conductor is usually also at/near ground potential.

Switching transients are mostly single phase. In MCCs they occur when power is removed from the coils on large contactors. The path they seek is to the neutral back to the coil such as via a nearby PLC input transistor base to drain/emitter or across a diode or SCR. On transmission lines and distribution systems it’s due to overvoltage that occur due to capacitance in the switching mechanism discharged by the arc as the switch opens. In transmission lines they have a resistor in parallel to stop this, if they don’t fail. This is seeking ground.

Surge arresters can be wired phase to phase or phase to ground. The MCOV can be lower on phase to ground. Like 300 V vs 500 V for a 480 V grounded wye system.

 

[paulengr]

A surge suppressor is a blob of semiconductors that doesn't do anything till voltage gets high enough to make it conduct, I don't see where a connection to earth helps it do it's job in any way.

Simply not true especially today.

No surge arresters turn surges into heat. They are all based on nonlinear resistance but they are shunt, not series devices.

One type is the spark gap. It is two plates separated by a calibrated distance. When the voltage exceeds the flashover voltage, they arc (short circuit).

A similar device is a gas discharge tube. Same principle but uses gasses such as neon for more control. These can have very low (telecom) trigger voltages.

Another type are SiC diodes. Invented in the 1930s. High current and voltage, used in series pairs for AC.’Diodes are in opposite polarities. Below the knee/avalanche voltage they are open conductors. Above that voltage they become almost dead shorts. They don’t switch very fast so they are combined with surge capacitors that temporarily absorb steep wavefronts.

But by far the most common today is the metal oxide varistor. This is a block of sintered metal oxides, mostly zinc oxide. It is a voltage variable resistor. Unlike the others it does not “switch on” but rapidly changes resistance as voltage increases. There is almost no “delay” so it works better without need for capacitors. Two downsides. First is there is not a specific “trigger” voltage because it is not a “trigger”. The resistance simply exponentially decreases at a certain point. It is more broad. Second issue is a little metal oxide is consumed every time it gets above its threshold voltage and they tend to fail shorted instead of open. MOVs were first introduced in the 1970s. The technology has slowly improved and taken over the market,

No intentional heat conversion at all. But infrared cameras are used to find bad ones. A utility surge arrester has some heat all the time. If it is hotter than the others it might be starting to fail. If it is colder it blew its guts out and already failed. Utilities usually use MOVs. Basically bigger versions of a “power strip”. Older ones were SiC or gapped SiC (spark gaps mixed with diodes).

This is generally true of all surge arresters. Bigger surges shorten arrester life. Small “power strip” models last a few years. By UL they have a series fuse and most have a “arrester good” light although some are just “fuse blown” lights that don’t detect open circuit surge arresters.

Large pole mounted ones last sometimes decades. Bigger (fatter, not higher voltage) is better. There are ways to test whether they are functioning (Megger it) but no way to tell how much is “used up”. On utility grade surge arresters visual inspection often catches most failures.

Fuses and PTCs (polymer resettable fuses) work based on electrochemical conversion to heat. Surge arresters do not work that way at all.

 

[retirede]

Simply not true especially today.

No surge arresters turn surges into heat. They are all based on nonlinear resistance but they are shunt, not series devices.

One type is the spark gap. It is two plates separated by a calibrated distance. When the voltage exceeds the flashover voltage, they arc (short circuit).

A similar device is a gas discharge tube. Same principle but uses gasses such as neon for more control. These can have very low (telecom) trigger voltages.

Another type are SiC diodes. Invented in the 1930s. High current and voltage, used in series pairs for AC.’Diodes are in opposite polarities. Below the knee/avalanche voltage they are open conductors. Above that voltage they become almost dead shorts. They don’t switch very fast so they are combined with surge capacitors that temporarily absorb steep wavefronts.

But by far the most common today is the metal oxide varistor. This is a block of sintered metal oxides, mostly zinc oxide. It is a voltage variable resistor. Unlike the others it does not “switch on” but rapidly changes resistance as voltage increases. There is almost no “delay” so it works better without need for capacitors. Two downsides. First is there is not a specific “trigger” voltage because it is not a “trigger”. The resistance simply exponentially decreases at a certain point. It is more broad. Second issue is a little metal oxide is consumed every time it gets above its threshold voltage and they tend to fail shorted instead of open. MOVs were first introduced in the 1970s. The technology has slowly improved and taken over the market,

No intentional heat conversion at all. But infrared cameras are used to find bad ones. A utility surge arrester has some heat all the time. If it is hotter than the others it might be starting to fail. If it is colder it blew its guts out and already failed. Utilities usually use MOVs. Basically bigger versions of a “power strip”. Older ones were SiC or gapped SiC (spark gaps mixed with diodes).

This is generally true of all surge arresters. Bigger surges shorten arrester life. Small “power strip” models last a few years. By UL they have a series fuse and most have a “arrester good” light although some are just “fuse blown” lights that don’t detect open circuit surge arresters.

Large pole mounted ones last sometimes decades. Bigger (fatter, not higher voltage) is better. There are ways to test whether they are functioning (Megger it) but no way to tell how much is “used up”. On utility grade surge arresters visual inspection often catches most failures.

Fuses and PTCs (polymer resettable fuses) work based on electrochemical conversion to heat. Surge arresters do not work that way at all.

So when an MOV experiences a surge, its resistance drops, and current flows through it in an attempt to clip the voltage surge.
There is current flowing through the device and voltage across it. How is that energy dissipated if not by heat?

 

[paulengr]

Surge protectors work by turning electrical energy into heat. The surge protectors in the panel reduce the energy in the surge, leaving a much smaller surge for the point of use protectors to deal with. Surge protectors installed in the panel are NOT useless. To provide good protection both panel and point of use protectors should be used.

In combination I would agree. The marketing is “whole house” implying that one little $100 surge arrester protects everything, with or without point of use protection. If you are at all concerned though skip the whole house arrester unless you put in a substantial one.

But there are two glaring issues. The first one is line length. Surge arresters are NOT the same as say fuses. It does not protect “downstream devices”. An arrester is wired in shunt or parallel to the devices. It is in series with the leads going to the load. So the voltage as seen by the load is the voltage across the surge arrester PLUS the voltage across the leads.

As approximate formula for the lead voltage is:
Vlead, kV = 0.34 (microhenries/foot) x length (feet) x di/dt x 2 where di is the surge crest current in kA and dt is the time to crest in microseconds.

So say we trip a breaker with a 25 kVA transformer with a 2%Z. That’s 25 / 240
/ 0.02 or 5.2 kA short circuit current and the surge will be maybe 1.5 times that by utility estimates or 7.8 kV. Switching surges will be 100 KHz or more or about 10 microseconds. So if we are 10 feet from the main panel that’s 20 feet of length (count hot and neutral). So we get:
0.34 x 20 x 7.8/10 x 2 = 10.68 kV or 680 V. We have to then add this to the MCOV of the panel surge arrester at a common rating of say 175 V so final surge is around 10-11 kV. If the surge arrester is right next to the load so we have say just 1 foot away to a wall wart, then by the same math lead surge voltage is only 0.53 kV and we are at an easily survivable 705 V, 530 for the cable and 175 from the cheap power strip.

If the point was to take the surge possible at the panel from the utility transformer (tens of kV) down to 10 kV for the power strip then to 700 V for the load, mission accomplished. Without the panel arrester the power strip does the heavy lifting and has a shorter life. But going the other way, I stand by my statement that the panel surge arrester is worthless.

Second beef with panel surge arresters is that the name brand ones are little better than power strips. You can save substantial money putting a receptacle on the side of the panel and just plugging a power strip into it! It just looks impressive.

Take those components apart and physically compare it to what you find taking apart a power strip. Then compare those to say a MacLean Z-Force or Surge Tec surge arrester.

Surge Tec Secondary Arresters > MacLean Power Systems

www.macleanpower.com

This is what utilities put out on the transformer. You should consider why they use one of these instead of that cute thing with lights and a device the size of the quarter that it costs. Just saying...big difference.

I have no affiliations with MacLean. I buy their products through supply houses. It’s just a brand that I know pretty well. Cooper and GE products aren’t as compact for panel installs. The SOLA stuff isn’t bad either. If the actual element doesn’t look like a “J” fuse it’s not worth much.

To find the little $0.25 MOV which is the same one you find in the power strips and in the Square D and AB products you need to look on the electronics web sites (Mouser, Digikey, Allied, Newark). It’s fine for point of use but not much else. Size does matter. The fatter the better.

Hence my beef with most of the whole house surge arresters on the market is the way they are sold. They are basically distribution equipment right at the demarcation line between feeders and service entrance. Just as the system bonding jumper and ground rods are important safety equipment to both feeders and service entrance, so is the panel surge arrester if one is used. You wouldn’t use a #16 ground wire (Code or not), why use a dinky one time use surge arrester? And if surges are actually important why protect the one part of the system that has a better chance of survival (higher BIL) when the loads fail all the time as another poster stated? And if you look at the math it’s obvious why they fail if surges are the culprits.

 

[ActionDave]

...But by far the most common today is the metal oxide varistor. This is a block of sintered metal oxides, mostly zinc oxide. It is a voltage variable resistor. Unlike the others it does not “switch on” but rapidly changes resistance as voltage increases. There is almost no “delay” so it works better without need for capacitors. Two downsides. First is there is not a specific “trigger” voltage because it is not a “trigger”. The resistance simply exponentially decreases at a certain point. It is more broad. Second issue is a little metal oxide is consumed every time it gets above its threshold voltage and they tend to fail shorted instead of open. MOVs were first introduced in the 1970s. The technology has slowly improved and taken over the market,

So like I said in not so many words,,,,,"a blob of semiconductors that doesn't do anything till voltage gets high enough to make it conduct,"

Nothing about a connection to earth makes a bit of difference on how these things function.

 

[kwired]

Second beef with panel surge arresters is that the name brand ones are little better than power strips. You can save substantial money putting a receptacle on the side of the panel and just plugging a power strip into it! It just looks impressive.

2020 NEC requires surge protection at the service or first panel in dwellings.

Though I agree that receptacle and power strip next to panel is probably just as effective in many cases - how many of those will pass inspections?

Why they add this requirement, IDK but kind of wonder if all those AFCI's required maybe need some surge protection to possibly make them last longer?

 

[Coppersmith]

I live in the "lightning capital of the USA". In one of the counties I work in, surge protectors have been required in panels by local rule for quite some time. This does not make it any easier to sell whole house surge protectors in the other counties. Most people think you are trying to sell them snake oil.

 

[kwired]

I live in the "lightning capital of the USA". In one of the counties I work in, surge protectors have been required in panels by local rule for quite some time. This does not make it any easier to sell whole house surge protectors in the other counties. Most people think you are trying to sell them snake oil.

Probably partly because they see damage in direct hits whether there is protector or not. What they don't realize is how much surge damage they may be experiencing from nearby strikes. It takes time for it to lead to total failure, but having that protector can lessen the unseen damage.

Still best idea to have the protector at the service/main panel and then supplemental protectors at equipment.

 

[GoldDigger]

So when an MOV experiences a surge, its resistance drops, and current flows through it in an attempt to clip the voltage surge.
There is current flowing through the device and voltage across it. How is that energy dissipated if not by heat?

I think that the design is to conduct at a low enough resistance that the impedance (including reactance) of the branch and feeder wires is high enough to dissipate most of the heat. I agree that there will still be some heat dissipation in the shunt device, but not necessarily corresponding to full line voltage.
I think that the limited amount of applied surge energy before replacement is needed is indeed an indication of thermal damage to the device with each event.

 

[retirede]

I think that the design is to conduct at a low enough resistance that the impedance (including reactance) of the branch and feeder wires is high enough to dissipate most of the heat. I agree that there will still be some heat dissipation in the shunt device, but not necessarily corresponding to full line voltage.
I think that the limited amount of applied surge energy before replacement is needed is indeed an indication of thermal damage to the device with each event.

Yes, but you can’t break ohm’s law.
P=IE.

 

[LarryFine]

Yes, but you can’t break ohm’s law.
P=IE.

That's "Watt's Law"; Ohm's Law is E=IR.

One volt is capable of pushing one amp through one ohm.

 

[acelectric103]

Educate me, please.

My only experience with surges in homes has been ones that cause multiple devices (computers, game consoles, microwaves, TVs) to fry at the same time. They have all been caused by a lost or faulty neutral. Easily identifiable and fixed. Catastrophic failures.

Where I currently live, people have "surge" damage all the time but usually only on one device in the house. I find it hard to believe with the number of electronic devices in most homes.

1. Why would that be true?

2. Wouldn't the power supply fail in addition the PCBs or other internal components?

3. Don't some of these devices have internal overload protection incorporated?

Yes, I have an SPD in my panel.

Electronic devices don't require a huge surge to be destroyed.

 

[GoldDigger]

That's "Watt's Law"; Ohm's Law is E=IR.

One volt is capable of pushing one amp through one ohm.

Actually, E=IR is just the definition of resistance. Ohm's original law was that if you define R(V) for a circuit element by R=V/I, then over a wide range of voltages (as well as other variables) R is a constant. He did not encounter examples of what we now see as variable resistance and negative resistance materials.

 

[retirede]

That's "Watt's Law"; Ohm's Law is E=IR.

One volt is capable of pushing one amp through one ohm.

Yeah, I know.
I had an instructor refer to to the power formula as “Advanced Ohm’s Law”. It kind of stuck and right or wrong, I lump them together.

 

[PQ Man]

We put together a video a few years ago for surge protection and tried to keep it non-commercial. This video highlights the benefit of short lead length that several people have talked about.

It also briefly shows what MOV's look like and in a very simple way, how they work. For simple calculation purposes, we use 15-25V/inch of let through for extended lead length on an SPD. SPD's are tested with 6" of lead so if you have 36" of lead length and it is supposed to clamp at 400V, it will clamp at 400V plus about 20VX30" or 1000V at the load center.

This is not uncommon on a main load center - people mount the SPD at the bottom and connect to the top of the panel 2 pole breaker because they think they need to have it next to the incoming main. However, the busbar in the load center is a very low impedance (inductance) compared to the 10 AWG connecting the SPD so mounting the SPD with 6-10" of conductor at the bottom of the load center (or top) could be much more effective. The SPD's MOV's actually clamp at about 400V (or slightly less) at the SPD device itself but the surge current (which is very high frequency and could be on the order of 1 MHz) has to travel out the SPD and back on 10 AWG wire which is why there is a significant voltage drop reducing its effectiveness at the load center where the wires are terminated on the breaker. The impedance of the conductor is based on 2XpiXfXL where the frequency f, can be very high compared it its 60 Hz impedance. This means that the other loads from the panel see the higher let through voltage.

The benefit of having a main SPD is to help with things coming in the front door (from the utility or through the main panel). The benefit of the secondary (plug location SPD) is that it will help with internally induced transients from other loads in the home or from the let through of the main SPD. Since as much as 80% of transients come from inside you home or business (even though they are smaller in magnitude), this 2nd level is especially important.

 

[paulengr]

We put together a video a few years ago for surge protection and tried to keep it non-commercial. This video highlights the benefit of short lead length that several people have talked about.

It also briefly shows what MOV's look like and in a very simple way, how they work. For simple calculation purposes, we use 15-25V/inch of let through for extended lead length on an SPD. SPD's are tested with 6" of lead so if you have 36" of lead length and it is supposed to clamp at 400V, it will clamp at 400V plus about 20VX30" or 1000V at the load center.

This is not uncommon on a main load center - people mount the SPD at the bottom and connect to the top of the panel 2 pole breaker because they think they need to have it next to the incoming main. However, the busbar in the load center is a very low impedance (inductance) compared to the 10 AWG connecting the SPD so mounting the SPD with 6-10" of conductor at the bottom of the load center (or top) could be much more effective. The SPD's MOV's actually clamp at about 400V (or slightly less) at the SPD device itself but the surge current (which is very high frequency and could be on the order of 1 MHz) has to travel out the SPD and back on 10 AWG wire which is why there is a significant voltage drop reducing its effectiveness at the load center where the wires are terminated on the breaker. The impedance of the conductor is based on 2XpiXfXL where the frequency f, can be very high compared it its 60 Hz impedance. This means that the other loads from the panel see the higher let through voltage.

The benefit of having a main SPD is to help with things coming in the front door (from the utility or through the main panel). The benefit of the secondary (plug location SPD) is that it will help with internally induced transients from other loads in the home or from the let through of the main SPD. Since as much as 80% of transients come from inside you home or business (even though they are smaller in magnitude), this 2nd level is especially important.

I try to explain it this way.

A fuse or circuit breaker is in series. It protects everything in the circuit with it.

A surge arrester is in parallel. It SHARES everything, just not necessarily equally.

On a white board then draw it this way, two devices in parallel. Now draw in the impedance of the wiring going to the load. Now draw in the voltage across the load. See...it sees three impedances in series...the surge arrester PLUS the wiring. Then we can get into surge impedance, why it is inductance not resistance, and representative values.

Then taking that example from breakers in the panel to say a computer located on #14 50 feet away it becomes obvious while house surge protectors don’t protect any critical loads at all.

 

[xformer]

So when an MOV experiences a surge, its resistance drops, and current flows through it in an attempt to clip the voltage surge.
There is current flowing through the device and voltage across it. How is that energy dissipated if not by heat?

IMHO I think the dissipated energy is dissipated as heat, but... the definition of a Surge is for an extremely short period of time. A time so short that the material wouldn't have a chance to absorb the heat or radiate the heat.

 

[oldsparky52]

I live in the "lightning capital of the USA". In one of the counties I work in, surge protectors have been required in panels by local rule for quite some time. This does not make it any easier to sell whole house surge protectors in the other counties. Most people think you are trying to sell them snake oil.

Which brand or model surge protector do you have in your home?

 

[Coppersmith]

I have a GE plug-on. The GE panel came with the house unfortunately.

 

[drcampbell]

... Surge arresters are consumables. They don’t last forever. ...

If I recall correctly, there was a surge absorber (not an arrestor or suppressor) on the market a while back called the "Brick Wall" which was made entirely of passive components -- inductors, capacitors & resistors. No varistors or anything else likely to accumulate incremental damage.

At Chrysler, (and maybe throughout the car biz) we adopted the same strategy for protecting the inputs of brain boxes. Purely passive. They reduce voltage spikes by soaking up energy in capacitors, but there isn't any threshold voltage, like a MOV has.

... Tiny surges eat away at the tiny traces on the circuit boards and also cause cumulative damage ...

The even-tinier traces inside integrated circuits are much more likely to fail.