surge protection built into ordinary household electronics

[Matt]

Well, "Yes", that is what it means, but think about what has to happen
to get there.

You have a lot of stuff in front of this spec. Your transfomer,
filter caps, the bridge rectifier, the fuse. The 60V/ 100mS has to
get through all that to do any damage.

It would be a good exercise for you to look at just the transformer.
Imagine 60V on the secondary. How much would that be on the Primary?

The transformer gives 6.3VAC RMS from a 120VAC input. So there would
have to be about (120/6.3)*60V = 19*60V = 1140V at the primary.

Do you think this is a reasonable value you would ever expect to see
coming out of the wall socket?

I guess you're saying that it isn't reasonable. But if a motor kick
can't do that, surely lightning can ... I guess ...

But Graham says that lightning protection is hopeless ... okay ...

 

[Eeyore]

The transformer gives 6.3VAC RMS from a 120VAC input.

I can see why you need an LDO regulator then !

So there would have to be about (120/6.3)*60V = 19*60V = 1140V at the primary.

You won't get that for 100ms.

I guess you're saying that it isn't reasonable. But if a motor kick
can't do that, surely lightning can ... I guess ..

The magnitude of the two is billions of times apart. If lightning actually hits
your house, what happens next is plain luck (or otherwise) for the most part.

But Graham says that lightning protection is hopeless ... okay ...

Close in protection from lightning is indeed pretty much hopeless. But the point
is, it'll mainly cause a *COMMON MODE* voltage on the ac input, not a
differential one.

Your main problem seems to be that you have very vague, generalised and poorly
informed / defined ideas of what to expect on the ac mains.

Graham

 

[mpm]

Your main problem seems to be that you have very vague, generalised and poorly
informed / defined ideas of what to expect on the ac mains.

It's a little off-topic, but probably the worst in-service power setup
I ever saw personally, was our 3.2 mile 1-Ph line buried down the
center of a rock/dirt (mostly rock) road on a remote mountain
broadcast transmitter site just NW of Denver, CO.

The line was 20-years old (at least!) and because the terriain was
difficult, was not buried too deeply or in conduit.
The insulation was (guesstimated) rated for 7200 volts, but was
running higher than that due to demand. The local power company would
take no ownership of this baby!!

At the top, the broadcast transmitters all wanted 3-phase, so insert
multiple motor-driven phase converters into this discussion. Not to
mention a heavy 1-ph load to boot. And (2) autostart diesel
generators for backup that could be a real bit^h to start, let alone
re-fuel in blizzard conditions on those steep mountain passes and
switchbacks.

And as if that weren't enough.... I'm pretty sure the snow plow hit
it in a couple of places over the years!!

And even that place got lightning hits and survived.
It is not impossible to make something lightning-proof, but it is
expensive for all intents and purposes.

Site: Eldorado, Mtn.
Jefferson County, CO.

 

[w_tom]

I guess you're saying that it isn't reasonable. But if a motor kick
can't do that, surely lightning can ... I guess ...

But Graham says that lightning protection is hopeless ... okay ...

Trivial surges such as motors must be made irrelevant by protection
already inside your power supply design. That is not found in
protector components. For example, you power supply must have
galvanic isolation for human safety. And that galvanic isolation is
made sufficient to made surge damage irrelevant.

Maybe once every seven years, a surge may occur that might overwhelm
protection that is standard in all appliances. Therefore earthing
makes that surge irrelevant. For example, a TV cable connects
directly to earth ground - not protector components - for effective
surge protection. That earth ground (quality of) determines the
protection. Surges so well earthing that protection inside all
appliances is not overwhelmed.

For other incoming wires (telephone, AC electric) direct earthing
is not possible. Therefore we install a 'whole house' protector with
that 'less than 10 foot' earthing connection. Even wire length too
long destroys the protection. Protection installed so that direct
lightning strikes find earth ground via that earthing electrode; not
destructively via electronics appliances or your design.

These requirements for electronics protection were even standard 30
years ago. Protection that is exceeded (as demanded by Intel specs)
in a properly designed ATX power supply.

Don't believe those myths that lighting damage is normal. When was
the last time your telco, their computer connected to overhead wires
all over town, was down for 4 days (no phone service) as they replaced
that computer? With maybe 100 surges during each thunderstorm, they
don't have damage. They also use a properly earthed 'whole house'
protector on each incoming wire. And up to 50 meter separation
between that protector and electronics. That separation also part of
the protection.

You may also want to learn from manufacturer data sheets about
standard ICs now required to meet 2000 and 15000 volt surge
protection. Those who don't learn from datasheets and other numbers,
instead, assume protection is not possible.

Those who posted nothing can protect from lightning? You now know
who has not first learned electronics design and other standards. But
they know anyway.

Meanwhile the fuse will take milliseconds to blow. Surge damage
completes in microsecones. Anyone with basic electronics knowledge
should have known this and posted it. Just another benchmark for
those who are making recommendations and yet don't even have basic
electronics knowledge. Fuses are not for electronics protection.
Their function - to protect you from a fire after electronics has been
damaged. Those other posters should have known this.

 

[Eeyore]

electronics knowledge. Fuses are not for electronics protection.
Their function - to protect you from a fire after electronics has been
damaged. Those other posters should have known this.

They do.

Fuses aren't for protection, hence not mentioned.

Graham

 

[Eeyore]

Meanwhile the fuse will take milliseconds to blow.

Or even seconds.

Surge damage
completes in microsecones. Anyone with basic electronics knowledge
should have known this and posted it. Just another benchmark for
those who are making recommendations and yet don't even have basic
electronics knowledge. Fuses are not for electronics protection.
Their function - to protect you from a fire after electronics has been
damaged. Those other posters should have known this.

Of course they know it.

Fuses weren't emntioned because they're not for surge protection.

Graham

 

[D from BC]

Or even seconds.


Of course they know it.

Fuses weren't emntioned because they're not for surge protection.

Graham

Wouldn't an inline fuse be also useful to soak some spike energy when
it's combined with a shunt type overvoltage circuit such as a MOV,
TVS, SCR, spark gap,zener or discharge tube.
The fuse acts like a resistor.. Doesn't blow... justs turns spikes
into heat.
D from BC

 

[Eeyore]

Wouldn't an inline fuse be also useful to soak some spike energy when
it's combined with a shunt type overvoltage circuit such as a MOV,
TVS, SCR, spark gap,zener or discharge tube.
The fuse acts like a resistor.. Doesn't blow... justs turns spikes
into heat.

A fuse is a conductor. It won't even notice a transient. Even super fast fuses
take (tens of) milliseconds to blow.

In any event there's probably not actually that much energy in the typical
transient.

Fuses are there to interrupt gross *fault currents*. That's their only purpose.

Graham

 

[Matt]

I can see why you need an LDO regulator then !

Now that you mention it, I'd better move to an 8V transformer. Dropout
for LM2937 is about 0.7V at full load.

I'll use a LM2937 to supply the microelectronics, and I'm switching a
150mA@5V relay using a PNP that supplies a second LM2937 that supplies
the relay. That regulator is a $2 part, but I want the relay to get an
even 5 volts when it needs it. I figure the relay contacts (rated
16A@277VAC) will last longer that way. Can you mention a better/cheaper
way to switch the relay?

 

[mpm]

What are you talking about??
I completely disagree with almost everything you say

Protection takes many forms:
- Fire Protection
- Physical Electrical Shock Hazard (Electrocution)
- Ground Fault Interruption
- Equipment Protection
- Component-Level Protection (ESD, etc..)

You speak of these as it they are all the same, or make inappropriate
comparisons.
THEY ARE NOT THE SAME!!!!

And the reason you don't see your phone or electric go down is because
they are REDUNDANT systems, not because they are immune to lightning!
(In the case of Electric, I mean the physical layout of the grid
itself. This is not always redundant depending on how your
neighborhood is served, and where the pole disconnects are located...)

I have seen cracked circuit boards caused by flying chucks of concrete
blocks walls that failed because of the percussive wave (Arc- Blast)
of lightning. Explain to me how a ground rod is going to help that?

As for your most egregious comment: That fuses are "not for
electronic protection", I suggest you take apart your TV set and put a
5,000 amp fuse in it. Then open the back cover and throw in about 2
pounds of glitter or aluminum foil. After the traces on the circuit
board lift (or melt) from the heat generated, you can explain to all
of us how fuses don't matter.

Regarding your "Earthing" comments, I don't even know where to
begin....
Please re-post your thoughts using the proper NEC Article-250
terminology. That would be a really good start.
As it stands, they seems to assume household-type knowledge of a
single ground rod driven according to local building codes. (Which,
as we know is a life-safety, fire-based code for clearing faults on
service entrances and branch circuits - not protecting anything
"electronic".)

When lighting strikes, there is a fault current flowing in the ground
circuit.
Do you really think that ground rod is doing anything for you at that
point?

I agree, you can make something nearly "lightning-proof", but it is
often economically or physicaly impractical to do so.
It is often cheaper to install redundancy, and this generally provides
other benefits besides up-time.

-mpm

 

[mpm]

On Jun 8, 3:13?am, Eeyore <[email protected]>
wrote:

A wire-wound fuse would exhibit significant reactance to lightning
frequencies.
But I wouldn't count on this for "protection".

 

[mpm]

Now that you mention it, I'd better move to an 8V transformer. Dropout
for LM2937 is about 0.7V at full load.

I'll use a LM2937 to supply the microelectronics, and I'm switching a
150mA@5V relay using a PNP that supplies a second LM2937 that supplies
the relay. That regulator is a $2 part, but I want the relay to get an
even 5 volts when it needs it. I figure the relay contacts (rated
16A@277VAC) will last longer that way. Can you mention a better/cheaper
way to switch the relay?

Is there a reason you want to use an LDO here?
The LM2937 is only good for 500mA, and you seem to want 150mA of that
for relay switching.
Which might be OK, but I don't know what your other loads are.

I would suggest a "beefier" regulator.
Then you can tie the relay high, and switch the other end to ground
through a FET.
That way, you only need one regulator.

I want to confirm your relay coil is 5 VDC, not AC right??

Why not take a look at a standard "7805" regulator which you can get
for about $.0.30 or less.
For a FET, it depends on what you want to drive it with, but a
RFP12N10L won't give you any trouble (n-channel enhancement) mode.
In fact, it's probably overkill, but it will work. They run about
$0.75 in small quantities.

A 2N7000 would also work.
Assuming DC Coil, don't forget a reverse bias diode on the relay coil
to absorb the collapsing magnetic field (relay kickback).

One more thing, the relay might require 150mA to pick the contacts,
but a lot less to actually hold them.
If your regulator filter caps can handle this, the extra instaneous
load when picking the relay might not be as bad as you envision. Test
it and see. Good luck. -mpm

 

[D from BC]

On Jun 8, 3:13?am, Eeyore <[email protected]>
wrote:

A wire-wound fuse would exhibit significant reactance to lightning
frequencies.
But I wouldn't count on this for "protection".

I'm still learning surge suppression.

Someday I'll play around on LTSpice to simulate line spike control.
But what level and duration of spike to simulate?
40 000 Vpk for 10mS?
100 000V pk for 1mS?
2000Vpk for 1S

I think it depends an how much PCB space is available and what the
budget is...

One time I gave myself a 1sqin space to fill with line surge
suppression. What can I get?
How much POP energy can a 1 sqin space handle before exploding? (A
good explode to form isolation.)

I just picked a MOV..not perfect but helps..
I think there was many other alternatives.

If lightning strikes the line...Hopefully the strike will be far away
such that the energy level is low enough for the MOV to shunt.

Also
For my region here in BC...
Electrical storms are very rare here (about 1/year)
Power failures are very rare (about 1/year)

I'm more worried about pops from inductive loads..
Air conditioners and AHU's?
12amp power tools?
40 or so light ballasts on the same circuit?
Machinery?

D from BC

 

[D from BC]

What are you talking about??
I completely disagree with almost everything you say

Protection takes many forms:
- Fire Protection
- Physical Electrical Shock Hazard (Electrocution)
- Ground Fault Interruption
- Equipment Protection
- Component-Level Protection (ESD, etc..)

You speak of these as it they are all the same, or make inappropriate
comparisons.
THEY ARE NOT THE SAME!!!!

And the reason you don't see your phone or electric go down is because
they are REDUNDANT systems, not because they are immune to lightning!
(In the case of Electric, I mean the physical layout of the grid
itself. This is not always redundant depending on how your
neighborhood is served, and where the pole disconnects are located...)

I have seen cracked circuit boards caused by flying chucks of concrete
blocks walls that failed because of the percussive wave (Arc- Blast)
of lightning. Explain to me how a ground rod is going to help that?

As for your most egregious comment: That fuses are "not for
electronic protection", I suggest you take apart your TV set and put a
5,000 amp fuse in it. Then open the back cover and throw in about 2
pounds of glitter or aluminum foil. After the traces on the circuit
board lift (or melt) from the heat generated, you can explain to all
of us how fuses don't matter.

Regarding your "Earthing" comments, I don't even know where to
begin....
Please re-post your thoughts using the proper NEC Article-250
terminology. That would be a really good start.
As it stands, they seems to assume household-type knowledge of a
single ground rod driven according to local building codes. (Which,
as we know is a life-safety, fire-based code for clearing faults on
service entrances and branch circuits - not protecting anything
"electronic".)

When lighting strikes, there is a fault current flowing in the ground
circuit.
Do you really think that ground rod is doing anything for you at that
point?

I agree, you can make something nearly "lightning-proof", but it is
often economically or physicaly impractical to do so.
It is often cheaper to install redundancy, and this generally provides
other benefits besides up-time.

-mpm

I read that grounded post too..
It got me thinking though..

When lightning hits a power pole the energy is going to ground..
At the point of impact it branches like tree roots taking every little
indirect route to ground.
Most of the energy goes down the pole but some energy branches to
homes..
Let's say a home is lucky enough that that the spike voltage is below
the arcing voltage to earth grounded conduits and earthed "green"
wires.
What remains is a spike that could cause some damage..

Imagine connecting a 1 million volt AC supply to a house.. I think
there will be many arcing paths long before reaching a bedroom outlet.

D from BC

 

[Roger Hamlett]

I read that grounded post too..
It got me thinking though..

When lightning hits a power pole the energy is going to ground..
At the point of impact it branches like tree roots taking every little
indirect route to ground.
Most of the energy goes down the pole but some energy branches to
homes..
Let's say a home is lucky enough that that the spike voltage is below
the arcing voltage to earth grounded conduits and earthed "green"
wires.
What remains is a spike that could cause some damage..

Imagine connecting a 1 million volt AC supply to a house.. I think
there will be many arcing paths long before reaching a bedroom outlet.

Some years ago, I saw a person electrocuted in a factory, by a lightning
strike. He survived, but only with a lot of luck. What happened, was that
he was working on a lathe a few feet inside the factory building, which
was earthed to a very good earth point, near the side wall. Lightning
struck a tree, just outside the building only a few yards from the earth
point. The machine's 'earth', was suddenly massively elevated, relative to
the floor on which he was standing. It stopped his heart, but good medical
training from another staff member, saved him. Other effects, were to blow
apart some of the electrical conduits in the wall, and disintegrate quite
a few safety connections.
Where I worked at the time, we used to test electrical 'survivability' on
the kit we were making, to 250,000v, at 10,000A. for 1uSec, and then lower
currents and voltages for longer times. Every bit of our kit on this
particular site survived fine. However I have seen the same type of kit in
one case, literally having 'exploded', when installed on a site that
received a sequential multiple 'strike'. It is when you visit a building,
and see the remains of a 3/4 inch solid copper lightning conductor, that
has been vapourised, destroying parts of the brickwork behind, that you
beging to get an idea of just how powerful a really close strike can be.

Best Wishes

 

[mpm]

On Jun 8, 4:56?pm, "Roger Hamlett"

It may not even have been the lathe?
Could have been were his feet were planted.

Aerial cranes sometimes get energized when they accidentally contact
overhead power lines.
(This energizes the cab).

There are many documented cases where a crane operator jumps out of
the crane to relative safety but then gets electrocuted when he walks
away from the crane. (Why? Because a fault current is flowing in the
earth; the potential difference between his feet can be lethal.) The
correct procedure is to hop away with both feet together...

When strikes are close, I mean really close, you really need to think
of strikes as electromagnetic energy launched on a 3-dimensional
transmission line. The lines are made up of your physical body, the
atmosphere, the trees, the buildings, the Earth, you name it!. It's a
lossy medium, with multiple dielectric constants, multiple dielectric
breakdowns, multiple current paths (not all of them to ground), and
multiple frequencies - though substantially all of the energy is
contained in the VHF range.

Then you have the heat and Arc Blast to deal with. Nasty, nasty
stuff!.

Sounds like you were making coaxial surge suppressors..????
Glad to hear your co-worker survived.

 

[Matt]

Thanks for your reply.

Is there a reason you want to use an LDO here?

Mainly because I was using a 6.3V transformer. Now it seems wiser to
switch to a 300mA 8V transformer.

The LM2937 is only good for 500mA, and you seem to want 150mA of that
for relay switching.
Which might be OK, but I don't know what your other loads are.

Relay -- 200 mA holding current
Power LED -- 20 mA
Relay-on LED -- 20 mA
microelectronics -- <25 mA

I would suggest a "beefier" regulator.

Did you think I was using two regulators because one LM2937 doesn't give
enough current? No, it was because I assumed a current drop through the
relay-switching transistor, and I wanted the relay to get a full 5V. I
barely knew what a FET was before you mentioned it.

Then you can tie the relay high, and switch the other end to ground
through a FET.

Any reason I shouldn't tie one end of the relay coil low and use the FET
to switch the other end high?

That way, you only need one regulator.

I want to confirm your relay coil is 5 VDC, not AC right??
Yes.


Why not take a look at a standard "7805" regulator which you can get
for about $.0.30 or less.

Thanks, I'll do that.

For a FET, it depends on what you want to drive it with, but a
RFP12N10L won't give you any trouble (n-channel enhancement) mode.

Ooh, I think that hits the spot. Only 0.15 ohms resistance at 200 mA,
so only a 0.03V drop across the transistor.

Now do I understand that I don't need to limit the switching current
going to the gate? I guess I just need a pull-up or pull-down resistor
there.

In fact, it's probably overkill, but it will work. They run about
$0.75 in small quantities.

A 2N7000 would also work.

Looks like that part has about 10 times the resistance as the RFP12N10L.

Assuming DC Coil, don't forget a reverse bias diode on the relay coil
to absorb the collapsing magnetic field (relay kickback).

I'm thinking of a TVS: SMBJ5.0CA
http://www.fairchildsemi.com/ds/SM/SMBJ20A.pdf
http://relays.tycoelectronics.com/appnotes/app_pdfs/13c3264.pdf

 

[Gerhard Fiedler]

Any reason I shouldn't tie one end of the relay coil low and use the FET
to switch the other end high?

In that configuration, you'd use an n-channel FET, which have lower Rdson
and are cheaper than equivalent p-channel FETs (which you'd have to use to
switch the load high). Also, the gate driver is simpler for the low-side
configuration when you have load supply voltages that are higher than the
logic supply (which isn't your case here).

Gerhard

 

[whit3rd]

Now that you mention it, I'd better move to an 8V transformer. Dropout
for LM2937 is about 0.7V at full load.

Not required. 6.3V is the RMS voltage, the peaks are 10V
and (after bridge rectifier) you get 8V on your filter capacitor.

And if you need to switch a relay, why do that with the regulated
power? Just use a $0.06 PN2222 transistor with a base resistor,
emitter to ground, and tie the terminals of the relay to unregulated
8V
and to the collector of the transistor. Your relay winding has
tolerance
for overvoltage, surely!

 

[Jasen]

All household goods have filters in now any way to pass EMC regs.

even toasters, lamps, alarm clocks, and jugs?

Bye.
Jasen