600 volt, 3 phase surge protection

Status
Not open for further replies.

mike_kilroy

Senior Member
Location
United States
1)Surge current is not the same as short circuit current.

2)It's a heat issue whichever way you cut it.
V*I*t.You simply cannot treat them as independent variables as you seem to have done.
That's basic physics.

3)If you had ever examined the nature of the damage to the silicon chip from a failed large diode or SCR, you would have observed that the damage, whatever the mode of failure, is some or all of the chip had melted. You can sometimes determine the cause of the failure by the position and area of the chip that melted. But melted it has. And melting needs heat.
It really is that simple.

1) For reference, I never once said short circuit current was surge current; I always said what I meant, max current, ie., max surge current. The short circuit words were other posts you are for some reason attributing to me.

2) no, it is not. Read the ref in post above this and perhaps you will agree that there is a real rating limitation called max surge current that degrades and kills MOVs - not watt-seconds, max current number only.

3) unfortunately you choose to ignore my posted comment that exceeding a rating - PIV - not a watts for some seconds as energy rating - will kill a transistor. THEN after it shorts, the hi watt-seconds flow, melt it, and blow it up. You will never get a 3"x3" semiconductor device make a shock wave while failing with a 2 joule spike - it needs the 460v@1000amp short circuit current that follows to make that sound. Yes, a semiconductor can overheat and fail shorted and thus melt and blow, but quit a few failures are actually from voltage spikes - without much energy behind them - that exceed PIV causing the semiconductor to fail - when it fails in a short circuit mode it THEN AFTER THE FAILURE FACT melts and blows up. You have to be ready to look beyond the obvious "it melted." Same with MOVs - quite often they degrade from the surge currents (pronounced "low joules") that degrade them so their turn on voltage reduces - per above article typically 10% per over max current surge, until it turns on at the AC voltage it is meant to protect, THEN conducting, melting, and failing fully.
 

Besoeker

Senior Member
Location
UK
1)The short circuit words were other posts you are for some reason attributing to me.

Um.........

Post #9
SPD's have short circuit ratings and need to be applied appropriately.

Post #18
If there was a fault in the SPD, maybe an MOV that fails short instead of open, how much current can the SPD safely handle without catastrophically failing. That is why a SCR is needed.

Post #41
Just like busway, transfer switches, safety switches, etc, the manufacturer indicates a withstand rating (short circuit rating)




2) no, it is not. Read the ref in post above this and perhaps you will agree that there is a real rating limitation called max surge current that degrades and kills MOVs - not watt-seconds, max current number only.
Varistor. Variable resistor.
You have current through it, you have power. Have that for a duration, you have energy. Heat.
You simply can't treat the quantities as independent variables.

3) unfortunately you choose to ignore my posted comment that exceeding a rating - PIV - not a watts for some seconds as energy rating - will kill a transistor. THEN after it shorts, the hi watt-seconds flow, melt it,[/QUOTE]
Not so. I responded thus:

If you had ever examined the nature of the damage to the silicon chip from a failed large diode or SCR, you would have observed that the damage, whatever the mode of failure, is some or all of the chip had melted. You can sometimes determine the cause of the failure by the position and area of the chip that melted. But melted it has. And melting needs heat.
It really is that simple.
You're puting the cart before the horse.
Exceeding PIV does not lead to catastrophic failure in semiconductors unless there is sufficient energy to cause heating to the extent that causes damage.
We used to routinely test SCRs in a rig that increased voltage to the point where forward or reverse voltage ceased to be blocked. That gave us a handle on the device withstand capabilities whilst not delivering enough energy to result in junction failure.

And, just another little thing for you to think about. If a device truly goes short circuit, it won't support 460V - 460v@1000amp short circuit just ain't gonna happen.
 

mike_kilroy

Senior Member
Location
United States
guess there is no sense in my responding to ur posts since u don't read them.

1) None of your referenced posts are mine. So accusing me of saying something I didn't has no bearing on the topic.

2) u r simply not wanting to see that yes, there are other ways to fail than pure long term heat. so be it.

3) guess u know better than all the MOV and TVSS manufactures who say a 2 joule (read again, low heat) HI CURRENT spike can kill an MOV/TVSS as well as a max joule rating 100-1000x higher. so be it. Exceeding PIV will kill a semiconductor before hitting that semiconductors heat dissipation rating - hence another example of more than just simple max dissipation rating on a device. So be it. All the mfgrs and i r wrong. ok.
 

ELA

Senior Member
Occupation
Electrical Test Engineer
Just look at any TVSS device spec - for example quick digikey search:

http://media.digikey.com/PDF/Data%20Sheets/Thomas%20Research%20Products/BSP3%20LC%20Series.pdf

p
ick the 480v model for an example. good for 360 joules. also limited to 10,000amps max amps. industry test appears to be 8/20 usec pulse, so 550v*10,000a*8/20E-6=2.2 joules. MUCH less than its max heating power rating. But it will blow up if we hit it with 20,000a*550v*8/20E-6= 4.4 joules.

.

mike_kilroy,

It is about the heating effect in each case.
With the high amplitude currents for short durations you get the intense localized heating damage because the power is so intense it does not have time to spread out via thermal conduction.

Often times manufactures play games with joule ratings. From the data sheet note that the 360 joule rating is at 2 milliseconds.

Regarding your above example:
The 8/20 figure is not a division. It is stating an 8 usec rise time and 20 usec decay to 50% level waveform. To be totally accurate you need a complex equation but you might approx. it by using 20 usec. Then the numbers come a lot closer. Next consider that the clamping voltage will be much greater at the higher current (Perhaps as much as a factor of 3x).
 

Besoeker

Senior Member
Location
UK
1) None of your referenced posts are mine. So accusing me of saying something I didn't has no bearing on the topic.
Agreed. My error and for that I apologise.

2) u r simply not wanting to see that yes, there are other ways to fail than pure long term heat. so be it.
I did not claim that it was long term heat. Just heat. Energy.

Exceeding PIV will kill a semiconductor before hitting that semiconductors heat dissipation rating
You seem to have ignored this:
Exceeding PIV does not lead to catastrophic failure in semiconductors unless there is sufficient energy to cause heating to the extent that causes damage. We used to routinely test SCRs in a rig that increased voltage to the point where forward or reverse voltage ceased to be blocked. That gave us a handle on the device withstand capabilities whilst not delivering enough energy to result in junction failure.
It is heat that causes some or all of the chip to melt thus it's demise.
A pic of a failed device that sits on my desk as a paper weight - the tungsten backing makes it quite heavy.
It's a typical dv/dt failure.

FailedSCR02.jpg


The current flows in a relatively small area, it gets hot, and loses the plot. It melts. Joules.
 
Last edited:

mike_kilroy

Senior Member
Location
United States
ela, thanks for the 8/20 correction; i should never had taken the bait to try to become a MOV wizard which I am not. I just tried to correct some stubborn wrong notions stated here that there is no such thing as multiple ratings that cause a failure that are not in units of joules. The fact is that excess voltage or current above some max defined value will often cause a device to fail, irregardless of how long it is applied or what the other terms are.
 

kwired

Electron manager
Location
NE Nebraska
For reference, a shot gun blast is about 3,000 joules. Dirty Harry's 44 magnum produces 900 joules.
Just like the semiconductor devices mentioned it depends where the energy is or where it travels. Is those figures at the cartridge, at the muzzle, specific distance from muzzle? Each point mentioned will have a different amount of energy pass by it. The chamber and barrel absorb some energy, ask anybody that has been around 50 cal machine guns. Fire for too long at a time you can melt down the barrel, they issue you a spare barrel with the weapon in the military, and a hot mitt and carrier to handle a hot barrel.

guess there is no sense in my responding to ur posts since u don't read them.

1) None of your referenced posts are mine. So accusing me of saying something I didn't has no bearing on the topic.

2) u r simply not wanting to see that yes, there are other ways to fail than pure long term heat. so be it.

3) guess u know better than all the MOV and TVSS manufactures who say a 2 joule (read again, low heat) HI CURRENT spike can kill an MOV/TVSS as well as a max joule rating 100-1000x higher. so be it. Exceeding PIV will kill a semiconductor before hitting that semiconductors heat dissipation rating - hence another example of more than just simple max dissipation rating on a device. So be it. All the mfgrs and i r wrong. ok.

I just want to comment that we (or mother nature) don't put a 55kamp surge or a 4 joule surge on a line. The surge may be a specific amount of energy but to our circuit in question it first is just an increase in voltage, how much current flows or how much energy was dissipated depends on how much voltage was present along with impedance of the current path. Any energy from the surge that does not pass through our particular circuit traveled some other path and the impedance of that path will dictate how much goes through that path, as well as any other paths involved.

The more loads there are parallel to the point we are discussing the more energy can be dissipated through those parallel paths, whereas a single load will have to attempt to conduct all that is imposed on it.
 

Besoeker

Senior Member
Location
UK
ela, thanks for the 8/20 correction; i should never had taken the bait to try to become a MOV wizard which I am not. I just tried to correct some stubborn wrong notions stated here that there is no such thing as multiple ratings that cause a failure that are not in units of joules. The fact is that excess voltage or current above some max defined value will often cause a device to fail, irregardless of how long it is applied or what the other terms are.
It isn't a stubborn notion.
It's simple physics.
 
T

T.M.Haja Sahib

Guest
Bes, why fixated on the idea of SPD destruction by thermal energy of current only? Consider the possibility of its destruction by mechanical force of the current and either can occur independently.
 

pfalcon

Senior Member
Location
Indiana
Take the same device as above, 360 joules AND 10,000amps max limits....
scenario 1: spike comes thru that amounts to 359 joules & less than 10,000amps: device lives.
Device functions as designed and lives to see another day.

scenario 3: spike comes thru that amounts to 4.4 joules & 20,000amps: device dies.
Device fails catastrophically creating an arc flash.

scenario 4: spike comes thru that amounts to 400 joules and less than 10,000amps
Device fails non-catastrophically and has to be replaced.

Yes? No?
 

Besoeker

Senior Member
Location
UK
Consider the possibility of its destruction by mechanical force of the current and either can occur independently.
I've never seen a solid state device fail owing to mechanical force. So, the probability of such an event is vanishingly small in my experience.
What's your experience of it?
 
T

T.M.Haja Sahib

Guest
I've never seen a solid state device fail owing to mechanical force. So, the probability of such an event is vanishingly small in my experience.
In a SPD with multiple MOV's in parallel, this could happen and so I think your experience is questionable.
 

Besoeker

Senior Member
Location
UK
Thanks.:thumbsup:
You're welcome. Your particular brand of charm is worth of a mention.

None. But, does it matter here?

Well now, what do you think?
Whose opinion do you think most people would consider to be the more valid?

That given by someone with some decades of experience in the field of solid state power?
Or that given by someone who has no experience in this field but is opinionated on absolutely everything and has a propensity to denigrate others on absolutely anything?
 
Status
Not open for further replies.
Top