Surge / Ground / Lightning
in [Home Built PC]
|
From: phil-news-nospam on 5 May 2008 15:25 In alt.tv.tech.hdtv spamfree(a)spam.heaven wrote: | I'm curous to know how surge suppression can work without a ground | (earth) of any sort. Does the "black box" detect overvoltage and | disconnect the power like an earth leakage safety switch? Without a ground of any sort, not all types of surges can be protected against. But some can. If the surge is a differential one (some use the term transverse), then what the surge suppressor can do is cancel it out by effectively short circuiting it to itself. A differential surge involves two wires with the voltage on each being of opposite polarity and equal level. The MOV component inside the suppressor will normally not be conductive. But when the voltage is high enough, it becomes a conductor. The arrival of a high voltage differential surge will result in the MOV between those 2 wires to become conductive. If the surge is a common mode one, AND if the surge has a slow rise time, then a device that is interconnected to other wires or other devices can be protected by allowing the surge to pass to all devices at the same level. As long as the rise is not too fast, keeping all the incoming wires, and all the interconnected devices, at the same level results in insignificant current flows. That surge will either reflect back from the protected equipment to the suppressor, and from there go back through all the connected wires (which could be more than where the surge arrived from). Most strikes have lower energy levels at high frequencies than what would cause damage. The exact frequency level that needs to be considered depends on the internals of the equipment. For example, where it has inductance to one end of a sensitive component like a CMOS chip, and no inductance to some other end, this could result in a very brief fast rise of voltage high enough to damage the CMOS chip. In some cases an LC circuit can actually increase the voltage level of high frequency components (at the resonant frequency). For example if you have energy at some voltage at 200 MHz, an LC series circuit resonant at 200 MHz will result in a higher voltage being present at the connection between the L and the C. So even in cases where there is not enough energy at high frequency in a surge to cause direct damage, it can still happen on some devices (think of them as having a lower threshold of damage to simplify this). | This might be fine for a TV, but surely not for a computer. If everything the computer is connected to is protected at a common point in the same surge suppressor, you can have this kind of protection, even on a computer. That does mean if you have a phone line connected to a modem, you need to protect both the phone line itself and the power to the modem, in common with the computer. | I don't recall any computer I've owned that did not have a three wire | connection to the mains. That and a MOV is OK for smallish surges, but | I believe that for a large surge, the sort that will blow a telephone | off the wall, one needs a large, short-path earth for the surge | detector to dump the extra power down. Such a surge is likely to have high levels of high frequency energy. The effective protection against these rare events is a combination of somewhere to divert the energy (like a ground path), and something (like an inductor) to ensure the energy does get diverted. One problem is that at the point of use, an alternate ground path is not practical. The grounding wire of the power circuit coule be as much a source of the surge as the neutral wire would be. The place to put the diversion system is at the entrance to the building. Most surges that come in by other paths besides the entrance to the building are induced surges that will not have so much energy and even less at high frequencies. Still, I have seen three incidents in which an induced surge damaged a device that was not connected to anything at all (in two cases they were battery powered devices, and in the third, it was disconnected before the storm but suffered damage anyway). | I've got a few plug in protectors here and there to sop up a small | spike, but when a storm is within a few km, I pull the phone wire out | of the ADSL router, and the plug out of the mains. If I'm working at | the time, I might just keep a watch on the weather radar and count | lightning fashes to thunder times. It's rare that I get interrupted. I | have underground power and phone lines so that gives a little extra | protection, I believe. I've been told that Australian phone lines are | the most vulnerable, and the most urgent to protect or disconnect. | I hope to be going wireless soon which obviates this problem. Disconnecting provides even better (but still not 100%) protection. Yes, the underground wiring helps. I don't know the issues with Australian phone lines. I do the wireless thing myself and feel much more comfortable with it. Most of the past damaging surges I've seen come in do that on phone and cable wires, and much less often on power wires. That may be due to the more sensitive aspect of equipment where it connects to these wires. -- |WARNING: Due to extreme spam, I no longer see any articles originating from | | Google Groups. If you want your postings to be seen by more readers | | you will need to find a different place to post on Usenet. | | Phil Howard KA9WGN (email for humans: first name in lower case at ipal.net) |
From: phil-news-nospam on 5 May 2008 15:45 In alt.tv.tech.hdtv bud-- <remove.budnews(a)isp.com> wrote: | The last standards for simulating typical surge waveforms I have seen | (IEEE) were | 1.2 us rise time, 50 us duration | 8 us rise time, 20 us duration | a ring wave with a frequency about 100kHz. So now you are saying these figures represent a typical surge waveform, as opposed to the worst case waveform you said a long time ago. The term typical is generally accepted as a median. That means half of the surges would have a slower rise time, and half would have a faster rise time. My concerns are not the typical surges. I suggest that half the surges don't even need protection at all; they won't cause damage even if there is no protection. But that also means half can be damaging and need the protection. And a fraction of those surges need _substantial_ protection. | All are long relative to 0.2 microsecond, so wave propagation should not | be relevant for household circuits. Maybe for the typical surge. How about for the most energetic 1% that are the ones I'm most concerned with because they are hard to protect against. | A favorite article from w_ also uses a "8x20 us impulse as a very rough | representative pulse" with most harmonic content from 20kHz to 100kHz. | | Martzloff, using the shorter rise time, has written: "For a 1.2/50 us | impulse, this means that the line must be at least 200 m long before one | can think in terms of classical transmission line behavior." And this statement is only using 1.2/50 us as an example. If you think such a timing is the standard, why not offer a quote that actually says that? What does the "/" mean in that case, anyway? I never got to ask you that before. Does it mean "divide 1.2 by 50"? | What reason is there to believe wave propagation is relevant to house | circuits? The most damaging surges (not the typical ones) have substantial fast rise high frequency energy (such as due to a very close direct contact strike). In these cases, even if you can remove all of the low frequency energy, there is still damaging energy in the higher frequencies that do follow transmission line behaviour not only in wiring lengths of typical homes, but even in wiring lengths inside a small appliance like a computer modem. |> As to the advantage of "whole house" vs local surge protection, "whole house |> protection depends on distances to all "protected" items being small. | | Longer distances make the system more subject to effects like direct | induction from lightning into the wiring. I don't see why, in general, | the distance has to be small. I believe he was referring to the distance between the whole house protection and the ground/earth electrode. For things like the service drop distance and the branch circuit distance, it can be a tradeoff between different kinds of surges. The longer wiring will, through its self-inductance, reduce the high frequency energy and slew the rise time of the wavefront ... especially for common mode surges. However, that same longer distance increases the potential level of induced surges where the wire is effectively an antenna. -- |WARNING: Due to extreme spam, I no longer see any articles originating from | | Google Groups. If you want your postings to be seen by more readers | | you will need to find a different place to post on Usenet. | | Phil Howard KA9WGN (email for humans: first name in lower case at ipal.net) |
From: bud-- on 5 May 2008 17:01 phil-news-nospam(a)ipal.net wrote: > In alt.tv.tech.hdtv bud-- <remove.budnews(a)isp.com> wrote: > > | The last standards for simulating typical surge waveforms I have seen > | (IEEE) were > | 1.2 us rise time, 50 us duration > | 8 us rise time, 20 us duration > | a ring wave with a frequency about 100kHz. > > So now you are saying these figures represent a typical surge waveform, > as opposed to the worst case waveform you said a long time ago. Still missing - your source that indicates nanosecond rise times and 100MHz spectrum. > > What does the "/" mean in that case, anyway? I never got to ask you that > before. Does it mean "divide 1.2 by 50"? It is standard notation in the surge field. 1.2 us risetime and 50 us duration -- bud--
From: bud-- on 5 May 2008 17:07 phil-news-nospam(a)ipal.net wrote: > In alt.tv.tech.hdtv bud-- <remove.budnews(a)isp.com> wrote: > > | Previously you said Martzloff "flubbed the experiment". > > I remember that. You were telling me about some information he had > obtained from some experiment. > > | Now you agree with Martzloff that branch circuit must be 200m for > | transmission line behavior with 1.2 microsecond rise time. > > That's not a result of an experiment. "*From this first test*, we can draw the conclusion (predictable, but too often not recognized in qualitative discussions of reflections in wiring systems) that it is not appropriate to apply classical transmission line concepts to wiring systems if ..." As usual, you don�t know what was written. > I'm not so sure the exact distance > is 200m for that exact rise time. But that is a subjective thing. Quit equivocating. Where is your cite. Like for nanosecond risetimes. > > | You say that doesn't apply because surges are faster. Martzloff uses 1.2 > | us because that is a standard rise time for surges produced by lightning > | as defined in IEEE standards. > > Martzloff did not say that was a defined standard in the statement you > quoted. He just used it as an example to come up with the 200m figure. He used it because 1.2/50 (voltage) is an IEEE standard. The 8us from w_�s engineer is another standard (8/20 current). > > | w_' professional engineer source says 8 micoseconds with most of the > | spectrum under 100kHz. > > Even with 1 nanosecond rise time, most of the energy will be present in > the spectrum below 100 kHz. That means nothing when the surge is strong > enough to have energy above some frequency that is relevant to the whole > system involved that can do damage. That frequency might be 100 Mhz for > some thing, and 1 GHz for other things. Still missing � your source. Nanosecond risetime. 100MHz spectrum. > > | You still have *no sources that support your belief* that risetimes are > | far faster. > > I have experience and observation for that. I need no more. Lots of people have experience and observation with flying saucers. The rest of us want a source. -- bud--
From: phil-news-nospam on 5 May 2008 17:00
In alt.engineering.electrical bud-- <remove.budnews(a)isp.com> wrote: | phil-news-nospam(a)ipal.net wrote: |> In alt.tv.tech.hdtv bud-- <remove.budnews(a)isp.com> wrote: |> |> | The last standards for simulating typical surge waveforms I have seen |> | (IEEE) were |> | 1.2 us rise time, 50 us duration |> | 8 us rise time, 20 us duration |> | a ring wave with a frequency about 100kHz. |> |> So now you are saying these figures represent a typical surge waveform, |> as opposed to the worst case waveform you said a long time ago. | | Still missing - your source that indicates nanosecond rise times and | 100MHz spectrum. Observation. |> What does the "/" mean in that case, anyway? I never got to ask you that |> before. Does it mean "divide 1.2 by 50"? | | It is standard notation in the surge field. 1.2 us risetime and 50 us | duration And what does the duration time have to do with high frequency energy? Hint: nothing -- |WARNING: Due to extreme spam, I no longer see any articles originating from | | Google Groups. If you want your postings to be seen by more readers | | you will need to find a different place to post on Usenet. | | Phil Howard KA9WGN (email for humans: first name in lower case at ipal.net) | |