Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Stephen Swaffer]

Re: APC step down transformer
« Reply #12 on: Yesterday at 03:54:04 pm »
ReplyReplyReplyQuote
Quote from: Stephen Swaffer on September 23, 2016, 04:58:40 pm
"This rule is in place because in the vast majority of cases it makes for a much safer system."

David Buckley  "I'd be interested in any examples you may know of where carrying the ground through makes for a safer system."


Any time you have metallic/conductive objects that could become energized you have an unsafe situation.  If two objects are at a differnent potential and you get between them current will flow-simply ohms law.  For that reason code mandates bonding of anything that could become energized.  How can you not carry the ground "through"?

If you bond the primary ground, but allow the secondary to float, then perhaps it might be safer-as long as that  isolation is maintained.  But what is there to tell you that some fault hasn't already occured (and faults will occur)?  Normally, the ground-neutral bond provides a fault path and circuit breaker trips or a fuse blows.  Lacking that bond, "nothing" happens-until a second fault occurs.  Hopefully that isn't someone relying on the "isolation" of the circuit for safety-or the second fault is a "hard" fault while the first is an arcing fault creating tremendous heat until a fire starts.  An intential bond is a good bond.  Incidentally,
isolation transformers are allowed in patient areas in health care-with monitoring to make sure the isolation is maintained.

Just as there are no perfect conductors(normally), there are no perfect insulators-so how much isolation do you really have with an ISO transformer?

Especially in the world of providing pro audio, there are better, more dependable ways to deal with noise than relying on a (hopefully) isolated power circuit.

[John Roberts {JR}]

Re: APC step down transformer
« Reply #12 on: Yesterday at 03:54:04 pm »
ReplyReplyReplyQuote
[Quote from: Stephen Swaffer on September 23, 2016, 04:58:40 pm]
This rule is in place because in the vast majority of cases it makes for a much safer system.

  "I'd be interested in any examples you may know of where carrying the ground through makes for a safer system."

Any time you have metallic/conductive objects that could become energized you have an unsafe situation.  If two objects are at a differnent potential and you get between them current will flow-simply ohms law.  For that reason code mandates bonding of anything that could become energized.  How can you not carry the ground "through"?

If you bond the primary ground, but allow the secondary to float, then perhaps it might be safer-as long as that  isolation is maintained.  But what is there to tell you that some fault hasn't already occured (and faults will occur)?  Normally, the ground-neutral bond provides a fault path and circuit breaker trips or a fuse blows.  Lacking that bond, "nothing" happens-until a second fault occurs.  Hopefully that isn't someone relying on the "isolation" of the circuit for safety-or the second fault is a "hard" fault while the first is an arcing fault creating tremendous heat until a fire starts.  An intential bond is a good bond.  Incidentally,
isolation transformers are allowed in patient areas in health care-with monitoring to make sure the isolation is maintained.

Just as there are no perfect conductors(normally), there are no perfect insulators-so how much isolation do you really have with an ISO transformer?

Especially in the world of providing pro audio, there are better, more dependable ways to deal with noise than relying on a (hopefully) isolated power circuit.

I had to guess about the proper quote organization, let me know if I guessed wrong.

Ground bonding is pretty much the law. I have wrestled with this for decades even through examples where people were killed by bonded chassis plugged into a RPBG outlet.

The only viable alternatives I can anticipate involves a more sophisticated safety strategy like combining a GFCI in the mains feed with a stinger cap ground sized for enough current to trip the GFCI but below the current harmful to humans.

I mentioned this to UL reps in passing but to discuss it seriously would involve serious money that I do not choose to throw at this one just now. To approach this seriously I would first have to buy the UL spec for GFCI power drops ($1k?), then propose a specification change to the ground bonding details in that spec, which involves convincing a committee of participants involved with that spec.

I can't imagine how many zeros that could cost to better protect a pretty narrow class of potential victims (performing musicians who sing with wired microphones between different FOH and backline safety grounds), a shrinking target demographic. 

JR

[Stephen Swaffer]

I apologize for the confusion and misspelling.  I was attempting to respond to David without hijacking this thread:

http://forums.prosoundweb.com/index.php/topic,160837.0/all.html

I feel like the understanding of bonding (as opposed to and relating to grounding) as it relates to safety is a critical topic for anyone dealing with power distribution.

[John Roberts {JR}]

I apologize for the confusion and misspelling.  I was attempting to respond to David without hijacking this thread:

http://forums.prosoundweb.com/index.php/topic,160837.0/all.html

I feel like the understanding of bonding (as opposed to and relating to grounding) as it relates to safety is a critical topic for anyone dealing with power distribution.

Ground "bonding" is all about providing a low impedance path that can carry enough fault current to open fuse/breakers without allowing enough voltage rise to be a hazard itself.

I've had to redesign (beef up) PCB grounds inside a product to meet UL "bonding" criteria.

The joker in the deck is RPBG that can turn a safety ground into a shock hazard.

JR 

[David Buckley]

Have a look at this IEEE paper: Boat Dock Exposure Voltage Mitigation (PDF).  They experimentally test isolation transformers configured the way I advocate and which is the standard practice in places where the NEC doesn't apply, and "the NEC way" of bonding the secondary of isolation transformer.

You may find the results..... shocking.

Now you can argue that this is all about marinas and pools, but people holding onto SM58s and guitars with cables to amplifiers are well connected too.  So is a person at a festival in bare feet that touches some metalwork that is "grounded" and gets shocked, because there is a potential difference between the soil and the earthed metalwork.

Ground-to-ground shocks are insidious and evil, because no standard circuit protection device detects them.

The places I worry about G2G shocks are farms, bodies of water, and rock and roll.  All have the propensity to catch the unwitting.  And for farms its also an economic issue; tingled cows produce less milk. 

[Jean-Pierre Coetzee]

Ground "bonding" is all about providing a low impedance path that can carry enough fault current to open fuse/breakers without allowing enough voltage rise to be a hazard itself.

I've had to redesign (beef up) PCB grounds inside a product to meet UL "bonding" criteria.

The joker in the deck is RPBG that can turn a safety ground into a shock hazard.

JR

JR I'm not sure if it's just a terminology thing but I understand bonding as connecting all conductive materials back to the primary ground to provide a safe path away from equipment for lighting to pass to earth and grounding to be the safety ground in an electrical circuit which would be used for what you described.

From what I have learn't faults always want to return to the source(ie the electrical service) where lighting or other ground potential differences always want to return to earth so as to be neutralised and the bonding / ground differentiation is the distinction between these two shock hazards.

This would be the reason for the separately derived system to be grounded at the transformer which I believe would be to code as David's linked article shows and therefore the incoming ground and the separately derived ground do not need to be bonded together.

This should theoretically be safe near water where the potential difference between two earth points would be low but could be a shock hazard in dry areas where the difference could potentially be quite high between two earth points .

So in the original article the grounds would need to be bonded together unless the op was going to dig in a ground rod every single time he deploys the transformer and even then there could be a potential shock hazard should the two different ground come into contact if that is even up to code.

[Mike Sokol]

The places I worry about G2G shocks are farms, bodies of water, and rock and roll.  All have the propensity to catch the unwitting.  And for farms its also an economic issue; tingled cows produce less milk.

There's been a ton of experiments about how low current shocks (a few mA) will reduce milk production on dairy farms. Here's the first one I've read that states the internal resistance of cows being around 500 ohms. So for farm "stray voltage" tests they use a 500 ohm resistor to read current flow. https://msu.edu/user/hillman/elecshok.htm

Since human beings are around 1,000 to 1,500 ohms resistance hand-to-hand (once you have enough voltage to punch through the non-linear epidermis skin layer) perhaps we should be using a 1,500 ohm resistor between gear and ground to measure fault current rather than high-impedance circuit voltage. While a high-impedance voltage test will certainly indicate a failed ground-bond, it won't necessarily indicate an immediately dangerous electrocution condition. So I'm thinking a two-part test. First a high-resistance one using a High-Z DMM (many megohms), and if a voltage is indicated, then a second low-resistance test with a 1k ohm resistor inline with the meter set to mA range. That way you could evaluate the immediate danger of the shock condition. A few mA through a 1,000 ohm load would be a tingle. From 5 to 10 mA would be a substantial shock hazard. And 30 mA on up to 120 mA would indicate a potential electrocution hazard. And yes, if you read 120 mA through a 1,000 ohm load that indicates a hard line-to-chassis short or an RPBG miss-wired outlet.

Now I'm not suggesting that you don't do anything to correct a 5 to 10 mA or even 1 to 5 mA fault. Because those low currents show there's a failure in the EGC, those small currents can easily go to deadly levels at any time. So they should be corrected immediately. But this sort of test would lend credibility to our diagnostics.

[John Roberts {JR}]

Have a look at this IEEE paper: Boat Dock Exposure Voltage Mitigation (PDF).  They experimentally test isolation transformers configured the way I advocate and which is the standard practice in places where the NEC doesn't apply, and "the NEC way" of bonding the secondary of isolation transformer.

You may find the results..... shocking.

Now you can argue that this is all about marinas and pools, but people holding onto SM58s and guitars with cables to amplifiers are well connected too.  So is a person at a festival in bare feet that touches some metalwork that is "grounded" and gets shocked, because there is a potential difference between the soil and the earthed metalwork.

Ground-to-ground shocks are insidious and evil, because no standard circuit protection device detects them.

About a year ago I built a prototype backline safety power drop...  I put a 3 pole relay in front of a GFCI outlet. This approach combined multiple safety strategies.

One was to sense the current flowing in the ground lead and disconnect power and ground if the ground current exceeds several mA.

The relay winding was powered from the output side of the GFCI so any GFCI fault would also disconnect ground.

The start-up circuit used a touch contact that wouldn't even connect power if outlet was miswired.

I abandoned this project as too expensive for a penny pinching musician market, and too hard to convince UL that is was a useful strategy.

The places I worry about G2G shocks are farms, bodies of water, and rock and roll.  All have the propensity to catch the unwitting.  And for farms its also an economic issue; tingled cows produce less milk.

I hope shocks from live performances are diminishing as more bands use wireless mics, and older (faulty) guitar amps get retired.

JR 

[John Roberts {JR}]

JR I'm not sure if it's just a terminology thing but I understand bonding as connecting all conductive materials back to the primary ground to provide a safe path away from equipment for lighting to pass to earth and grounding to be the safety ground in an electrical circuit which would be used for what you described.

Perhaps and I am not the last word on terminology.

I learned the term ground bonding in connection with product design to pass UL safety testing. In that context bonding was pretty much as described... the ability to pass tens of amps current in a ground path with low single digit voltage drop, long enough to trip fuse/breakers. It seems there is a similar current carrying requirement in mains wiring external to products. 

From what I have learn't faults always want to return to the source(ie the electrical service) where lighting or other ground potential differences always want to return to earth so as to be neutralised and the bonding / ground differentiation is the distinction between these two shock hazards.

Yes, while lightning mitigation (earthing?) is probably more about preventing an up-stroke than dealing with the tens of thousands of amps in a direct hit down-stroke. Once that much current is moving typical size wires are pretty much just a suggestion. In the earth is where lightning ultimately ends up, via the lowest resistance path it finds. Sometimes ionized air is lower Z than wires. 

This would be the reason for the separately derived system to be grounded at the transformer which I believe would be to code as David's linked article shows and therefore the incoming ground and the separately derived ground do not need to be bonded together.

This should theoretically be safe near water where the potential difference between two earth points would be low but could be a shock hazard in dry areas where the difference could potentially be quite high between two earth points .

So in the original article the grounds would need to be bonded together unless the op was going to dig in a ground rod every single time he deploys the transformer and even then there could be a potential shock hazard should the two different ground come into contact if that is even up to code.

This is beyond my personal experience and I am still digesting the IEEE paper, so I won't muddy these waters further..

JR

[Stephen Swaffer]

So I have taken the opportunity to cut and paste the post I origianlly responded to by saying it was a bad idea to advocate not tying grounds together on the web, where people might not consider all of the other conditions that need to be met to be safe:

"If the negatives (like weight) don't make this unrealistic, then yes, a transformer with a locally derived neutral (and thus ground) is by far and away the best solution for a PA distro. 

You're in the USA, so you need to be aware of the NEC: what you're building the NEC calls a separately derived system, and 250.30(A).1 requires that you bond the ground you've magic'd from the secondary of your transformer to the incoming ground.  This is a f**king stupid rule, and does much to destroy the value of isolating transformers, and has particular relevance to cow sheds, and boats and docks.  But RnR where one is often presented with a dodgy ground connection are another example where the need for this bond makes the situation less safe than it would be without the bond.

GFCIs on output circuits recommended."

The IEEE article lists proposed code changes that enumerate 5 conditions that must be met in order for this to be an acceptable practice.  These code changes also apply specifically to boat dock conditions.  I am surprised that they do not specificlly prohibit metal objects being adjacent to both systems-but that is somewhat mitigated by requiring a GFCI on the input side of the transformer.

I possibly see some value in this-but it also requires a non metallic sheathed wiring method to keep the 2 systems and I am isolated still digesting the IEEE paper as well.  But it sill requires everything on the boat dock to be bonded to that systems ground-which is bonded to the neutral for that system.  I am still trying to wrap my head around how this would work in a stage situation.

I still don't see how safety is compromised comparatively by "continuing the ground through"?

[ProSoundWeb Community]