Fail-Safe Galvanic Isolator vs. Isolation Transformer

[GatorDontPlay]

I've seen a few threads of various ages discussing both GI and ITs, and referencing Nigel Calder and also SmartGauge Electronics - Isolation Transformers or Galvanic Isolator? 1/2

I've looked at both of those references, but they both seem to predate the 2008 version of ABYC A-28 that defines Fail-Safe GIs (Calder's last version was 2005, and the Smart Gauge site talks about open circuit faults, not mentioning Fail-Safe GIs at all).

I'm trying to decide what to add to a boat that currently has one 30A circuit (but I've contemplated adding a second, or changing to a 50A circuit before the winter) and no GI or IT. My options seem to be the ProSafe FS galvanic isolator, for which I can cover the 30A, 30A X 2 = 60A, or 50A scenarios with one 60A GI for under $300 (30A only for $200), or the Charles Industries ISO-G2 for around $600 (covering only a single 30A scenario; no good 50A option, and doing 2 X 30A would entail a second $600 ISO-G2 for a total of $1200 spent on isolation, vs. $300 for the ProSafe).

Given the Fail-Safe advancements in the state of the art for GIs, is it still worth the 3X-4X cost for an IT? The main difference is that if it fails (closed) I'll lose galvanic protection, but as long as I'm testing regularly and checking the anodes, it seems like that risk can be managed (and is certainly no worse than not having a GI in the first place).

[goboatingnow]

Fundamentally GI can not offer the same protection as IT. There are many scenarios where GIs can conduct and allow currents to flow. An IT completely separates shore ground from boat ground , eliminating current corrosion from fault shore ground wiring.

Dave

[GatorDontPlay]

Calder says "Galvanic isolators that have a capacitor and comply with ABYC standard A-28 will do an effective job of safely blocking galvanic currents in most circumstances."

I know (and understand why) an isolation transformer would be preferred... but if a galvanic isolator is 1/4 the price (saving me close to $1k), is effective in most circumstances, and I spend most of the year in freshwater (where galvanic corrosion is less)... wouldn't the $1k be better spent on something else?

[goboatingnow]

Originally Posted by GatorDontPlay

Calder says "Galvanic isolators that have a capacitor and comply with ABYC standard A-28 will do an effective job of safely blocking galvanic currents in most circumstances."

I know (and understand why) an isolation transformer would be preferred... but if a galvanic isolator is 1/4 the price (saving me close to $1k), is effective in most circumstances, and I spend most of the year in freshwater (where galvanic corrosion is less)... wouldn't the $1k be better spent on something else?

good answer, you pays your money and you takes your chances.

from Smartgauge site
"For this reason, some galvanic isolator manufacturers have put a capacitor inside the galvanic isolator so that the capacitor conducts the AC currents thus preventing the diodes from conducting. This idea would work, if the capacitor was the correct size.
However, whilst this could be done to cope with the currents produced by one piece of equipment, it cannot be done for those produced by, say, 15 separate items installed round the boat. For technical reasons, there is a limit to the size that this capacitor can be. And it would be necessary to exceed this limit in order to keep the AC voltage below the "conduction" voltage of the galvanic isolator. Unless the AC voltage is held below this level, there is no point in fitting the capacitor to the galvanic isolator. It will not achieve anything.
The only conclusion is that if an installation has enough of this "modern" equipment to cause AC voltages across the galvanic isolator that are sufficient to cause it to conduct then the galvanic isolator is not doing it's job properly."

see SmartGauge Electronics - Isolation Transformers or Galvanic Isolator 2/2 for the whole story on GI and IT


Dave

[colemj]

If you are not in a marina much, then the GI will save money and weight and not have an appreciable risk over a IT.

Mark

[GatorDontPlay]

" For technical reasons, there is a limit to the size that this capacitor can be."

It would be nice to see some of the technical reasons and math behind what seems to be an assertion that all galvanic isolators produced today are useless on boats that have an array of modern electronics. I tend to trust Calder and ABYC more than that dated website.

[Andina Marie]

Over the last 15 years we have sold over 20,000 galvanic isolators, all on unconditional warranty. To date we have not had a single warranty return. We do have some fail safe features but I would never claim it is fully fail safe, give me any galvanic isolator and I can make it fail UNsafe.

We don't just type test, every one we make goes through the ABYC tests including running at 70+ amps AC until the temperature stabilizes.

Very few installations need the capacitor. We sell a capacitor separately if you need it but we've only sold about 50 in the last 15 years. Install the GI first and then measure the AC voltage across it, if it is more than about 0.7 volts, first try to track down which item has the leakage and correct it. With accumulated salt over many year some older boats cannot be corrected in which case a capacitor is needed. Our capacitor will keep AC voltages below 0.5 volts with AC currents up to 5 amps. Compare that to galvanic isolators with built in capacitors if you can get them to quote you any numbers .

[Andina Marie]

Originally Posted by GatorDontPlay

" For technical reasons, there is a limit to the size that this capacitor can be."

It would be nice to see some of the technical reasons and math behind what seems to be an assertion that all galvanic isolators produced today are useless on boats that have an array of modern electronics. I tend to trust Calder and ABYC more than that dated website.

For the math behind capacitors on Galvanic Isolators, see the article "Does My Galvanic Isolator Need a Capacitor to Conduct AC?" at http://www.yandina.com/GIsolCap.html

The reason is financial not technical.

[transmitterdan]

Originally Posted by Andina Marie

The reason is financial not technical.

Exactly right!

[GatorDontPlay]

Andina Marie (Yandina):

Does your GI pass the ABYC A-28 test in Appendix C / 28.13.5, without a capacitor, as tested by an independent laboratory (e.g. UL)? If not, it doesn't meet ABYC standards as advertised, even if connected to a remote monitor (which it looks like you don't supply either, and would add extra expense and hassle that would again make me lean towards the ProSafe FS GI).

Also, from your page on the reasoning behind the capacitor (http://www.yandina.com/GIsolCap.html):

"What is acceptable leakage? My philophosy is you shouldn't be trying to cover up a problem by conducting it to ground. I maintain you should check every item on your AC supply and if any show any leakage at all they need to be repaired. That is a safer alternative to just allowing the leakage to flow to ground.

You can easily determine if you have any AC leakage - take an AC voltmeter and put it across your galvanic isolator. The reading should be less than about 0.25 volts. If it is not, start turning off breakers until you track down the culprit and repair it."

What if the culprit is not on your boat? Per Calder, "it is extremely common to find AC leakage voltage on dockside grounding wires that exceeds 1.0 volt AC." Then it's outside of your control, regardless of the age or saltiness of your boat.

You also say:

"Our capacitor will keep AC voltages below 0.5 volts with AC currents up to 5 amps. Compare that to galvanic isolators with built in capacitors if you can get them to quote you any numbers"

Again, is that confirmed by an external test lab on a repeatable setup? Seems like you're creating your own standard there -- don't know why any other manufacturer would test to that rather than Appendix C of A-28, which specifies the test circuit, AC leakage of 3 Amps RMS, simulated galvanic potential of 0.5V, and maximum DC current allowed through the GI of 5 mA.

This is fascinating that I've gotten two opposite answers to the question I posed -- goboating now says that even ABYC A-28 (2008) compliant fail-safe GIs are mostly useless and you have to go with an IT, while Andina Marie seems to say that a simple 4-diode GI with no capacitor is sufficient. Which is making me think of Calder's preface to his chapter on the topic:

"Corrosion and grounding are two of the hardest issues to get a handle on in the boating world, not the least because the science is often incomplete, disagreement sometimes arises among the experts, and all kinds of self-anointent 'experts' are pushing their own theories. . . I am essentially relying on positions developed by the ABYC over many years for two reasons:

1. I believe these positions are well grounded in the available evidence and represent a consensus position

2. The nature of the standards development process within the ABYC is such that it responds in a measured (although sometimes rather slow) manner to emerging science, and as such, at any given time ABYC standards form a sound basis for installation decisions on boats."

Makes sense to me.

[transmitterdan]

Originally Posted by GatorDontPlay

I've seen a few threads of various ages discussing both GI and ITs, and referencing Nigel Calder and also SmartGauge Electronics - Isolation Transformers or Galvanic Isolator? 1/2

Phrases that include absolutes like "cannot" in papers like this should invoke some skepticism. It is possible to put a properly sized capacitor across the diodes of a GI. But that costa money and in my opinion does not make them fail-safe.

Also I'm not sure I believe the premise that there are lots of devices that shunt line (L wire) current to earth (E wire). If that were so then ground fault devices could not work. And the purported interference reduction benefit of such shunting is dubious. If RF is shunted to the earth wire then it will surely radiate and thus defeat the purpose. Why would a designer shunt to E when N is available and is in parallel with the E wire? I have seen capacitor shunting from L to N and N to E but that in and of itself does not create leakage current. So I am also skeptical of this aspect of the article.

To answer your original question, I would go with what Calder recommends.

[Saltyhog]

Calder etc. recommend following ABYC. That is one line of thinking for electrical safety and impressed current corrosion prevention. There are other lines of thinking on the subject.

If planning on adhering to ABYC, pick either an isolation transformer or a galvanic isolator. You have the answer as to which provides better safety. If your boat is big enough and will support the size and weight, Id' go with an IT. If you're like me though, the IT is just going to be too big and heavy for the light use it will see. Note that an IT can also be used to convert local power voltage to your boats native voltage. Won't convert frequency however and that can be a problem with things like air conditioners.

The galvanic isolators from Yandina seem to be a good value. You'd need to spend about what they charge just for the components to make one yourself. I did look into that option. Yandina is careful to never claim ABYC compliance. I think it's a quality GI and I know of lot's of folks who are happy with them. They should be tested occasionally and for correct operation

If you need to feel better about ABYC compliance, I believe this is a very good ABYC compliant (IE failsafe) galvanic isolator

To really add some safety, consider an RCD or ELCI. There are those who argue that with a whole boat RCD, no connection between green wire safety ground and vessels DC ground or underwater bonding is needed, and therefore no galvanic isolator would be needed. (What would it be isolationg?) I believe this is currently acceptable in the EU, and frankly I'm beginning to think it makes the most sense. I hope the idea spreads this way (US) but I'm not holding my breath.

[GatorDontPlay]

Thanks for the thoughts, Saltyhog.

It does look like Yandina claims ABYC compliance for their Galvanic Isolator:

Galvanic Isolator Data Sheet

"Meets ABYC standards. (When fitted with a remote monitor, not included)"

DEI looks like a good product, from what I can tell roughly equivalent to the ProSafe but perhaps a bit more expensive.

Calder mentions the RCD (equating it to GFCB, GCFI, and RCCB terms; doesn't use the ELCI term but sounds like the same thing also) but specifically writes (in bold all caps italics):

"it is a BAD IDEA to break the AC-to-DC ground connection" just because you have a RCD/GFCB/GFCI/RCCB/(ELCI?) installed -- as you
are then dependent on said electrical device working (and they apparently have a failure rate in the teens, mostly due to vulnerability to lightning strikes) in order to be safe for swimmers (the trip point, whether 5 mA per US requirement or 30 mA for Europe, is also a safety question). Also says attempting whole-boat protection (as opposed to at each particular AC outlet/device) is likely to cause a lot of nuisance tripping.

[Andina Marie]

We passed all ABYC specifications when we first started manufacturing our Galvanic Isolator. We have not updated the design to meet the latest revisions regarding a capacitor or remote monitoring (and can sell for $100 less). With over 20,000 sold on unconditional warranty and zero returns, our track record indicates their reliability. They were tested independently by West Marine and sold under their brand name for about 10 years.

We do have fail safe functions in our GI but I'm not going to say it is 100% fail safe because NONE of the GIs are fail safe. I can cause any GI, including ours, to fail "unsafe" however in normal installations with a circuit breaker on the shore power, we are fail safe.

The CAPACITOR is rarely needed. You quoted AC leakage on docks up to 1 volt but leakage ON THE DOCK has nothing to do with the Galvanic Isolator on a boat. It is AC leakage ON THE BOAT that flows down the ground cable and attempts to flow through the GI. The AC voltage from the leakage is added on every half cycle to the DC voltage from electrolytic sources and if the total of the two exceeds 1.2 volts then some DC current flows through the diodes on the GI. As the AC voltage gets higher, a higher percentage of the time DC will sneak through on top of it, decreasing the GI performance up to a maximum of 50%. You can easily determine if this is a problem by measuring the AC voltage across your GI. We've only sold about 50 capacitors in 15? years so it is rare that they are needed. I've not checked lately but the capacitor used to be an option in the ABYC specifications.

If you have 3 amps of AC leakage you have a serious problem with some of the equipment on the boat. You should be eliminating the cause of the leakage rather than trying to bypass it across the GI. However if you can't, our Capacitor can carry 5 amps and maintain AC voltage below 0.5 volts, a 66% increase over the ABYC specification. I'm not going to submit it for independent evaluation, selling only 2 or 3 a year would not pay for the testing.

We are currently in production of a Double Galvanic Isolator. It is the equivalent of two GIs in series to give 2.4 volts isolation. This not only helps protect aluminum boats and boats using magnesium anodes but doubles the rejection of AC leakage.

[Saltyhog]

Like I said, Calder and ABYC agree on the bonding of AC green wire protective earth and DC ground. I'm only saying there are other views and there is at least *some* controversy on the subject. There are no ABYC police. It's not like UL or IEC. In the EU, ISO 13297 specifically permitted not connecting green wire and DC ground if a whole boat RCD is fitted. Also, your quote of Calder mentions 5ma trip current in the US. No longer true as ABYC now falls in line with EU with 30ma trip current RCDs. This increases the reliability of the the RCDs. I do agree that the risk of an RCD failure and thus not tripping is not to be taken lightly. I general, for safety I much prefer passive components to active. An RCD is active and therefore must work to provide protection. A wire is about as passive as you get electrically. There, I just talked myself back into connecting the green wire to the DC ground