Galvanic Isolator

[Liam]

I have been putting off this thread for a bit but I thought I'd bite the bullet and just go for it. After all the talk about them my Dad wants to buy a Galvanic Isolator for his boat but knows absolutely sod all about them. I've heard about them on here and only have a very vague idea on how they work as I have always kept clear of the topics in the past...So here it goes... I am looking for a suitable unit for a 50ft Narrowboat with 4 135Ah batteries, twin alternators, 1500W Inverter and landline hookup. What would you suggest?

 

Please keep this one on topic!

Edited November 24, 2008 by Liam

[Keeping Up]

There is a high-spec unit which has 4 diodes instead of two and has LEDs which light up if the performance is being comprimised by stray currents. Unfortunately I can't remember which this one is, I'm sure somebody will. It is well worth buying such a unit, because the extra diodes will improve the isolation against stray earth currents that may be generated on shore or by other people's boats.

 

I won't repeat all the technical bits, but we did a fairly comprehensive set of experiments and measurements when the discussions were raging last time. We demonstrated beyond reasonable doubt that there is actually no basis in reality for the often-raised fear that stray currents produced by your own on-board equipment may disable the effectiveness of your own isolator. There is however a significant risk that other people's stray currents can enter your boat through the earth-lead and negate the effectiveness of your isolator; hence the recommendation to use the 4-diode unit.

[WJM]

I dont think it really matters much what is on your boat. The only thing that matters is that you are going to connect your boat to the earth via the electricity supply earthing connection. The problem with doing that is when connected, your boat will slowly dissolve. To prevent this a Galvanic Isolator will break your earth connection and only rejoin it when a big surge of current tries to get through, as will only happen when a fault occurs. Some GIs do that invisibly, while others have little lights to indicate their function. I dont think it really matters which one you buy as long as it does the job. They are very basic with no moving parts, so I am tempted to say they are all very much the same regardless of what you spend on them. I paid about £70 for mine.

 

I have heard that some modern devices like washing machines and computers will negate the protection of a GI and you then need to consider an isolation transformer. If you are not a live-aboard, then the best strategy is to only plug the landline in when you need it, then transformers and isolators aside, the boat just does not have much opportunity to dissolve.

[churchward]

I fitted a Sterling unit to my boat after a recommendation for our Survey Engineer. However, If I was as to do it again I would opt for a unit that had an external indicator to show any stray current attempting to flow through the earth.

 

an example of one like that can be found here

 

http://www.marinemegastore.com/product.asp...at_id=ECLSWI004

[chris w]

There is a high-spec unit which has 4 diodes instead of two and has LEDs which light up if the performance is being comprimised by stray currents. Unfortunately I can't remember which this one is, I'm sure somebody will. It is well worth buying such a unit, because the extra diodes will improve the isolation against stray earth currents that may be generated on shore or by other people's boats.

 

I won't repeat all the technical bits, but we did a fairly comprehensive set of experiments and measurements when the discussions were raging last time. We demonstrated beyond reasonable doubt that there is actually no basis in reality for the often-raised fear that stray currents produced by your own on-board equipment may disable the effectiveness of your own isolator. There is however a significant risk that other people's stray currents can enter your boat through the earth-lead and negate the effectiveness of your isolator; hence the recommendation to use the 4-diode unit.

Actually they all have 4 diodes in them. Two diodes in series, in parallel with two diodes in series facing the opposite direction.

 

It wouldn't work as a GI with two diodes (ie: just one each way). Galvanic voltages can be no higher than 0.8v if one considers the physics of what is going on. A single diode has a forward voltage drop of 0.6v which is not enough, under worst case conditions, to block all galvanic action. Two diodes have a forward drop of 1.2v which is larger than any possible galvanic voltages and so are perfectly suitable.

 

The unit needs two diodes in each direction, to ensure that you are not corroded nor do you corrode anyone else (and to ensure that it is irrelevant which way round they are fitted).

 

To preclude their efficacy being affected by possible stray ac earth currents from chargers etc, mine is fitted with a bypass capacitor (25,000uF bipolar)which routes all such ac currents around the GI whilst blocking all the dc galvanic currents.

 

Although the ones with LEDs are useful, in that you can see if the GI is being affected by stray ac currents, it doesn't help solve the problem as the ac stray currents will still be there. The bypass capacitor will solve this issue and remove the need for LEDs as there will never be an issue to indicate. In the USA, in order to meet the ABYA specifications (one of the governing bodies), a GI MUST be fitted with a bypass capacitor.

 

Chris

[WJM]

"...knows absolutely sod all about them. I've heard about them on here and only have a very vague idea..." "What would you suggest?"

 

 

A plain one, one with lights or one with a dial - take your pick.

[Liam]

A quick search on the old internet has come back with one with the following spec... It's got the indicitator at the front with a simple "Fail" and "Pass" on the front... Does that one look OK? £93.50.

 

 

 

GALVANIC ISOLATOR with Indicator For use with 120/240V 50//60Hz 16A supplies Correctly installed the galvanic isolator will block galvanic currents in the shore earth connection up to 1.1V. Above this level current is allowed to pass maintaining the safety earth connection. For almost every combination of metals likely to be found at a mooring the 1.1V maximum is sufficient to block corroding currents. When the voltage at the earth connection is greater than 1.1V, current is allowed to flow through the galvanic isolator potentially corroding submerged metals on the boat. In this situation the shore power should be disconnected to avoid hull corrosion. The meter shows PASS – the voltage is below 1.1V, and FAIL – the voltage exceeds 1.1V. When the indicator is within the GREEN zone the galvanic isolator is blocking D.C. galvanic current. D.C. galvanic current can also be caused by other shore earth connections such as telephone and cable television.

[chris w]

A quick search on the old internet has come back with one with the following spec... It's got the indicitator at the front with a simple "Fail" and "Pass" on the front... Does that one look OK? £93.50.

It will do the job perfectly, with the proviso that IF the meter shows a high level of stray ac currents (from chargers et al), there is nothing you can do about it. ie: if the needle swings over, what are you then going to do? You are no longer galvanically protected if the needle goes above 1.1 volts.

 

That's why I still strongly advocate getting one with a bypass capacitor like this one: click here . Scroll down to the 30A version with capacitor. It's a Sterling GI.

 

Although more expensive (£113) it does have the bypass capacitor. It's not much more to pay (£10) against the value of your boat. The Adverc GI also has a bypass capacitor. click here

 

Here's a short article from the US magazine "Yachting Monthly" click here which discusses capacitors in GI's. However, a word of caution, the author's reference to "So far as I know all galvanic isolators currently sold do have a capacitor" is only true in the USA and NOT in Europe. Most don't over here because it's not a governing body requirement whereas it IS in the States.

 

Chris

Edited November 24, 2008 by chris w

[Keeping Up]

Actually they all have 4 diodes in them. Two diodes in series, in parallel with two diodes in series facing the opposite direction.

 

It wouldn't work as a GI with two diodes (ie: just one each way). Galvanic voltages can be no higher than 0.8v if one considers the physics of what is going on. A single diode has a forward voltage drop of 0.6v which is not enough, under worst case conditions, to block all galvanic action. Two diodes have a forward drop of 1.2v which is larger than any possible galvanic voltages and so are perfectly suitable.

 

The unit needs two diodes in each direction, to ensure that you are not corroded nor do you corrode anyone else (and to ensure that it is irrelevant which way round they are fitted).

 

To preclude their efficacy being affected by possible stray ac earth currents from chargers etc, mine is fitted with a bypass capacitor (25,000uF bipolar)which routes all such ac currents around the GI whilst blocking all the dc galvanic currents.

 

Although the ones with LEDs are useful, in that you can see if the GI is being affected by stray ac currents, it doesn't help solve the problem as the ac stray currents will still be there. The bypass capacitor will solve this issue and remove the need for LEDs as there will never be an issue to indicate. In the USA, in order to meet the ABYA specifications (one of the governing bodies), a GI MUST be fitted with a bypass capacitor.

 

Chris

 

Absolutely right Chris, my mistake, what I meant to say is that there a unit with four diodes IN EACH DIRECTION. It was one of these which advertise a 2.6 volt turn-on voltage. With one of those, I do not believe that the capacitor is required (that's just my opinion, not a cue to start a debate )

 

A quick search on the old internet has come back with one with the following spec... It's got the indicitator at the front with a simple "Fail" and "Pass" on the front... Does that one look OK? £93.50.

 

I have reservations about the ones with a meter. I would imagine that the meter would be able to pass a small DC current at all times, in which case it partially negates the benefit of the isolator. The LED types, on the other hand, would not pass any DC under normal circumstances.

[chris w]

Absolutely right Chris, my mistake, what I meant to say is that there a unit with four diodes IN EACH DIRECTION. It was one of these which advertise a 2.6 volt turn-on voltage. With one of those, I do not believe that the capacitor is required (that's just my opinion, not a cue to start a debate )

The reason for having 4 diodes in each direction is not to preclude the need for a capacitor Allan but so that LEDs can indeed be used as an indicator.

 

This is because LED's have a forward voltage drop of around 2 volts and so 4 diodes are needed each way to ensure the diodes' total voltage drop (which will be around 2.4v total) is greater than the LED's forward voltage drop. Otherwise the LED indication wouldn't work.

 

Chris

[Keeping Up]

The reason for having 4 diodes in each direction is not to preclude the need for a capacitor Allan but so that LEDs can indeed be used as an indicator.

 

This is because LED's have a forward voltage drop of around 2 volts and so 4 diodes are needed each way to ensure the diodes' total voltage drop (which will be around 2.4v total) is greater than the LED's forward voltage drop. Otherwise the LED indication wouldn't work.

 

Chris

That is very probably the reason they did it. However it turns out that it increases the effectiveness of the isolator, and they naturally emphasise that virtue in the description. It strikes me as a double bonus, not one to be discounted just because they did it for the wrong reason in the first place.

[chris w]

That is very probably the reason they did it. However it turns out that it increases the effectiveness of the isolator, and they naturally emphasise that virtue in the description. It strikes me as a double bonus, not one to be discounted just because they did it for the wrong reason in the first place.

I agree with your logic and it will of course give added protection up to 2.4v instead of 1.2v. Galvanic corrosion will only ever account for 0.8v so the 4 diodes in series give an additional headroom of about 1.6v as opposed to the two diodes in series additional headroom of about 0.4v. However, who knows how large a stray ac voltage might be on the earth line?

 

Chris

[Lesd]

I have the Adverc unit Chris mentioned above, just noted this from their website which may be useful advice for GI users;

 

GALVANIC ISOLATOR TEST.

 

NOTE: It is recommended that this device is tested (as shown below) every six months and after any RCD trip.

 

1. Unplug any connections to the shore power before commencing the test.

 

2. Remove both leads from the Galvanic Isolator.

 

3. Using a multimeter, set to the diode test function Diode Test, test continuity across both terminals using positive and negative probes.

 

4. Note the reading on the display.

 

5. Short both terminals with a metal object such as a small screwdriver, to discharge an internal capacitor.

 

6. Repeat 3, with the multimeter probes reversed.

 

7. Note the reading on the display. This should be the same as 4 give or take a few digits e.g. 1032-1029, 783-786.

 

8. Re-connect the original earth leads to the Galvanic Isolator terminals, as before.

The readings can vary, depending on the multimeter used. For a successful test, it is essential that both readings, with the probes reversed, are virtually the same.

[Gibbo]

A quick search on the old internet has come back with one with the following spec... It's got the indicitator at the front with a simple "Fail" and "Pass" on the front... Does that one look OK? £93.50.

 

If you have an RCD on board protecting the entire system then that should be ok. If you don't have one then avoid it.

 

The reason being that a few amps of earth leakage due to a fault will melt that.

 

Think about a 3 amp fault going up/down the earth conductor...... 3 amps * 1.4 volts = 4.2 watts. That is one HELL of a lot of heat in a plastic case and with no metal case or fins it will overheat.

 

Proper GIs are supposed to be able to handle the full fuse current indefinately. That's why they are huge metal things with heatsinks on them.

 

But as I said, with an RCD that should trip instead.

 

But you still have the problem of AC currents making the GI conduct and thus useless. A capacitor can help in this regard but there is no certainty. Despite what others might tell you.

 

Gibbo

[blackrose]

I have the Safeshore (70amp) GI that Allan mentioned. It has duel led status indicators which show whether there is any DC or AC leakage (1 led = DC leak, 2 leds = AC leak).

 

Chris, if you have a GI with a capacitor how does it work? I assume that if the diodes are forced into circuit by a fault condition then the capacitor takes over, but what does it do exactly and what is the advantage? Also is there any way of putting a capacitor onto my GI?

Edited November 24, 2008 by blackrose

[Liam]

Thanks for all the replies so far, but...

 

How do I find out if we have an RCD?

 

Gibbo, you've lost me completely with the maths and the electrickery talk... I know how to wire a plug if that helps!

[chris w]

I have the Safeshore (70amp) GI that Allan mentioned. It has duel led status indicators which show whether there is any DC or AC leakage (1 led = DC leak, 2 leds = AC leak).

 

Chris, what's the advantage of your GI with the capacitor? How does it work?

The diodes in any GI will totally block any galvanic currents. No question - job done. However, there can be other (stray) currents on the earth lead notably from any piece of equipment that uses "switched-mode" technology like modern chargers and inverters. An amount of low and high frequency ac current can flow down the earth lead and through the GI.

 

If the voltage of these stray currents exceeds 1.2v (or 2.4v for your Safeshore model) then the diodes in the GI will switch on and act like a piece of wire...ie: they will block nothing, not even the galvanic currents, rendering your GI totally useless (but not damaging it).

 

This of course is what they will do anyway if a large fault current flows through the GI. The GI acts as a switch - blocking very small dc currents (galvanic currents) but allowing large currents to flow to earth as a normal earth wire would do. However, you do not want the GI to switch on due to small stray ac currents but only big fat fault currents.

 

A capacitor is an electronic component that allows ac currents to flow but blocks dc currents. (How it does this, is not relevant at this stage). So by fitting a capacitor across the GI, any stray ac currents will bypass the diodes and not switch them on thus ensuring the GI carries on totally blocking any galvanic currents.

 

Although having a meter or LEDs will tell you whether there are any stray ac currents or not, they will not help you to get rid of them. What they will tell you is that you have, at that moment, a GI that cannot do its job. A capacitor on the other hand will get rid of all ac stray currents.

 

To make reference to Gibbo's comment above, viz: "A capacitor can help in this regard but there is no certainty" It is important that the capacitor, and its connecting leads, are chosen carefully or otherwise the capacitor will also have what is known as "inductance" a property which actually blocks high frequency ac currents - the opposite of what we want the capacitor to do.

 

The astute amongst you may have thought, "well, why have a GI at all, why not just have a big capacitor?" The reason is that the capacitor in the GI needs to be an extremely large value (electronically speaking) in order to allow low frequency (50Hz) currents to pass but it only ever has to operate at 1.2v (the forward voltage drop of the two diodes). If the voltage exceeds that, the diodes switch on and act like a piece of wire carrying any dangerous currents to earth. If we only had a capacitor and no GI, the capacitor would need to be able to handle the full PEAK mains voltage of 339v plus some reserve (say 400V). That together with the large value needed, would mean that the capapcitor would be enormous and would need to be towed behind on a butty!!!!

 

Chris

Edited November 24, 2008 by chris w

[blackrose]

Thanks for all the replies so far, but...

 

How do I find out if we have an RCD?

 

Gibbo, you've lost me completely with the maths and the electrickery talk... I know how to wire a plug if that helps!

 

Have a look on your consumer unit (your mains panel), you should see either an RCD or RCBO trip switch.

[chris w]

Have a look on your consumer unit (your mains panel), you should see either an RCD or RCBO trip switch.

Although don't mistake an MCB (miniature circuit breaker) for an RCD. The devices are different but similar looking. An RCD will have a "test button" with it. If it has no test button, it is an MCB.

 

The RCD trips on earth faults, even tiny currents like 30mA flowing between live and earth, the MCB trips on excessive currents, like more than 16A, flowing between live and neutral.

 

Chris

[Liam]

Well there's no mains panel as such... the way we have it at the moment is that when we're out on the boat, we switch our inverter on, which the switch box plugs in to. A cable then runs from this switch around the boat. When we're on the moorings, we unplug the cable from the switch to the interverter and simply plug this in to the moorings through an outdoor extension lead.

 

I am crap with electrics so although I amazed at everybodys knowledge and am grateful for the help, it doesn't mean a thing to me as I don't understand it.

[chris w]

Well there's no mains panel as such... the way we have it at the moment is that when we're out on the boat, we switch our inverter on, which the switch box plugs in to. A cable then runs from this switch around the boat. When we're on the moorings, we unplug the cable from the switch to the interverter and simply plug this in to the moorings through an outdoor extension lead.

 

 

Can you tell me where you are moored and the name of the boat to ensure I never come on board. Liam... what you have described is very dangerous. You MAY be safe when you plug into the shore ASSUMING the shore point has an RCD. However, if your AC earth is not connected to the hull then even that won't always protect you.

 

Your inverter, without an RCD on board, is an accident waiting to happen. You must get someone who knows what's what to make it safe. Inverters can kill you just like real mains voltage.

 

Chris

Edited November 24, 2008 by chris w

[Liam]

It's not just that basic system, there is other stuff but thats the only way I describe it. I know enough to switch things on and off, not the whole system in detail.

 

I think I have a picture of all the electrics on the engine room bulkhead which I'll post when I get home.

[chris w]

It's not just that basic system, there is other stuff but thats the only way I describe it. I know enough to switch things on and off, not the whole system in detail.

 

I think I have a picture of all the electrics on the engine room bulkhead which I'll post when I get home.

Even if your "switch box" incorporates an RCD (does it?) this RCD WILL NOT FUNCTION unless the inverter earth and neutral are bonded togethet at the inverter and also bonded down to the hull.

 

Chris

[Liam]

Hi Chris,

 

No I don't think it does... the switch is a simple switch with fuse in.

 

There is something which looks like an ammeter which has a square face on it with dial which is usually in the central (12 o'clock position but does twitch backwards and forwards when the engine is running/charging. No idea what this is, its probably not related but if I remember rightly there is a very small/think cable/wire which runs from this to the inverter.

[blackrose]

Assuming Liam's inverter in neutral-earth bonded and his AC system is hull-earth bonded, couldn't he just buy a plug in RCD from B&Q for £8?

Edited November 24, 2008 by blackrose