Galvanic Isolators, again sorry

[wakey_wake]

 I've read about this and thought about it too much. We have three electrical points all claiming to be "earth" and a device with two terminals. It's this thinking which is leading me to ask more questions... 🤔

 

I have read the often-instructed... "Gibbo's Smartgauge website here " and it agrees with the instructions on any (all?) GIs I've seen.

 

I've also read a lot of old GI threads here, and I understand this-

On 08/11/2010 at 13:01, Gibbo said:

A GI is totally unnecessary with an on board generator. The whole purpose of a GI is to interrupt the galvanically generated current between your hull and shorepower via the shorepower lead earth conductor. You don't have that connection with an on board generator unless you put the generator on the shore with an earth spike, which is just daft and is also no longer an on board generator.

 

I understand back-to-back double diodes as working by

1. Accepting voltages generated by dissimilar metals, in either direction, without significant conduction
2. Conducting fault currents (such as Live touching accessible metalwork) with insignificant voltage drop, so that breakers can trip promptly.
3. Being able to continue conducting fault currents indefinitely? Else the specs and heatsinks on some are ridiculous overkill.
4. Sometimes, using any developed voltage to light diagnostic LEDs. I believe this isn't going to be triggered by dissimilar metals as the ones we use won't develop quite enough voltage.

 

I understand that capacitor(s) across the GI are to conduct AC at significantly higher frequency than 50Hz, so those currents bypass the GI.

 

I also know an isolating transformer with a shore-earthed screen is the gold-plated solution, but I don't have the budget for that. Money, weight or space.

 

 

All instructions say to insert the GI between shore-earth and (consumer-earth = hull). Sorry, this scribble is the most convenient just now.

 

This protects against ordinary galvanic corrosion in the absence of bad equipment, and it offsets you against leakage currents generated elsewhere on the shore-earth.

 

It doesn't protect against in-boat leakage currents. Nor does it provide status lights to warn about that.

 

So why do we not insert the GI between (shore-earth = consumer-earth) and hull? I think this must be because one could accidentally short the GI and not realise, an thereby lose the hull protection. Electrocution protection remains solid.

 

After this I got into thoughts which surely are heading towards madness, but I can't really see why. 🤕
Is there some justification for a star or delta type GI, in order to have protection from and detection of onboard DC leaks?

Yeah this would have been quicker with pencil, paper and camera... I give you the "wacky delta wired GI". I know you won't like it. I don't like it!

But I cant actually see why it's wrong.

 

If each basic GI has status lights then you know what's leaking where.

Obviously it's overkill to do it with three ordinary GIs. If this schematic were actually sensible you would want three bridge rectifiers on the same heatsink, because only one will ever be conducting fault current.

[nicknorman]

The purpose of the GI is to stop low voltage current flowing between the shore earth, and the hull, to prevent electrolytic corrosion. So your first diagram is correct. The others are not correct.

In boat leakage currents don’t cause electrolytic corrosion and are “dealt with” by having an RCD.

Edited September 30 by nicknorman

[beerbeerbeerbeerbeer]

Idiots question; if on occasion a boat ties up in a marina and uses an extension lead from battery charger to shoreline socket,

would one still advise a Galvonic Isolator?

 

the boat’s not using it’s 240v mains sockets,

the charger is plugged directly to land and wired to battery bank

[nicknorman]

6 minutes ago, beerbeerbeerbeerbeer said:

Idiots question; if on occasion a boat ties up in a marina and uses an extension lead from battery charger to shoreline socket,

would one still advise a Galvonic Isolator?

 

the boat’s not using it’s 240v mains sockets,

the charger is plugged directly to land and wired to battery bank

It depends on whether the dc negative terminal is connected to earth within the charger. Probably not but you never know…

Even if it is, if “on occasion” just means for a day or so every several months, even if there is earth current it’s unlikely to do much to a 15 tonne steel boat.

[beerbeerbeerbeerbeer]

 

🤷‍♀️

it’s an IP22 Victron. 
wired straight to battery bank and I plug it in to shore via an extension lead. 
 

 

[nicknorman]

I don’t know, but with the thing unpowered you could measure the resistance between the negative dc terminal and the earth pin. Of course it may not have an earth pin (ie it could be plastic) in which case no problem. If it has a metal earth pin and a low resistance connection to dc negative then problem. But most likely it will be open circuit in which case no problem.

[wakey_wake]

1 hour ago, nicknorman said:

The purpose of the GI is to stop low voltage current flowing between the shore earth, and the hull, to prevent electrolytic corrosion.

Yes. I agree that is necessary, and the "standard" way will help a lot with that.

I think the "non-standard" way would help just as much because in this analysis we're ignoring the consumer arm of the circuit, and looking at the shore-GI-hull circuit.

 

1 hour ago, nicknorman said:

So your first diagram is correct. The others are not correct.

I know the first is conventional and normal, that's why I labelled that that way.

 

I'm hoping to understand the ways the others are "not correct"... and for that reason we hope, not conventional.

 

1 hour ago, nicknorman said:

In boat leakage currents don’t cause electrolytic corrosion and are “dealt with” by having an RCD.

If a diabolical device in the boat takes 10mA of electrons from hull-Earth and deposits them in Live and/or Neutral it won't quite trip the RCD. No status lights will light to warn me.

I'm pretty sure it will cause electrolytic corrosion, because those electrons will mostly come from Fe(solid) becoming Fe²⁺.

 

 

This thought was prompted by -

I made up a mains extension lead with the Earth interrupted by a pair of shrouded 4mm connectors and ten turns into which the current clamp can be inserted. That's three ways to see what's going on, two of which don't insert extra resistance.

As an aside: I highly recommend https://www.amazon.co.uk/gp/product/B00O1Q2HOQ/ which has precision to a milliamp on the current clamp.

 

Then I surveyed a fair fraction of the stuff onboard and concluded,

1. many devices leak enough to see, even when "off" eg. an older microwave at 0.15mA
2. long runs of cable leak too, which surprised me until I considered there must be capacitance between the conductors... but it was more than I expected (without calculations). I forgot the details.
3. I'm at the limit of what my equipment will show me, and at modest risk of things like meter impedance skewing the results.
4. I don't have a convenient way to tell whether any of this is simple AC 50Hz in-phase, some other strange AC, or with DC components.

So I asked a question and got more questions!

 

 

55 minutes ago, beerbeerbeerbeerbeer said:

Idiots question; if on occasion a boat ties up in a marina and uses an extension lead from battery charger to shoreline socket,

would one still advise a Galvonic Isolator?

 

the boat’s not using it’s 240v mains sockets,

the charger is plugged directly to land and wired to battery bank

Assuming the boat already has GI, circuit breakers, RCD etc. then my concern would be that you're bypassing all of that protection. For what benefit?

 

If the boat doesn't have all that, then you're just charging a battery which happens to be connected to a boat. You might be relying on an RCD in the bollard for your personal safety. Check it's there, check its ratings, consider bringing your own (inline) on the 16A cable?

 

You might want to check the DC current is balanced, ie. the IP22 is well behaved. You might want to check (when it is disconnected) there is an open circuit from DC negative to all charger mains pins. Being Victron I guess it would be OK?

 

You might want to get an inline GI anyway? But it sounds like it would not be worth it in your case.

 

Or borrow an inline GI with status lights, and see them being off for peace of mind, and then later wake up worrying that maybe they would be on now while you're not able to watch? Oh dear, time to refill my 🍺🍺🍺🍺

[nicknorman]

Your non-standard diagram is wrong for a couple of reasons, firstly one of those consumers with a metal box might end up with its earthed case connected to hull, then you have bypassed the GI. Secondly, and bearing in mind most boats have an onboard source of mains power (inverter) you have placed a device between the consumer’s earth line and the hull which gains you nothing. A device which might quite easily fail open circuit. 

 

As to your “diabolical device”, it doesn’t exist. Don’t forget that ac doesn’t cause electrolytic corrosion, only dc.

[DandGNWales]

15 hours ago, nicknorman said:

Your non-standard diagram is wrong for a couple of reasons, firstly one of those consumers with a metal box might end up with its earthed case connected to hull, then you have bypassed the GI. Secondly, and bearing in mind most boats have an onboard source of mains power (inverter) you have placed a device between the consumer’s earth line and the hull which gains you nothing. A device which might quite easily fail open circuit. 

 

As to your “diabolical device”, it doesn’t exist. Don’t forget that ac doesn’t cause electrolytic corrosion, only dc.

AC can cause  corrosion it's a known thing on Pipelines where they pick up induced ac off overhead power lines and when it leaves the pipeline it takes  a little metal each time, leaves a nice smooth pit, should be ok on the boats lol