Galvanic Isolators

[GUMPY]

Single phase 230 volts, the product of a diseased brain.

Not at all, it allows the use of half the copper in cables compared to running 110v, which considering the worlds copper reserves have less than 10 years left might be considered to be a good thing.

 

I think you will find that with the exception of the Americas and Japan most countries are now 230v L-N.

There used to be the occasiuonal place in Europe ( Oslo, some parts of Belgium, some parts of Paris) that had Delta power i.e. three phases of 127v giving 220v between the phases but we havent come across one like this for a while.

 

Another thing is that running two phases means double pole breakers on everything and all switches have to be DP as well, thus increasing the cost. This also applies to circuits fed from an isolation transformer if you do not bond one side of the secondary of the transformer to the boat ground.

 

 

J

Edited November 11, 2006 by idleness

[Grasshopper]

Also (ref your Sea Otter) while aluminium is less prone to chemical corrosion than steel, aluminium and aluminium alloys are actually above steel in the galvanic list of metals. ie: aluminium is more susceptible than steel to galvanic corrosion.

 

 

Yup. Hence the fact that I was following the thread and erring on the side of caution with regards to implementing shore power.

 

If your boat sits between them, this current may take a short cut by going in an item near one boat, and exiting via your anode near the other. This will eat up your anode too even though you are not connected to them

 

Now you're just trying to give me nightmares............

 

Zinc fish huh? First I've heard of them I'll confess. I'll have a google.........

[chris w]

Grasshopper

 

I'm really not trying to give you nightmares.

 

Here's a good article http://www.ocean.udel.edu/seagrant/publica.../corrosion.html

 

and http://www.performancemetals.com/faq.asp

 

and a third good one http://www.boatus.com/boattech/MarineCorrosion.htm

 

Chris

 

PS: you may moor your ali boat next to mine anytime!!!!!!!

Edited November 11, 2006 by chris w

[Keeping Up]

The idea of using a transistor to detect the change occurred to me too (in order to obviate the need for a second GI). The issue would be getting some hard switching to occur based on the rather flaky characteristic of diode forward voltage drops.

How about this circuit? I doodled a basic circuit on the back of an envelope, then realised that it would be a good idea to protect it in case the GI were faulty. That gave the opportunity to indicate that fact as a safety check, so I added the high-power 10k resistor (good for 300v rms, but you'd probably have to use several 2W metal-film resistors in series because a wirewound one wouldn't sensibly pass RF) and extra LEDs in pairs so they won't conduct if the GI is working.

 

Basically, one LED alight means there's DC present, 2 LEDs alight means there's AC present, and the bunch of 4 alight means beware your GI is faulty.

 

As a portable tester you'd need to power it from a pair of PP3 batteries (which would be the most expensive part of the unit) or you could make a simple +/- supply to run it from the incoming mains (using a double-insulated transformer of course) and then it could be left in place. You couldn't, of course, run it from the boat's batteries.

 

Would it work?

 

Allan

[chris w]

Allan

 

Good thinking. A couple of points......

 

1. I don't think you'll have enough base drive to illuminate the LEDs. The GI voltage we're trying to detect is only 1.2v. Applying this to the 10K & 1K resistors in series means, that after deducting the BC548 Vbe, there's only about 50uA base drive.

 

2. To get hard switching you should reckon on a value of about 10 for the HFE value which won't produce enough drive current in the collector circuit. Even if you use the actual HFE (90 min) it's barely marginal. Because of the low value we're trying to measure (1.2v) you can't use a Darlington or Schmitt arrangement due to Vbe's adding up but a simple 2 stage common emitter circuit should solve it I guess.

 

3. You would need to connect the 9V battery common to the boat earth to be able to reference the GI voltage.

 

4. If the GI failed open circuit, the "fault" LEDs would not indicate this (due to the voltage drop across the 10K resistor) unless mains voltage (well > 50v at least) existed on the earth line, in which case they would illuminate. However, presumably the RCD would have tripped immediately this happened.

 

 

I am working on a similar circuit but incorporating a bridge rectifier to be able to detect AC or bipolar DC with one transistor and be able to use the boat battery to power it.

 

How do I add a file attachment to these posts so that I can post the circuit diagram? I can't seem to find any info on this. Must a file be hosted on a website or can it be posted from my PC?

 

Cheers

 

Chris

[Audi Fan]

How do I add a file attachment to these posts so that I can post the circuit diagram? I can't seem to find any info on this. Must a file be hosted on a website or can it be posted from my PC?

Upload it to your gallery Chris

 

Click on 'My Controls' above and start an album

[chris w]

Thanks

[Keeping Up]

Allan

 

Good thinking. A couple of points......

 

1. I don't think you'll have enough base drive to illuminate the LEDs. The GI voltage we're trying to detect is only 1.2v. Applying this to the 10K & 1K resistors in series means, that after deducting the BC548 Vbe, there's only about 50uA base drive.

 

2. To get hard switching you should reckon on a value of about 10 for the HFE value which won't produce enough drive current in the collector circuit. Even if you use the actual HFE (90 min) it's barely marginal. Because of the low value we're trying to measure (1.2v) you can't use a Darlington or Schmitt arrangement due to Vbe's adding up but a simple 2 stage common emitter circuit should solve it I guess.

 

3. You would need to connect the 9V battery common to the boat earth to be able to reference the GI voltage.

 

4. If the GI failed open circuit, the "fault" LEDs would not indicate this (due to the voltage drop across the 10K resistor) unless mains voltage (well > 50v at least) existed on the earth line, in which case they would illuminate. However, presumably the RCD would have tripped immediately this happened.

I am working on a similar circuit but incorporating a bridge rectifier to be able to detect AC or bipolar DC with one transistor and be able to use the boat battery to power it.

 

How do I add a file attachment to these posts so that I can post the circuit diagram? I can't seem to find any info on this. Must a file be hosted on a website or can it be posted from my PC?

 

Cheers

 

Chris

OK. Answering each point in turn:

 

1&2. I reckon 50 microamps should be enough base drive, the hfe is quoted as 200-450 which should give at least 10mA drive to the LEDs which is plenty. You could use a BC547C instead (still only 17p from Maplins) which has more than double the gain. It wouldn't actually matter if it didn't switch on hard, as long as a few mA go through the LED, and in fact it would respond better to high-frequencies if it didn't saturate.

 

3. You don't need any connections other than the ones I've shown. One side of the GI is connected to boat earth, the other to 'msins' earth, but it doesn't matter which probe goes to which side. The whole point is that it simply connects across the GI.

 

4. You may be right, the 'fault' LEDs could be a little insensitive. 50v would pass 5mA through the LEDs; it wouldn't make sense to reduce that 10k resistor (or it gets awfully warm should there be 240v across it) but many LEDs will give a good glow at 2mA. All the same, you may not get that much earth leakage and then they wouldn't glow. It really needs a bit of amplification without losing the protection aspect (I'm not sure how the transistors would cope if you put 30mA into the base)

 

If you can draw your circuit and save it as for example a jpg, you then need to post it on a server somewhere (photobucket for example); if you haven't got access to any such server, PM me and I'll tell you an address you can email it to, then I'll put it up on my own web space. I hope you've got an easier way of drawing it than usong MS-Powerpoint which is the best I've got these days.

 

Allan

[chris w]

You're right of course about the battery common connection. I somehow overlooked the fact we're just measuring across the GI!! The BC548B is spec'd with an HFE of 220 min so it should be enough to light the LED. I was originally just looking at the basic BC548.

 

I have a great piece of software for drawing AND running circuit simulations. It's called "LiveWire" and is available (inter alia) from Maplins at about £25. It has virtually any circuit component you could need except PICs and a full range of instrumentation. Thoroughly recommended. It links into a companion piece of software called PC Wizard (again Maplins ~£25) which allows one to lay out a PCB for fabrication.

 

I'll draw your circuit on LiveWire and run it. The BC548B and BC557B are both in a long list of available simulation transistors in its contents. It is a truly superb tool. It even allows one to run "explosions" (!) which simulates the circuit using real life parameters of the chosen components which if exceeded cause components to blow!

 

I do have a photo site to which I can upload jpegs although "Audi Fan", in his post above, indicated that I can create a forum album and upload from there.

 

Chris

Edited November 11, 2006 by chris w

[cheshire~rose]

Ok! I think I have finally grasped this!

 

Let me see: The hypothesis of the isolation transformer causing radio frequency interference due to the clamping action of the diodes depends upon the resistance of the complete circuit. How much RFI there was from the SM PSU depends upon the resistance of the complete circuit providing sufficient AC voltage (be it 50Hz or RFI or whatever).The source impedance will be high but It's just Ohm's law (or the ac equivalent V=IZ). Of course, the impedance of the capacitors that decouple the supply due to the relative values of the effective potential divider and combined pulse voltages via EMC filtering may cause dc superimposed on the ac current!

 

See it’s easy!

 

What I am saying is that is ever I do buy a boat I will have to make certain that whoever I pay to fit the thing had better:

• Be very well qualified!

• Be extremely experienced!

 

This forum is a truly wonderful place for ordinary people to get really great advice from people who know a lot more than they do. The problem is how can you judge if it IS great advice unless we can qualify that by having some knowledge of the ability of the advisor?

 

Do any of the advisors fancy sharing their credentials with me?

[Guest]

If ever I get a shore supply, I will have a wooden boat built

[chris w]

As with most things in life, there is rarely one correct answer but a continuum of "expert" opinion

 

Do any of the advisors fancy sharing their credentials with me?

Your place or mine!!

Edited November 11, 2006 by chris w

[DHutch]

I know i risk getting caught in the crossfire, but...

...If a 110v transfomer is £53 from Screwfix, Why are the isolating transformers so expensive?

Yeah, its stupid isnt it. Basicaly, as far as im aware, its simply a case of of the usally "marine" markup on prices. Live caravans, only more floaty...

 

 

I believe such things are generally what I used to know as an "autotransformer", and have only one winding in them that is shared by input and output circuit...

As it transpired below, i have always been under the impression that they where two coil, isolated transformer. So that they deliberatly isolate the output so that, like bathroom shaversocket transformers, you can actaully touch eather of the two powerwires and not get a shock, as long as you dont touch both. (which is where the 110v bit comes in...)

 

Sorry this is a little late in reply.

 

 

Daniel

[chris w]

Allan

 

I drew up the circuit on "LiveWire" and it works really well.

 

The LEDs monitoring the voltage across the GI start to glow at about 0.7-0.8v and are full on at 1.1v. The bipolar DC monitoring works fine (ie: regardless of the polarity of the voltage across the GI)

 

The LEDs monitoring for o/c on the GI diodes work fine too but the AC voltage needs to be about 80v rms before they glow appreciably, as we suspected. Obviously, because the latter LEDs are in inverse parallel it is possible to tell which side(s) has gone o/c. (although that's probably of academic interest!)

 

The next thing would be to get rid of the 9v batteries. Total quiescent current draw is practically zero and when lit with, say, a 1.1v input the current draw from each battery is 7mA. Since a fault might occur while away from the boat for a few days, the batteries might then be exhausted and one would not see an indication of a fault, so running the thing from a step down transformer and rectification would be the way to go as you suggested.

 

Great stuff

 

Chris

[Keeping Up]

Do any of the advisors fancy sharing their credentials with me?

OK, I believe in share and share alike. You show me yours and I'll show you mine

[cheshire~rose]

Your place or mine!!

 

 

OK, I believe in share and share alike. You show me yours and I'll show you mine

 

OOH! hang on fella's let me get the diary out...... I can think of no better way to spend a long dark evening than discussing inverse parallel diodes! Does this mean you want to compere notes on my source impedance?

let me just spell check that......

 

oh yes I think I meant sauce impudence!

 

Ok so who is buying the beer?

[Richard Fairhurst]

If ever I get a shore supply, I will have a wooden boat built

Believe me, compared to the hassles of protecting a wooden hull, a little corrosion from your shore supply is nothing to worry about...

[Keeping Up]

The LEDs monitoring for o/c on the GI diodes work fine too but the AC voltage needs to be about 80v rms before they glow appreciably, as we suspected.

Revised version posted to get around this problem by using a Press-to-Test switch. When the switch is not pressed the LEDs should detect a faulty GI if there are any small leakage currents (50 micro-amps upwards) or voltages (5 volts upwards)

 

Allan

[Osprey Sprinkler]

 

I reckon 50 microamps should be enough base drive...

 

Looks like you have this all figured out anyway but just a thought...

 

If you were worried about base drive you could replace the 10K resistor with an ordinary 5W (or less) incandescent light bulb. This would have the right resistance when hot (fault condition) but an order of magnitude or so less when cold, giving plenty of current to play with during everyday non-fault testing.

 

Of course you might want to check it doesn't have a coiled filament if you want to keep the inductance down.

 

Also, the LEDs might not stand the instantaneous forward current while the bulb warms up following a fault but you could replace them with suitably rated ordinary diodes, since you already have another light on. Might be cheaper too - not that the circuit is expensive in the first place .

 

Ashley

[Yoda]

Alan, Chris and Ashley et al,

 

You have lost the plot on this one.

 

It fine to make a tester for a galvanic isolater but that actually missed the point of doing away with this device because it will not work under most circumstances that it is being used under. The idea is to find a way of stopping galvanic action by other means and to ensure boat safety when using onboard generation or shore based power. This could be by isolation transformer or effectively two wire sytems.

 

The tangent that Alan has started does not stop the problems associated with this setup.

[chris w]

[the galvanic isolator] will not work under most circumstances that it is being used under

I disagree with you Yoda. A galvanic isolator WILL work perfectly at stopping galvanic currents providing the diodes don't go open circuit or they are not biased into conduction by small ac voltages. These latter situations are not "most circumstances".

 

What we are trying to monitor (and what Allan's circuit addresses) are:

 

a] any LOW level AC from onboard equipment on the earth line that might tend to bias the GI's diodes ON thereby decreasing their effectiveness against galvanic voltages (which tend to be around 0.8v max), and

 

b] any HIGH level AC voltages on the earth line which may blow or have blown the GI diodes thereby rendering the earth circuit open - a dangerous fault condition.

 

If, in our collective wisdom, we can solve these two issues then we have made the GI a far more useful and safe tool. If you want as close to absolute safety as is available (and money is no object), then simply buy an isolation transformer. (although if you're happy with just 1650VA, you can buy an isolation transformer from RS for £85+VAT; about the same price as a GI)

 

 

Chris

Edited November 12, 2006 by chris w

[Keeping Up]

a galvanic isolater ... will not work under most circumstances that it is being used under.

Tell us Yoda, upon what evidence (other than that which has previously been presented to this forum) do you base this sweeping statement?

 

Allan

[GUMPY]

I will add this again because the first part was ignored earlier.

 

One of the problems of self checking GIs is that one then effectively ends up with a testing system that is connected to a potentially lethal earth system. If it wasn't potentially lethal then why is it being checked? (think about it).

 

Now unless the system disconnects the power when a fault is found it’s going to be useless.

 

Think of this you are off the boat the diodes fail you are not there to see the monitor so there is no earth conected to the boat. You return to the boat to find the boat live and are killed as you step back onto the boat because the boatyard didn’t have an RCD on the supply. There is NO requirement in BS7671:2001 for marinas or moorings to have RCDs on the supply.

 

So while it might be a wonderful academic exercise it’s useless unless it disconnects the shoreline at the time of failure.

 

 

Julian

[lymmranger]

I have to congratulate all involved with this thread.

 

It has been both fascinating and informative.........

 

The potential game of tennis using a live handgrenade never materialised!

 

instead the "boffins" seem to be nearing agreement.....

 

A rarity on this forum these days

 

Can you guys please remember to translate this into english once you finally agree on the way forward!

 

[Gibbo]

Correct point well made Chris. I'll be interested to read Gibbo's response to this (are you there Gibbo?) because on his website he states "They simply do not operate as designed without the neutral-earth bond" which I have always disagreed with. And if the statement is incorrect it could have implications on the conclusions later in the page.

 

I come across this discusion all the time.

 

The easiet way to understand it is this.....

 

Draw yourself a circuit diagram for an AC supply into an RCD. Then try to get a different current down live and neutral without a neutral/earth bond.

 

You simply can't get one. Where can the difference in current go? It *has* to go back to the supply somehow.

 

Gibbo