Galvanic Isolator with LED status

[geoffschultz]

I'm down in Guatemala putting together a list of repair parts for my boat which was struck by lightning. Basically everything that was electrical was fried, including the galvanic isolator. My old one (a Guest 2433-p) had a led which indicated that it was working. I really liked that I could just look at the isolator and know that it was working. However, I can't find a replacement that has a status indicator. Can anyone give me a pointer?

-- Geoff

 

[dlochner]

There are 2 types of diode based galvanic isolators, fail-safe isolators and fail not safe isolators. ABYC requires the the LED on the Fail Not Safe Isolators, so that you know if it has failed.

Fail Safe isolators fail in such a way that the AC ground works, however, the galvanic protection is no longer there. Fail not safe isolators fail the other way, the boat is still protected from galvanic issues but the AC ground is broken.

Better to have ground protection on AC devices. Testing a GI is not that difficult. Testing A Galvanic Isolator Photo Gallery by Compass Marine How To at pbase.com

 

[geoffschultz]

Thanks for the info. I agree that having a ground is highly important, but I have to wonder why they can't incorporate a status LED into a failsafe galvanic isolator. Quite honestly, with all of the other maintenance that needs to be done on a boat, disconnecting my electrical system to test the isolator isn't something that I realistically will do on any regular basis.

-- Geoff

 

[Gunni]

I hear you Geoff. My assumption is that our GI status indicator is our zincs. When they evaporate in 60 days it is time to check and replace the GI

 

[walt]

You may notice that the GI that have the LED indicator require 12 volts also to power the detection/drive circuit.

The method on that web site to test GI is "useful" but it also can be way off.

We used to test earth ground on electronic equipment and UL required the ground continuity test to be done with 10 amps of current. Remember that while the earth ground normally conducts no current, when it does its safety thing, it has to pass the full current necessary to blow the fuse.

So if the GI is on the earth ground for a 30 amp circuit, when a fault occurs, the GI must pass 30 amps in order to blow the fuse. That is also the reason the GI is rather hefty.

Consider a fault in the GI where it still sort of looks like a back to back diode set but its impedance went from near zero to 10 ohms.

The DVM meter might use 1 to 10 ma for bias current on the diode test (guess - depends on the meter but probably ballpark). The bad GI with the 10 ohm resistor would test just fine for both of those tests since that very small current only causes a voltage drop of .01 to .1 volts. Everything looks good in the test.

But now take that same bad GI with a 10 ohm impedance and put 30 amps through it. That 10 ohm resistor will limit the current in the GI to 11 amps rms - and that is not enough to blow the fuse. So the safety ground does not do its job and the dangerous condition still exists.

FYI, I would probably be fine myself using those simple tests.. but UL wouldn't.. And now were getting real nit picky on a page which is useful..

 

[geoffschultz]

Walt, I'm lost. I understand from your description how the described test could produce a false sense that the GI is working properly, but I don't understand where the 10 ohm resistance that you describe comes from. Is this internal to the GI, or on the ground connection in the shore power plug, or where?

-- Geoff

 

[walt]

Every part of the safety earth ground from beginning to end must be very low impedance and able to handle the fuse/breaker threshold current for it to do its job. But if we are testing only the GI, the fault I mentioned would be internal. Ive seen similar types of failures in transient suppression diodes. Or corrosion might do that. A failure like that would likely be rare but could happen and hence the point that UL would not allow that test in production.

In your case where the GI was checked and failed, the simple test likely never gives a false failure (at least not something I can think of). Its only that it can give a false pass. Ie, if you did the simple test and it shows the GI is bad, assuming you did the test accurately, you can feel comfortable that the device is bad.

why they can't incorporate a status LED into a failsafe galvanic isolator.

With 12 volts to power some sort of circuit, this probably is possible. Why you cant find one.. dont know. Maybe there isnt a market? I would think GI are very rugged to begin with and have a very low failure rate. Its a very simple device that can handle a lot of current. I would think the circuit that is detecting and driving the indicator LED is much more likely to be damaged by lightning since it has an external 12 volt input (with possible lightning induced transients) than the actual GI diode set itself..