Current leakage?

[jes25]

So, the other day I went out to a job b/c one of my guys said there was an issue with voltage on this machine. The machines are in a vet clinic and there are two identical units. The machines have a tank that gets hot inside somehow with steam or something for sterilization. I'm not real sure how they work, but they have a series of heaters (not submersed) all around them. Anyhow, they are single phase 3 wire 208/240 rated and hooked up 240. One of the units is using the high leg of a delta service, which is of no consequence IMO. (although the machine tech insisted that was the problem)

I identified the problem, which was an open ground on the receptacle. I corrected it and the problem went away. I looked everywhere for a bootleg neutral connection inside these things to no avail. So my question is, can all those heaters "leak" voltage onto the machine with no ground? I'm talking like a 180V's with a digital meter that disappeared under the load of a wiggie. You could feel it tho. I still think there is a 120V connection or something inside but we took it all apart and I couldn't find one.

Sorry for the novel

 

[GoldDigger]

So, the other day I went out to a job b/c one of my guys said there was an issue with voltage on this machine. The machines are in a vet clinic and there are two identical units. The machines have a tank that gets hot inside somehow with steam or something for sterilization. I'm not real sure how they work, but they have a series of heaters (not submersed) all around them. Anyhow, they are single phase 3 wire 208/240 rated and hooked up 240. One of the units is using the high leg of a delta service, which is of no consequence IMO. (although the machine tech insisted that was the problem)

I identified the problem, which was an open ground on the receptacle. I corrected it and the problem went away. I looked everywhere for a bootleg neutral connection inside these things to no avail. So my question is, can all those heaters "leak" voltage onto the machine with no ground? I'm talking like a 180V's with a digital meter that disappeared under the load of a wiggie. You could feel it tho. I still think there is a 120V connection or something inside but we took it all apart and I couldn't find one.

Sorry for the novel

It was a good read, I couldn't put it down.

It is unlikely that there is an actual resistive leak from the heater element(s) to the metal frame, but it is quite probable that there is a capacitive current leakage. It does not take very much current, not even enough to trip a GFCI, to be felt when touching the machine. And that would show up fine on a digital meter and disappear on a wiggle. A real resistive leak that only shows up under high voltage is not impossible either.

Now, why would you see it on one machine and not the other, and why did the technician say what he did?

Elementary:
1. The machine was designed for 208 wye or 120/240 split phase. In the both cases the heater element leads would not be any higher than 120 volts from ground. So perhaps the insulation of the machine is not quite good enough for the tech to be happy with 208 volts to ground from your high leg delta wiring.
2. But much more likely: If the problem is capacitive leakage, then the average voltage from heater resistance wire to the shell at ground would be 70 volts or less in the case of the wye, and exactly zero in the case of the split-phase 120/240. So the leakage would be minimized. But in the case of the machine connected to the high leg, the average voltage from the heater element to ground would be 104 volts. If the leakage was not symmetric from one end of the winding to the other, then depending on which way the two leads were connected, you might even see as much as 200V to ground on the shell of the machine.

So even if both machines had no effective ground connection, the one on the 120/240 would have zero leakage current and the one on the high leg could have quite a lot.

I rest my (voltmeter) case.

 

[Sierrasparky]

WOW I needed to read that twice!

 

[GoldDigger]

WOW I needed to read that twice!

Take your time. I did. :happyyes:

 

[ActionDave]

WOW I needed to read that twice!

Take your time. I did. :happyyes:

I read it twice. I need pictures.

 

[GoldDigger]

I read it twice. I need pictures.

Sorry, I am graphically challenged. I just wave my hands and draw in the sand on floor of the shop.

Let's try this:

Go to http://en.wikipedia.org/wiki/High-leg_delta and look at the second illustration on the right side.

What is the average height of line ac above the ground/neutral point at n? It is zero.
Now look at the average height of the line ab above ground. Looks like about 104 volts.
Imagine those two lines are the heater coils on the two machines, and decide which heater will try to capacitively couple more voltage with respect to ground into the shell of the machine?

 

[jes25]

It was a good read, I couldn't put it down.

It is unlikely that there is an actual resistive leak from the heater element(s) to the metal frame, but it is quite probable that there is a capacitive current leakage. It does not take very much current, not even enough to trip a GFCI, to be felt when touching the machine. And that would show up fine on a digital meter and disappear on a wiggle. A real resistive leak that only shows up under high voltage is not impossible either.

Now, why would you see it on one machine and not the other, and why did the technician say what he did?

Elementary:
1. The machine was designed for 208 wye or 120/240 split phase. In the both cases the heater element leads would not be any higher than 120 volts from ground. So perhaps the insulation of the machine is not quite good enough for the tech to be happy with 208 volts to ground from your high leg delta wiring.
2. But much more likely: If the problem is capacitive leakage, then the average voltage from heater resistance wire to the shell at ground would be 70 volts or less in the case of the wye, and exactly zero in the case of the split-phase 120/240. So the leakage would be minimized. But in the case of the machine connected to the high leg, the average voltage from the heater element to ground would be 104 volts. If the leakage was not symmetric from one end of the winding to the other, then depending on which way the two leads were connected, you might even see as much as 200V to ground on the shell of the machine.

So even if both machines had no effective ground connection, the one on the 120/240 would have zero leakage current and the one on the high leg could have quite a lot.

I rest my (voltmeter) case.

Ok, well I admit I haven't had time to check out the link, but I have a few thoughts:

The problem was on both machines when plugged into the high leg / open ground receptacle. When I connected open ground recep to the 120V legs of the delta the problem still existed, but with only 80V to ground. So, it seems your hypothesis of capacitive leakage is correct.

Now, what is capacitive leakage and resistive leakage for dummies?

 

[GoldDigger]

Ok, well I admit I haven't had time to check out the link, but I have a few thoughts:

The problem was on both machines when plugged into the high leg / open ground receptacle. When I connected open ground recep to the 120V legs of the delta the problem still existed, but with only 80V to ground. So, it seems your hypothesis of capacitive leakage is correct.

Now, what is capacitive leakage and resistive leakage for dummies?

Simple uncluttered from scratch:

Two wires next to each other will form a capacitor. Not a very big one, but a capacitor just the same. For a couple of feet of wire, 1/4" apart, the impedance at 60Hz of that capacitor will be very large. probably megohms or more.
But if you move the wires closer the capacitance increases. And as you make the parallel wires longer the capacitance increases too.
If instead of two wires you have a heater winding (50 feet long when uncoiled?) and it is separated from a flat metal plate by a thin insulator, there is also capacitance between the wire and the plate. If you apply a voltage, current will flow. No power because PF=0, but current will flow.
If the capacitance gets high enough, the impedance at 60Hz may get down into the hundreds of thousands of ohm range. At some point the current flowing will get up into the milliamp level and you will start to feel tingling or buzzing when you touch the plate.
Ground the plate and there will be no voltage and the milliamp of current just flows to ground. The amount of current is inversely proportional to the frequency, so with a DC ohmmeter you measure infinite resistance.

If instead of looking at capacitance, you look at resistance because the insulation is not good, then you have a current flowing which is not dependent on frequency and you can measure it with an ohmmeter.

In a real world heater winding, there can be both capacitance and resistance in parallel. All that you will measure as fault resistance with your ohmmeter or DC megger will be the resistive part. But the capacitive part can still cause a shock and not show up on a DC test.

Keep in mind that I am not talking about the series resistance from one end of the wire to the other. We can ignore that except for the fact that it allows the voltage at different points along the wire to be different when the heater is turned on. That means that the leakage current will be coming from all of the different parts of the wire, as if in parallel, and not just from a wire at line voltage from one end to the other.

Questions from there?

 

[jes25]

Questions from there?

Well, I think I get the resistive leakage part fine; essentially a high impedance fault. And for the capacitive part, it sounds like this type of heater essentially becomes a capacitor and develops a charge that flows to ground. If we had a ungrounded section of conduit with a few hundred feet of wire inside would we get a similar effect? Also, capacitive leakage is the reason for the higher trip rating for deicing cable GFCI protection I'm assuming.

No power because PF=0, but current will flow.

This part I don't get. I thought capacitors were used to correct a low PF

 

[GoldDigger]

Well, I think I get the resistive leakage part fine; essentially a high impedance fault. And for the capacitive part, it sounds like this type of heater essentially becomes a capacitor and develops a charge that flows to ground. If we had a ungrounded section of conduit with a few hundred feet of wire inside would we get a similar effect? Also, capacitive leakage is the reason for the higher trip rating for deicing cable GFCI protection I'm assuming.

I think you have got it.

This part I don't get. I thought capacitors were used to correct a low PF

Both capacitors and inductors have a power factor of zero when used by themselves but their 90 degree phase shift is in the opposite direction. And when you have both a resistive impedance and a reactive impedance the true impedance comes from those two in parallel. You do not just add the two values with a j on one of them.

Anyway, the result is that when you add the same amount of capacitive reactance and inductive reactance into the load, you are left with only the resistive portion so the power factor will be 1.
(Making the whole thing harder to understand is that no capacitive reactance is not an impedance of zero it is an impedance of infinity. <sigh>)