Ground Detection for DC panelboard

[WastefulMiser]

Is there a relay for this or is this possible? You don't bond the positive or negative ... is this a mistake?

And can regular CT's be used for DC current monitoring?

 

[gar]

100218-1150 EST

WastefulMiser:

A standard current transformer, a transformer, as distinguished from some other device that is incorrectly labeled as a current transformer, can not directly measure a DC current.

Magnetic current sensors that use a Hall device can be used at DC. There is a DC drift problem that may be of concern as applied to its use to determine if there is difference between the currents in the + and - supply lines.

When using a magnetic material and DC there is always the problem of residual magnetic flux.

Can you more completely describe what you want to do with some numeric data? What is the maximum load current? What is the smallest increment of "lost current" that you desire to detect? What are the consequences of the "lost current", if any exists?

Note: the "lost current" does not have occur as a result of a ground fault.

.

 

[jim dungar]

Is there a relay for this or is this possible? You don't bond the positive or negative ... is this a mistake?

And can regular CT's be used for DC current monitoring?

Yes it is possible to monitor for a ground on an ungrounded DC circuit.

No, AC CT's cannot be used on DC circuits.

 

[WastefulMiser]

Thank you.

I have a spec in which they want to monitor the battery voltage and the bus voltage inside the panelboard via a switch, which I do not think will be an issue.

Next they want to monitor the charge and discharge current of the batteries inside the panelboard. Discharge being able to read up to 400A (25-0-100mV scale) and charge up to 40A (0-10mV scale) which we (vendor) will have to use Hall device that gar kindly educated me on which I guess will go to a transducer to the ammeters.

They also want a 27-relay and ground detection equipment consisting of 2 amber LED indication and 1 pushbutton. And 1 continuous ground detection relay with an output contact for remote alarm.

The 27 won't be an issue; however, how do you detect stray current when you have several inputs (chargers, batteries) and mulitple outputs (UPS, trip, charge, close, LED, 86's, relays, etc) without getting really complex? I've never seen ground detection on a DC system. Maybe I've never just noticed ...

 

[WastefulMiser]

Can you more completely describe what you want to do with some numeric data? What is the maximum load current? What is the smallest increment of "lost current" that you desire to detect? What are the consequences of the "lost current", if any exists?

Note: the "lost current" does not have occur as a result of a ground fault.

Let me more accurately answer your question: I'm not sure without consulting the end user. Their spec is very vague as I outlined in my previous post. What would be the ideal response to them?

"What is the smallest increment of 'lost current' that you desire to detect? Lost current doesn't necessarily mean a ground fault." Eh??

 

[Besoeker]

Is there a relay for this or is this possible? You don't bond the positive or negative ... is this a mistake?

On a fully floating system we connect two resistors (or resistor chains depending on voltage) between the DC+ve and DC-ve. The centre point is connected to ground and we put a voltage sensing relay across each resistor.
The relays are set such that neither picks up if the voltages are balanced. In the event of one side of the DC leaking to ground, the voltages will go unbalanced with one relay getting lower voltage and the other higher. With enough unbalance, the relay will pick up. It's simple and adjustable.
[/QUOTE]

 

[rcwilson]

Battery ground systems don't try to detect current, they detect voltage.

Measure positive to ground and negative to ground voltage. If the system is floating, it should be about +65V and - 65V on a 125 VDC system.

Google battery ground relays and you will find at least one relay that does this. (Arga). Areva (old GEC) also used to have a relay that did the 27 and the ground detection, and I thought Basler did too. The LED's and pushbutton in the specification are standard.

For your battery current monitoring, look at using a shunts and a mechanical meter with a custom scale. It is cheaper than a digital device. At least it used to be.

Battery ground detectors are also a standard part of most battery chargers.

 

[rcwilson]

DC ground detectors used to be two 130V lights wired in series across the battery with the midpoint grounded through a NC pushbutton. (Assume a 125V DC system). If there were no grounds on the floating DC system, both lights would be dim at half voltage.

When a ground occurred on the positive bus, the positive light went out and the negative light went full brightness. Pushing the test button lifted the ground connection and both lights would go back to dim. The "test" only proved the lamps weren't burned out.

The two LED's and pushbutton in the spec are more modern versions of the two light ground detector.

The purpose is to alarm that a ground exists so it can be repaired, not to trip on ground fault. That is why no leakage current monitoring is needed, unless you want to get real expensive. I've put those Hall effect monitoring systems in and they work, sometimes.

 

[WastefulMiser]

Thank you all. Makes a lot of sense. I appreciate your time and knowledge.

Our vendor just quoted a digital meter (Hawk 3) for everything. I will get them to update accordingly to meet the spec.

How would you detection the "stray voltage" without turning off the breakers one-by-one?

 

[gar]

100218-1316 EST

WastefulMiser:

I used my concocted term "lost current" to avoid the term ground fault.

Consider a GFCI device. This consists of a current transformer thru which both the source and return wires path. If there is no unbalanced current between these wires, then the net current detected by the current transformer is 0.

With a GFCI it does not matter how current between these two conductors is lost or supplied from some other source. When an unbalance of sufficient magnitude occurs, then a threshold level is exceeded, and the device is tripped.

The internal test resistor in the GFCI is not a test to ground, but simply a resistor between opposite wires on opposite sides of the GFCI.

It is true that most of the time concern is with shorts to ground.

The method described by Besoeker is a neat simple way and can realistically avoid the drift and zero bias problems of a Hall device and provide great sensitivity. It only measures lost current to or from ground. Maybe a more useful description is it measures an unbalanced leakage resistance to ground because it is a relationship of the leakage resistance to the test resistors. It is like an unbalanced Wheatstone bridge.

Your charge and discharge monitors probably do not require great accuracy. So does this mean an ordinary analog meter is satisfactory and to be used, or what.

If you have one current sensor, for example a DC shunt (just a low resistance 4 terminal resistor), then you can measure the net battery current flow by putting the shunt in a common lead to the battery. If you need to separately know the current from the charger and the current to the load, then two current sensors are required.

If a shunt is used to measure current, then there are direct wired connections to the high current cables. If a magnetic sensor is used to sense current, then isolation exists at the sensing location. A shunt can be used with an isolated amplifier as another coupling means. A resistor shunt will probably give the most accurate stable readings over the greatest dynamic range.

.

 

[rcwilson]

"How would you detection the "stray voltage" without turning off the breakers one-by-one? "

Turning off the circuits momentarily is the most common method of tracing down the fault location. On many systems, the short interruption is possible.

Halmar used to make Hall effect sensors ganged together with a controller that indicated which circuit was faulted. Each sensor encircled one circuit's + & - lead, just like a zero sequence ground fault system. Expensive due to all of teh sensors required, but it worked well except on some high resistance faults.

Another method is to inject AC current between ground and the faulted DC leg. We used a 120 V Variac feeding a 480-120V control power transformer wired for isolation/step down. The output ran through a limiting resistor and a timing relay that pulsed the current on and off about once a second. Using a clamp-on on the lowest range we could follow the pulsing current along the wiring to the fault. A pulsing current is easier to detect on the meter. Depending on the meter & clamp-on we used the AC current was in the 0.2 - 5.0 Amp range.

This was on old WWII vintage control circuit wiring. The impressed AC had no effect on the old equipment. I don't know if I would do it on a UPS system.

BTW- the most fun was when the 250V DC crane power supply ground detection lights indicated a grounded rotor winding on a crane motor. As the crane moved, the DC commutator would switch the faulted winding from positive to negative. During production, cranes were zipping up and down giving us quite a light show in the electric shop as we pulled the replacement motor out of storage.