DC power plant design question

Hello all:

I am looking for some opinions from the engineers here for a DC power plant design I am working on. This plant will be a nominal 48VDC plant with approximately 200 amp max load, for a small telephone system application.

The basic power plant will consist of :
1. 6, 30 amp AC-DC rectifiers working in parallel operation with load sharing
2. 24 cell Valve regulated battery string with approximately 450 amp hour capacity

The design questions are as follows:
1. Should I build the plant with a fixed earth reference, or let it float. I am leaning to the fixed reference on the positive side of the plant for simplicity sake, but will entertain the possibility of a floating system. Reliability is a key factor here.
2. I want to incorporate some current monitoring into the plant using 200 amp, 75mv shunts, so this gives rise to 2 questions.
a. Measure the current on the negative load side or the positive return side? I have seen it both ways but lately most always on the return side. May not matter in a floating design, but may matter in an earth referenced design.
b. Since I want to be able to monitor current on both the total rectifier outputs AND the total discharge load(s), it seems like I will need 2 shunts. Anyone have any idea how to get both with just 1 shunt?

If you reply, please give reasons (both pro and con) for your answers.

Thanks

1. Earth bond the negitive side, it will increase safety because if the mains is connected to it, a fuse will blow before anyone gets killed.

2. Don't bother with current shunts, go for hall effect sensors, they don't have a voltage drop and their output is isolated from the circuit being measured. Some sensors can also be simply clipped over the cable so you don't even need to break the circuit.

2a. It doesn't matter either way, especially if you're using hall effect sensors.

2b. I would think you need two sensors, you could have one on the battery, with a positive reading indicating charging and negitive discharging and another on the load indicating the total current draw.

It's been many years but I recall helping out on troubleshooting a telephone switch plant at our refinery. It used a bank of large wet cell lead acid batteries and was charged by two parallel 48vdc chargers for redundancy (not sure how they shared current, possibly just diode isolated, as either one was sized for max load and recharge rate.

Our system did indeed float the -48 volt system and I thought that was pretty normal in telephone systems as there are ground detectors on the subscriber output loops to help troubleshoot subscriber wiring by being able to tell if either tip or ring conductor was grounded.

Pretty hard to beat the simplicity and reliability of a simple large current shunts. If you wire the shunt from the positive battery terminal first that hook both the load and the charger positive leads to the other side of the shunt you will have a bidirectional current indication that the battery bank is either accepting a charge or supplying a current to the load. The chargers we had were designed to continuously float charge the batteries so the chargers were actually supplying both load current and battery float current, so a second current shunt would be needed if just load current information is needed.

Oh, by the way the problem we solved was that the original construction contractor did not torque down the battery terminals tight enough. As the chargers normally supplied the load all appeared OK until there was a AC power outage and the batteries had to supply full load current. We could see sparks flying among the battery terminals when we simulated a power failure. As these were wet vented batteries we were luck there wasn't an explosion

Lefty

Doing the load sharing is not an issue. The rectifiers I am using are designed with internal circuitry that drives / recieves a load share bus signal, which alters the internal regulate circuitry to ramp current output up/down till each is within 5% of the master. The master is the one with the highest voltage setting.
Link to PDF: Model HSP 48-30
https://www.electro-tech-online.com/custompdfs/2007/05/1461869.pdf

I like the hall effect sensors, but think the simplicity of the shunt(s) far outweighs all other factors. I have settled on using these panel meters and can get the meter and shunt for dirt cheap. They will work for both plant voltage and current (with appropriate shunts). By adding a simple switch, I can use 1 meter to display readings from either of the 2 shunts (or more) if I so desire.
**broken link removed**

Dialtone said:

I like the hall effect sensors, but think the simplicity of the shunt(s) far outweighs all other factors. I have settled on using these panel meters and can get the meter and shunt for dirt cheap. They will work for both plant voltage and current (with appropriate shunts). By adding a simple switch, I can use 1 meter to display readings from either of the 2 shunts (or more) if I so desire.

Click to expand...

Have you seen some of LEM's hall effect, just supply them with a 12V to 15V bipolar power supply and they give you a +-10V signal directly proportional to the current which you can use to drive a panel meter with a suitable scale.

The datasheet below is just an example - they come inall shapes and sizes and can be purchased from RS components. If the bipolar supply is an issue then you can buy little DC-DC convertors or you might be able to tap the batteries at some point.
https://www.ortodoxism.ro/datasheets/lem/LA205-S.pdf

I was told that the grounding of the (+) side of 48V was to prevent erosion of the conductors in case of leakage to ground.

A (-) ground system would erode the conductors in the event of leakage.

I'm thinking of automotive and on boat installations where you don't want the vehicle/vessel's body corroding.

What you're saying makes sense. I just assumed that mose DC installations were negitive earth not positive earthed.