Do real circuits need to be physically grounded?

[atlbraves49]

When we do labs in classes we never connect anything to physical ground (always just "common" node), but in an actual implementation, does a circuit need to be connected to physical ground?

For example, I have a circuit running 7 amps into a prototype board inside a metal box. Should I connect the "ground" side of the circuit to physical ground... i.e. should i run a wire from the prototype board to touch the metal box?

 

[waht]

That really depends on the application. If you ground a circuit to physical ground, then a potential difference between the circuit and the physical ground will be zero, hence no current will flow. In contrast if you don't ground, then some potential difference might develop ranging from a few millivolts to kilovolts, and that might damage sensitive electronics, or even shock you if were to handle it.

 

[Bob S]

For shielding purposes, the ground on the circuit board should be connnected to the shield (metal box). The common on the power supply should be connected to the shield as it comes through it, and for low noise-level circuits (~ 1 nV/root Hz) the dc voltages ideally should be bypassed to the shield as they come through the shield. If the power supply common is grounded at the power supply, then the metal shielded should not be grounded, or else there may be ground loops.

I had a particularly hazardous situation at MIT a few years ago. For cost reasons, the big power transformer outside our lab had a "Y" (wye) secondary (not delta) with 120 V three phase power, so there was no obvious ground for all the single phase power outputs to balance all outputs, so each 110 v plug outlet had hot, common, and ground as 110 volt 3 phase power on all the workbenches. Different work bench grounds sometimes were 110 volts apart!.

 

[atlbraves49]

heres an example of the circuit I am talking about:

http://img7.imageshack.us/img7/9600/samplecircuitdrawing.jpg

Where the thermoelectrics produce a voltage/current, which is dumped across the resistor/MOSFET combination. The voltage across the MOSFET/Resistor in series, is measured by a DAQ. So my question is, do i need that ground there (and the implementation just means running a wire from that side of the circuit to the metal box?), or should it just go straight from the thermoelectrics, through the resistor/MOSFET, back to the thermoelectrics?

 

[MATLABdude]

In this instance, to get the circuit functional, you just need the common ground. To get the circuit safe(r), you probably want the case ground going to earth. Note that the power supply ground (for your thermoelectrics) and/or the PC ground may be connecting to the Earth ground already.

 

[atlbraves49]

In this instance, to get the circuit functional, you just need the common ground. To get the circuit safe(r), you probably want the case ground going to earth. Note that the power supply ground (for your thermoelectrics) and/or the PC ground may be connecting to the Earth ground already.

The thermoelectrics are the power supply. They have two leads coming out of them, one goes to the top side of the circuit, the other goes to the bottom side.

I have 4 of these circuits in the metal box. Will it be safe for me to run a wire from the bottom side of each circuit to the metal box? And will it function correctly? Or should I just not.

 

[Born2bwire]

The thermoelectrics are the power supply. They have two leads coming out of them, one goes to the top side of the circuit, the other goes to the bottom side.

I have 4 of these circuits in the metal box. Will it be safe for me to run a wire from the bottom side of each circuit to the metal box? And will it function correctly? Or should I just not.

Not every circuit has a physical ground. For example, any portable battery device does not have a physical Earth ground. But if you really need to define a ground you can just use the negative terminal of a battery or create a virtual ground. In your case, your power supply is supplying the signal and ground terminals. You should not force the ground to Earth ground because they may not be the same. If that's the case, then there could be a voltage between the two and you'll keep dumping current.

 

[EEJaime]

It seems there is some confusion here regarding the grounded circuit conductor, (neutral leg), and the equipment grounding conductor, (earthed conductor). The "particularly hazardous" system Bob S described seems to be the standard 208Y/120V, 3ph,4 wire system we design into every building we do. The 3 phases, A, B and C are wye connected tothe center star point, (Neutral point). The single phase Voltages phase to neutral are at 120V, the phase to phase Voltage is 208 V. The Neutral point of the secondary side of the service is bonded to ground at the service entrance only. The feed to a 120V load will include a 'Hot', a 'Neutral' and an equipment ground conductor. the load is connected to the Hot and Neutral. All boxes, panel enclosures, outlet boxes, the ground lug on the receptacle, etc..., are grounded to the equipment ground conductor. The system nuetral is the current return path to complete the circuit. The equipment ground should never be connected into the circuit directly.

I don't know if I've clarified this or made matters worse, but there you go.

Regards