The deeper mysteries of 'ground'?

[Al (astral highway)]

Hi folks,

I'm trying to deepen my understanding of 'ground.' I know it's critical and perhaps also taken for granted in some ways. I can conceptualise it, of course as
a voltage reference - 0V, by which all other DC voltages, more positive or more negative, are defined and, with the help of associated passive or active circuitry, are kept constant if that is the goal.


It's very clear how a physically large ground plane is used both in RF and precision audio applications, to reduce 'noise' , by passing noisy wiring as closely as possible to the ground.

Some aspects of ground are not clear to me, however. What if a ground was composed of a less -than -perfect conductor, eg made of brass? What would be the effect on the circuit overall? This doesn't arise in practice, of course, but I like to do thought-experiments to deepen my understanding...

It's also not always clear when a component should be grounded. Would we ground a theorectical aluminium cap in a high voltage power supply which is isolated from its case, or just let it float?? When can components 'float'?

I'm sure 'ground' is often tagged as elementary but given its critical importance, I wonder what else might come up in a discussion here?

[ms660]

Depends what's meant by ground.

The earth (planet/soil etc)

Chassis.

Common signal rail such as HT-ve (in some cases)

This could be a long one.

Lawrence.

[Craig Sawyers]

The existence of a 0V reference need have nothing to do with ground. For example airplanes, or even more extreme satellites.

All you need to make a low noise system is an effective Faraday cage around the signal carrying circuits. That can be conducting sealed boxes connected together with shielded cable.

The difficulty comes when you connect to mains, since there is then a conflict that can give rise to either hum loops, where current flows in the cable shields and modulates the signal, or capacitive noise coupling. Capacitive coupling can be minimised by using an interwinding shield in the transformers.

Another way of defeating hum loops in unbalanced audio systems is to use a "ground lift" resistor between the signal 0V and safety ground. Around 1k to 4.7k 1.4W for small signal things like preamps etc, and 10ohms to 47ohms 5W ohms for power amps

All that notwithstanding, every chassis absolutely has to be connected to green/yellow safety ground. And there is a defined way of doing this - a chassis safety ground has to be to a dedicated screw that cannot be used for anything else. The safety earth wire has to be terminated in a crimped ring terminal. There is a sequence of nuts and spring washers that must be followed, and the gauge of wire has to take any fault current without melting before the fuse blows.

Craig

[Al (astral highway)]

Originally Posted by Craig Sawyers

The existence of a 0V reference need have nothing to do with ground. For example airplanes, or even more extreme satellites.

Ok! so I'm not defining (eg chassis) ground and earth as having the same meaning - just the opposite. Indeed, it will be interesting to consider the times when they are the same thing and when, categorically, they are not.

[Lucien Nunes]

Would we ground a theorectical aluminium cap in a high voltage power supply which is isolated from its case, or just let it float??

The can of an electrolytic is not always designed to be used as its negative terminal, but nor are the two usually isolated to withstand any voltage. If the negative lead is connected to the chassis nearby, the can can be too. If the capacitor is connected to some other location e.g. a star point, or its negative is not at ground potential, then the can must usually float. The reservoir of a negative supply will usually need the can isolated to withstand the supply voltage.

Conversely, a paper-in-oil will usually withstand working volts from the can to either electrode and the can would be grounded if beneficial as it will be capacitively coupled to the cap outer foil. For example, the metal-cased motor-run caps in my Compton organs are usually connected to mains earth both for safety and to stop them radiating mains hum to the high-impedance audio circuits.

[TonyDuell]

One of Don Vonada's 'laws' (from the computer manufacturer Digital Equipment Corporation') is 'There is no such thing as ground'.

This has 2 meanings, one obvious the other less so.

The first can be shown by the fact that voltmeters have 2 leads. In other words, when you measure a voltage, you have to specify what you are measuring it with respect to. If there is an obvious reference point, then that is commonly called 'ground' for that circuit. In a TTL-based logic circuit, it's the negative side of the 5V power supply.

The second is that even when you have defined what you mean by the common rail, and defined a point on it where you connect the voltmeter probe, all other connections have some non-zero impedance to that point. In theory you can make the resistance as low as you like, but at higher frequencies (I come from the computer hardware side, where even quite slow waveforms can have components in the 100MHz range) it's the inductance that causes problems. This can mean that 2 points -- say the 'ground' pins of 2 different ICs on a PCB -- can have a significant voltage drop between them, Actually, it's a nasty random-ish voltage signal with some very high frequency components. This 'ground bounce' can cause some very interesting problems.

Going back to the first point, in a switch-mode power supply there are often 3 grounds, not electrically conncted together (in fact if you did connect 2 of them, you would end up with blown fuses or worse). They are :

1) Mains safety earth -- the green/yellow wire in the mains cable. Connected to mains noise filter capacitors, etc

2) The negative side of the mains bridge rectifer. The reference point for the primary side of the supply -- the chopper circuit, etc. Needless to say this one is 'live' with respect to safety ground.

3) The common rail of the output side. Often (but not always) the negative side of the main output

[Malcolm G6ANZ]

High currents can also have interesting effects on 'ground'. I worked with linear accelerators part of them had a pulse modulator producing 40kV at several hundred amps. Everything in that part of the circuit was referenced back to a point called pulse earth. Due to the high currents there could be a considerable voltage drop between pulse earth and chassis 0v and even the mains earth.

[dseymo1]

This is an interesting discussion.
I've mentioned elsewhere that I'm reading about Tesla's research, and struggling with the theory.
One of the many concepts I find difficult is his equivalence of true ground and any reasonably-sized isolated plane. His 'single conductor' light sources are simple enough to to understand when the current is returned to Earth, but not so when the 'return' is just a metal lampshade, for example (but then he considers the whole setup to be electrostatic, which can't be right, can it?).

[RogerEvans]

The Rutherford Lab near Harwell used to have a proton synchrotron with several large pulsed magnets running at a few Hertz. Two ground probes a few metres apart would show around 50V difference in sync with the pulses due to inductive ground currents and soil conductivity.

Roger

[Jon_G4MDC]

TV studios are fun too.
There is Studio ground - lights, props etc and Technical Ground - cameras mics &c.
(I hope i remembered the correct terminology)

I found a fault with an antenna mounted on the lighting gallery. It was fed from equipment on technical ground but the mounting was studio ground. Only a few Volts of difference in potential but it was passing 60A. Fat sparks when connecting up!

I

[TonyDuell]

Yes, currents flowing in earth connections (which have impedance of course) can cause interesting problems. This is one reason for the 'star earthing' used in may devices where all earth connections are brought to one point (perhaps the -ve terminal of the smoothing capacitor). The idea is to ensure there is no impedance common to 2 different circuits, so the voltage drop caused by a particular current has no effect on anything else.

It's also why in some cases a ground plane (as an internal layer of a PCB) has 'slots' in it. To force currents to take a particular route back to the PSU (say). So that the HF noise from (say) a microcontroller circuit can't end up on the ground of a sensitive DC amplfiier.

[G8HQP Dave]

What if a ground was composed of a less -than -perfect conductor, eg made of brass?

All ground connections are made from less than perfect conductors. Forgetting this is what makes people make grounding errors. My usual advice is to design grounds as though every 'ground' connection is a resistor (and/or, where relevant, an inductor). This is good advice because it happens to be true of all ground connections.

Other advice:
current always flows in loops
a 'star ground' is not a star but a very short (usually random) bus
there is no such thing as 'a voltage'

[ricard]

Quite.

Another way of putting it is that 'ground' is an engineering simplification to have a common '0V' reference all through the circuitry which simplifies the design work.

[Al (astral highway)]

Originally Posted by TonyDuell

... even when you have defined what you mean by the common rail, and defined a point on it where you connect the voltmeter probe, all other connections have some non-zero impedance to that point. In theory you can make the resistance as low as you like, but at higher frequencies (I come from the computer hardware side, where even quite slow waveforms can have components in the 100MHz range) it's the inductance that causes problems. This can mean that 2 points -- say the 'ground' pins of 2 different ICs on a PCB -- can have a significant voltage drop between them, Actually, it's a nasty random-ish voltage signal with some very high frequency components. This 'ground bounce' can cause some very interesting problems.

I had an intuitive sense of this, but it helps me to see a clearly delineated practical problem that arises commonly -- thank you for this insight. Interesting thread developing here!

[Craig Sawyers]

There is no spoon ;-)

[G4_Pete]

I used to teach aprentices fault finding and correct fusing on DC distribution systems that incorporated ,Grounded systems, Foating distribution systems, earth balanced via low impedance, and earth balenced via high impedance systems. But I never ceased to get phone calls on "strange Voltages" on sytems when people used the wrong reference point for measurements, especially when leakage fauts can slew potentials in these systems in any direction about chosen refererence points. That was even without getting into discussions about potential differences between all the different grounding systems on a typical complex industrial plant and the affects of local circulating currents. I used to ask would you measure your car battery by putting one lead on the battery and the other by your feet on the ground , Ohh and what is the impedance of the meter are you using . Long pause followed by I will check again!!

Perhaps it is understanding the term "Ground/Earth" in the context of whatever system you are looking at that is important.
Pete

[ms660]

The term earth in vehicle electrics is another one.

Lawrence.

[Mr Moose]

Hello,
Leakage from damaged underground high voltage cables can cause the ground to be above 0V. This can be particularly dangerous for horses with their metal shoes and their legs far apart, as this incident from Newbury racecourse demonstrates.
http://www.telegraph.co.uk/news/ukne...acecourse.html
Yours, Richard

[mole42uk]

I wouldn't believe too much of this story: " And because Fenix Two and Marching Song were wearing steel shoes, the current passed through their bodies unlike the other horses that were wearing aluminium plates. "

[ms660]

Originally Posted by mole42uk

I wouldn't believe too much of this story: " And because Fenix Two and Marching Song were wearing steel shoes, the current passed through their bodies unlike the other horses that were wearing aluminium plates. "

That quote does not say that the horses that had aluminum racing plates escaped electrocution because of the fact that they had them fitted, maybe they weren't in the same gradient zone.

Lawrence.