Discussion - Earth, ground and chassis issues

Thread Starter

bloguetronica

Joined Apr 27, 2007
1,541
This is an issue that is rather concerning me. Not a long while ago, someone advised me that I should connect the earth to the ground of circuits, with respect to my power supplies. I didn't considered this, because I previewed that two of them could be connected in "series", so to speak, that is, the negative terminal of one to the positive terminal of another.
Also, I saw one project of a PSU that had the ground terminal (in that project, the earth was connected to the chassis, as I have) connected to both live and neutral by capacitors. It made some sense, but was strange.

With this topic I want to open a discussion about this theme, that is, about how, earth, ground and chassis should be connected.
 

beenthere

Joined Apr 20, 2004
15,819
Just as a general statement, there is never a situation in which the chassis/case should not be tied to earth ground. This positively eliminates any shock hazard. I don't believe there can ever be an exception to this rule.

The case you mention where the hot and neutral had capacitors to earth was for noise elimination. Look at the diagram for any power line filter, and you will see a similar arrangement. The capacitors are small in value, and provide a low impedance path for RF frequency signals.

With respect to the circuit inside, the connection of the PS ground to the chassis/earth is more optional. If the circuitry needs no outside leads, then letting the ground float will probably not be significant. There aren't too many devices where no external signals come in, though.

Your example of two power supplies connected in series for a positive and a negative rail did not preclude bonding their common ground to the chassis. It simply would have referred ground to the earth potential, and provided a path to earth for noise. If the output of the power supplies were properly isolated from the line input, no undesired effect would have resulted.

In the case of multiple devices connected together, using earth as circuit ground is mandatory. If any individual circuit floats, there is a strong likelihood of a circulating current in the signal path. This will almost certainly result in noise in the signal of interest, and a degraded output. this applies to scientific equipment, but audio equipment will have huge hum problems if any amplifier/preamp has floating circuitry.

Perhaps you could share your reservations about bonding earth and circuit /power supply ground together?
 

Thread Starter

bloguetronica

Joined Apr 27, 2007
1,541
Your example of two power supplies connected in series for a positive and a negative rail did not preclude bonding their common ground to the chassis. It simply would have referred ground to the earth potential, and provided a path to earth for noise. If the output of the power supplies were properly isolated from the line input, no undesired effect would have resulted.

In the case of multiple devices connected together, using earth as circuit ground is mandatory. If any individual circuit floats, there is a strong likelihood of a circulating current in the signal path. This will almost certainly result in noise in the signal of interest, and a degraded output. this applies to scientific equipment, but audio equipment will have huge hum problems if any amplifier/preamp has floating circuitry.

Perhaps you could share your reservations about bonding earth and circuit /power supply ground together?
Well, earth is connected to chassis in both power supplies. However, I left ground unconnected to earth. My concern on this was to cover a situation where you have two of these power supplies with the positive of one connected to the negative of the other. This would result in a short if the ground on both was connected to the earth.
 

beenthere

Joined Apr 20, 2004
15,819
Unless the power supplies in qustion were not isolated from the line input, how would bonding the common to earth create a short? As fas as I can see, it only makes ground common to earth.

Ring out a bench supply sometime - you'll find that the case and the common lead are bonded.
 

mrmeval

Joined Jun 30, 2006
833
Metal cases are tied to earth ground. If for some reason a wire inside breaks or other bad things happen then the case will not be hot and electrocute someone.

At one time I had one of the old Arvin "Widow maker" radios. They work excellent. The case is metal and it is directly tied to the AC line...using a non-polarized cord. As best I can tell they did this to use the case as one half of the wiring probably to save money.

Plug it in one way and all is well if the neutral isn't floating a bit high, other wise you might get little nips. If it's plugged in the other way then the case is hot. If you're touching the case and ground at that point you will meet St. Vidicon in person.

I think this is like the one I had but it's been a long time and I remember it's lethality but not exactly what it looked like
http://cgi.ebay.com/1948-Arvin-tube...7702441QQihZ009QQcategoryZ38034QQcmdZViewItem
 

Thread Starter

bloguetronica

Joined Apr 27, 2007
1,541
Unless the power supplies in qustion were not isolated from the line input, how would bonding the common to earth create a short? As fas as I can see, it only makes ground common to earth.

Ring out a bench supply sometime - you'll find that the case and the common lead are bonded.
You can see shematics of these PSU's in the projects section. As you can see, if you take two of these power supplies and tie the positive of one to the negative of other, a shortcut will not occur since the ground terminals (now at different potentials) are not connected throught the earth.
Now imagine that I had the ground tied to the earth in both power supplies. A short would occur now that their respective grounds, at different potentials, were connected throught the earth, witch is the same in one building. As you see, the only thing connected to the earth is the case, witch is to be made of metal.

But here is to be discussed the best way to deal with earth, case and ground.
 

subtech

Joined Nov 21, 2006
123
I've no comments just yet, but earth, ground, and common are about the three most misunderstood words in all of the electrical world.
This discussion should be an eye opener.
 

wireaddict

Joined Nov 1, 2006
133
By connecting the positive terminal from one side of the supply to the negative terminal on the other section of the supply AND IF [!] both sections are isolated from each other you form a series circuit just like connecting two batteries in series. The resultant voltasge between positive & negative terminals that aren't tied together are the sum of the two original P/S voltages. The wire that ties the two sections together is also roughly similar to US single-phase 240 VAC residential wiring [except that residence uses AC & the PS output is DC]. The common between the two 120 VAC mains or 'hots' in the house [called the 'neutral' wire] is also grounded. In both cases the two 120 VAC & the +/- DC PS outputs are hot with respect to to the common, neutral or ground & no short circuit condition exists.

However, having said all that, if the power supply uses two diode bridges which are connected to the same transformer [secondary] winding including one with a center tap, you absolutely can't connect any combination of positive & negative terminals together without at least two of the diodes attempting to short out the secondary winding. & this will happen whether this pos.-to-neg. connection is grounded or not. Another situation which will cause immediate fireworks for a similar reason is to ground either the positive or negative PS outputs if the secondary has a grounded center tap. Fortunately, these situations are rare, particularly on bench top power supplies.

So, to summerize, neither connecting sections of a PS with isolated outputs together nor grounding this connection will cause harm any more than doing so with the common between two batteries. I hope this makes sense!
 
Top