Interconnect Ground Cable Ground Loops

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I've bought some bulk Belden cable, got a deal on RCA cable ends, and before installing new single-end cables on my system, I've been reading up on avoiding ground loops, Whitlock, and others. Whitlock mentions bundling of signal cables - why not have shields connected on one side for all cables to a source, and then run a nice big ground/earth cable, the only one, between each chassis? Only one safety ground would then be run to the building power.
 
I am out of my comfort zone here and CBW but two things come to mind. First, chasis are often tied to the signal ground through a small resistor so you would need to tie to signal gound not the chasis. Secound, ground loops develop in the samll differences in impedence between the grounds so you would need to devise some form of star grounding.
 
Thanks for replying, jj - - On Rod Elliott's grounding/earthing page

Earthing (Grounding) Your Hi-Fi - Tricks and Techniques

figure 4 has a schematic for a high current safety loop breaker circuit that puts 10 ohms between the signal ground and the safety earth except if there is a fault. Installing these would be a project that would never be finished by me. I checked the schematic for one of my Adcom amps and there are a dozen chassis grounds. So, let's add the assumption that I've checked for those 5 ohm or 10 resistors between signal or zero voltage point and chassis ground, and that they have been bypassed if present.

1) Right now I have my gear plugged into 2 mains wall outlets, maybe a few ohms of impedance in the earth/ground between them, and of course it hums like it doesn't know the words. If I was reckless I could disconnect one of the ground pins to the wall outlet. Potentially, if there was a fault, 15 amps could burn through the cable shields on the way to the ground/earth. A big ground bus or a star, with all cable shields cut on one end, would cut this ground loop and allow a breaker to trip if the worst fault happened in a chassis box.

2) If all remains safe, I would place the severed end of each cable on, say, the far end from my control box. No current would flow through the shields, but through the fat green wire. The signal would be still shielded from environmental garbage. Elliot touches on this in the article referenced above while discussing switching noise: "It might be possible to reduce this noise by installing a heavy earth strap that joins each chassis."

What do you think? Are there any obvious pitfalls to cutting the shields and connecting a ground star, or should I say an earth strap?
 
This is what Audio Precision does:
 

Attachments

  • AP ground wire.png
    AP ground wire.png
    74 KB · Views: 666
Hey, there it is!
http://www.ap.com/download/file/149
"Floating BNC connectors help prevent ground loops."
I'm looking at a 6 gauge very fine strand (3000) cable for normally sold for lighting and portable power. This would act as a massive ground bus going down my equipment cabinet and over to the video monitor and cable box. Some nice Monster-type spade lug leads would then go to the back of each chassis. Also I'd be ready if I ever have to jump-start a locomotive.
 
Last edited:
Thanks for replying, jj - - On Rod Elliott's grounding/earthing page

Earthing (Grounding) Your Hi-Fi - Tricks and Techniques
There is a problem with Rod's recommendation.

A 35 amp bridge can fail open during a fault.

There are several types of bridge constructions out there I am familiar with intimately. One type has an aluminum case, and the diodes are arranged such that explosive forces will push against the side wall of the case. The case will constrain the structure sufficiently long that the silicon will fully short during melt. Some bridges are designed where the chip is horizontal to the mounting surface such that the forces are up in the case, where only epoxy is the constraint. The end user will not know what type they have selected, so may or may not have a safe arrangement.

I would recommend a fast blow fuse on the hot line just as it enters the case.

jn
 
Jn,
I tested and reported the Disconnecting Network (DN).
It consisted of the bridge rectifier, shorted across both diagonals. Tappings off adjacent terminals.
A 100nF ceramic cap (x7r) and a 10r 600mW metal film.
The fuse holder exploded as both fuses (live and neutral) ruptured.
None of the components in the DN were damaged in any way that I could measure.
I suspect a Fault current approaching a kA passed through the DN.
 
Jn,
I tested and reported the Disconnecting Network (DN).
It consisted of the bridge rectifier, shorted across both diagonals. Tappings off adjacent terminals.
A 100nF ceramic cap (x7r) and a 10r 600mW metal film.
The fuse holder exploded as both fuses (live and neutral) ruptured.
None of the components in the DN were damaged in any way that I could measure.
I suspect a Fault current approaching a kA passed through the DN.

Shorting across both diagonals completely removes the bridge. Did you mean shorting the positive terminal to the negative one? edit: ah, you said tappings off adjacent terminals..nevermind..

Never, ever...and i mean...NEVER put a fuse into the neutral line. Ganged breakers ok, not fuses.

What did you test? I was speaking about a hot to ground short, not hot to neutral. Hot to neutral doesn't exercise the bridge.

Was your test per the schematic?

jn
 
My report spelled out the conditions of the test.
I connected the Live to the Main Audio Ground. Then switched on at the socket outlet.
I think my test pre-dated ESP's webpage. ESP copied his information from this Forum. But he has a misleading diagram. The PE should be connected directly to Chassis. The DN should also be connected to Chassis, but this connection does not need to be near the PE connection. His diagram shows the PE running through the DN and then to Chassis . This "running through" is WRONG.

The Fuse holder incorporated into the IEC filter happened to have both poles fused. I fitted a T4A into the Live pole and a T8A into the Neutral pole (I think I described that as well).
Both fuses exploded. I never found much of the fuses.
One end cap was "gone". Much of the glass was "gone". One end cap, if I remember correctly, was still inside the fuse holder. Two others were on the floor on the other side of the room.
No one else volunteered to do this test when I suggested we should check what we were recommending. Eventually I stepped up to the plate and did it.

But the point:
none of the components in the DN were damaged, neither electrically, nor visibly as far as I could tell.
 
Last edited:
Shorting across both diagonals completely removes the bridge. Did you mean shorting the positive terminal to the negative one? edit: ah, you said tappings off adjacent terminals..nevermind..
I don't get much wrong and especially if it relates to Safety.
If I had this wrong I would post a big OOPS and correct my error.

Now that you have read it, the dual shorting achieves paralleling of the diodes and only one Vf as the voltage drop.
This Vf could be very high at 1kA or 2kA.
It would be nice to have some experimental data to complete this test & report.
Do you have facilities to measure the kA Fault Current and Vf while passing that almost instantaneous Fault Current?

ESPs single shorting and tappings off opposite sides leaves single diodes in series, for double the Vf and near half the current capability.
There is no advantage in having a doubled Vf.
There is a very definite advantage in having paralleled diodes.
 
Last edited:
I didn't understand your blowing the neutral fuse. I see you're across the pond. Your neutral is hot to ground? On this side of the pond, neutral fuses are frowned upon.

The fuses losing containment is not good, perhaps slow blow's aren't nice there.

The KBPC35 series of bridges will be in the 1.8 to 2 volt range at a kiloamp. But it is rated to survive a single 400 amp half sine superimposed on rated load (JEDEC method). It's rated up to 660 Amps squared seconds for fusing.

jn
 
Just for confirmation, we in the UK are also not supposed to fuse the Neutral.
It just happens that, that particular IEC filter came with fusing for both poles and a switch. I wanted the filter and the switch and this was the only IEC socket that I had in stock that gave this facility.

To help prevent blowing of the Neural fuse alone, I doubled the fuse rating from the Live T4A to T8A.

For the test I did not change that, I just linked the Live to Main Audio Ground to "prove" that my "concoction" was safely protecting from electrocution.

All my other IEC filters do not have any fuses. I see no problem, other than I could insert an even bigger Neutral fuse or even solder a 1mm copper link through a "dummy" fuse.

Unfortunate that the fuse door blew open, but the explosion was quite extreme.
 
I am still very very unhappy that the neutral fuse blew.:( From your description, it should not have done so. What's going on??

I also am unhappy that there was a containment breach. Fuses are not supposed to do that.

I do know that they voltage rate the bigger fuses for that reason. But I don't know about the fuses you chose. If they were v rated for your application, maybe it means the fuses should be fast blow. And rated to protect the wires only, not the equipment.

I also recall the load panel breakers I use as being 10kA fault current rated. Is it possible that the fuses or holder you chose have ratings like that as well?

jn
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.