Is it dangerous ... (dot dot dot)

[EFK]

... to NOT have the negative speaker terminal grounded? I was thinking about this today. MArshalls - for example - use floating jacks and ground the negative. On a Fender w/ the jacks grounded right at the chassis, the negative once again is therefore grounded. However, I have a couple of old 50s amps w/ the output transformer mounted right on the speaker, and the positive and negative leads are just connected right to the speaker terminals w/ no ground wire running off the neg. back to a chassis ground or any kind of ground. So, is this considered dangerous in any way?

[Chuck H]

Hmmmm.?. Good question. I'd be interested in other opinions too. My gut is telling me it's not dangerous. The OT is isolating the output from HV and even with it grounded the AC is still there. In fact, with the output grounded you can touch the speaker + terminal and ground (chassis, switch, handle screw, etc) and become part of the complete circuit. If the output is not grounded you have to touch specifically both speaker terminals to do this. I know it's common advice here for some hum abatement processes to isolate the output. It's my understanding that you can have some kind of trouble like noise, hum or instability isolating the output if your amp uses a NFB loop. The output and NFB circuit are supposed to be grounded in the same place, but I know people have isolated their output on NFB amps without a problem so perhaps it's just a best practice rule because problems pop up sometimes.

[J M Fahey]

Strictly speaking, by itself, no, it's not dangerous.
The voice coil itself is "floating", only cares about voltage applied *across* its terminal and the resulting current that flows *through* it.
The only case I've seen where there was some importance attached to grounding was in distributed-speaker PA systems, where speakers are roof , wall or pole mounted, and where speaker frames are grounded; I was told it was done to avoid static buildup.
I received 2 differing, non-consistent explanations:
1) to avoid static voltage buildup, which would discharge through sparks, which in some cases might damage the voice coil ... or ...
2) to avoid static voltage buildup, which would increase corrosion in humid climates , by electrolytic action.
Personally never could confirm either of them, although have seen such frame grounding.
Have also seen some speakers with a third lug, riveted to the frame.
Have also seen it in Japanese amps (think Roland JC120/G100 Yamahas) with said 3rd lug grounded either to the negative lug or straight to the chassis.
Speakers are, of course, "floating" by themselves since they are bolted to wooden panels.

[Enzo]

Like anything else, it is the overall system that matters, not the parts.

No, there is no danger from having the OT secondary connected to the speaker load without also being referenced to ground. It just usually is.

Hum and instability do not occur because the output is grounded or not, they occur when whatever good design they start with is compromised by some change. You can ground the output jack to chassis, or float it and ground the negative terminal with a wire to somewhere else, or not ground it at all. But whatever you do, it is part of the overall grounding scheme. Ground loops occur when two ground points get connected together but want to be at differing ground potentials. Current then flows through that connection, causing hum.

NFB? Yes! The NFB line samples the output signal and feeds it back to an earlier stage. Since those stages are generally referenced to ground, the NFB line needs to be, otherwise we are connecting one side of the signal, without reference.

[Chuck H]

Hum and instability do not occur because the output is grounded or not, they occur when whatever good design they start with is compromised by some change. You can ground the output jack to chassis, or float it and ground the negative terminal with a wire to somewhere else, or not ground it at all. But whatever you do, it is part of the overall grounding scheme. Ground loops occur when two ground points get connected together but want to be at differing ground potentials. Current then flows through that connection, causing hum.

Thank's Enzo. I did know that but I WAS vauge.

NFB? Yes! The NFB line samples the output signal and feeds it back to an earlier stage. Since those stages are generally referenced to ground, the NFB line needs to be, otherwise we are connecting one side of the signal, without reference.

Of course. The NFB circuit absolutely DOES have to have a 0V reference. But the output jack doesn't need to be grounded to provide voltage for the NFB circuit. The source impedance for the NFB circuit is sooo low that the comparably high resistance voltage divider for the NFB circuit should not be affected either way.

[Enzo]

Of course. The NFB circuit absolutely DOES have to have a 0V reference. But the output jack doesn't need to be grounded to provide voltage for the NFB circuit. The source impedance for the NFB circuit is sooo low that the comparably high resistance voltage divider for the NFB circuit should not be affected either way.

Not sure I understand what you are saying. For the NFB to work, the output signal must be ground referenced, I know you said that part. The output jack itself need not be the ground point, but SOMETHING on the output winding has to be grounded. Or as is seen rarely, a dedicated NFB winding could be separate from the speaker winding.

The output winding is a voltage source by itself - one end to the other - but for that signal to go anywhere and do anything, a complete circuit is required. Like a guitar pickup, we cannot just wire the hot end to an amp without grounding the other side, unless we have a differential amp input like a microphone jack.

[Chuck H]

I'll need to try it for myself I guess. It just seemed to me that at 8ohms (or four or sixteen, etc) the output impedance is low enough that compared to the voltage divider, that makes up the average NFB loop, an absolute 0V reverence wouldn't be needed. The voltage divider 0V reference resistor would be grounded so that it has a common relationship with the PI or preamp circuit it's connected to. I will check it out for myself the next time I have an amp open.

[Enzo]

But that low impedance floating won't send anything down the NFB at all, other than whatever couples through teh air as crosstalk of a sort..

This makes me think of the tube rectifier circuit with the 5v heater. That 5vAC floats, so it can be connected to 500v of B+ safely because there is no reference to ground. But that 5v of AC is not added to the top of the B+ because there is no common reference for it.

NFB sends the output signal back to the amp. The voltage divider does substantially reduce the amplitude of that output signal, yes. But it is still a signal. Inagine if instead of feeding that sample back to the same amp, you wanted to send it to another amp. You couldn't wire a line from the hot side without also sending along a ground.

This is a signal voltage, so regardless of the impedance, what we care about is the signal voltage across the output. That is what we sample. 23 volts of signal is 23v of signal, regardless of the impedance it is across. SO if we want 2.3v of NFB, we send that 23v signal through a 10 to 1 divider and then use the result.

Maybe I can put it that way, say we want a line out. Take my 23v example, and we want a very hot 2.3v line out. We can't just take a wire from the hot and run it to a couple resistors unless the other end of those resistors was conected to the cold.

[Enzo]

Have you been reading Lowell's problem with the distorted DI on his SVT over in repair? It really is the exact same thing, at least as regards to grounding or not on a transformer winding.

[EFK]

Thanks for the responses! Very interesting - very. I guess I was thinking of it in terms of a 'catastrophic' case issue. For example, what if in the above example w/ the OT mounted on the speaker, the OT self destructs or something and the HV punches through the insulation to the output. Then someone pokes around back there wondering what the heck just happened, maybe there is still juice running into it for some silly reason or another and our investigative 'someone' who didn't immediately hit the power grabs onto those speaker leads to see if "hmmm, maybe something became disconnected..." and then he becomes a path to ground. ???

[Jazz P Bass]

That right there is a very good reason for earthing /grounding the speaker frame itself.

[Enzo]

That is a very convoluted scenario of failure. A short from primary to secondary of the OT - that doesn't just short out in general taking a fuse - that leaves B+ on the speaker leads. Grounding the speaker frame then would not remove the B+ from the speaker terminals. But all in all, this seems to me like looking for weird exceptions at the expense of more common concerns. Like forgetting seat belts in your car but concerning yourself over whether a bowling ball thrown off an overpass could come through the windshield.

[Enzo]

Oh, and the bowling ball? It would.

[Chuck H]

Screw seat belts then. I'm betting bullet proof tempered glass for my windshield (I have a bowling ball phobia)

[Enzo]

And it is not just bowling balls, frozen turkeys are just as lethal. However, frozen turkeys are less likely to get you on the bounce.


And let us not even talk about live turkeys. "As God is my witness, I thought turkeys could fly."