Speaker output grounding in no-NFB amps

[Rhodesplyr]

In vintage tube amps with no negative feedback from the output transformer secondary, you still usually see one side of the secondary grounded to the chassis.

I think I remember at some point reading about why it was important to tie one side of the secondary to ground even when you don't need a ground reference for NFB, but I can't remember the reasoning. Does it have to do with the potential for RF pickup by the speaker wires? And sometimes the speaker frame itself is connected to the ground wire.

What happens if you don't connect the OPT secondary to circuit or chassis ground? And when grounding the secondary where should the connection go? Direct to chassis? To the main ground buss?

[J M Fahey]

In vintage tube amps with no negative feedback from the output transformer secondary, you still usually see one side of the secondary grounded to the chassis.

True.
That said, all of old amps used a Switchcraft ("Fender") type jack bolted to chassis , so like it or not, jack/speaker return terminal was grounded by definition.
Only the English amps used pastic Cliff type jacks, where grounding was optional.

I think I remember at some point reading about why it was important to tie one side of the secondary to ground even when you don't need a ground reference for NFB, but I can't remember the reasoning. Does it have to do with the potential for RF pickup by the speaker wires? And sometimes the speaker frame itself is connected to the ground wire.

In the old days, speaker frames had to be conncted to ground to avoid DC static causing arcing which was heard as random nasty clicks, that on distributed PA systems, fed from line transformers.
Remember in the old days there were no specific Guitar/Bass/PA/Organ/Home audio speakers, Jensen made P or C12"something" , Celestion made G12 ... and that was it.
Since speaker could be used "anywhere" some factories just grounded one terminal for you.

What happens if you don't connect the OPT secondary to circuit or chassis ground?

No big deal

And when grounding the secondary where should the connection go? Direct to chassis? To the main ground buss?

If itīs just " a reference" and no significant ground current flows, no big deal.
Meaning current exits an OT tap, flows through speaker and return goes straight to other end of OT winding on its own wire, no big deal.
Now if speaker jack is mechanically grounded where itīs screwed, and OT common wire goes to chassis "somewhere else" so significant current flows through chassis metal, yes, a voltage difference will appear through that speaker return current, chassis does not have zero ohms, and you **might** have stability problems.

[mozz]

How about this Admiral console el84 amp i picked up today? Negative of output transformer goes through cathode resistor to ground. Picture needs zoom.

[Rhodesplyr]

That said, all of old amps used a Switchcraft ("Fender") type jack bolted to chassis , so like it or not, jack/speaker return terminal was grounded by definition.
Only the English amps used pastic Cliff type jacks, where grounding was optional.

There is a third option -- where the speaker is hardwired to the OPT secondary, which is the case in a fair number of amps, particularly budget ones. I worked on one of these recently where one speaker wire was simply spliced to one of the OPT secondary leads. The other secondary lead and speaker wire were soldered to a chassis grounding lug right next to the OPT.

[R.G.]

How about this Admiral console el84 amp i picked up today? Negative of output transformer goes through cathode resistor to ground.

As with all grounding issues, you have to ask "where is the current going?". In this one the current runs through the loop from OT to speaker and back to the OT. As long as only one terminal of that current loop is attached to anything else, no current flows, so it's just referencing the output to the positive bias voltage for the cathodes of the output tubes.

By the way, that looks very similar to the output stage of the Magnavox hifi console I salvaged way back when. I still remember that PT, OT and output stage as producing the best overdriven amp signal I've ever heard.

There is a third option -- where the speaker is hardwired to the OPT secondary, which is the case in a fair number of amps, particularly budget ones. I worked on one of these recently where one speaker wire was simply spliced to one of the OPT secondary leads. The other secondary lead and speaker wire were soldered to a chassis grounding lug right next to the OPT.

Again, it's "where does the current go" and if only one point in a current loop is attached somewhere, no current can flow through the resulting "monode" connection.

[Gingertube]

In that console schemtic moz posted it is wired that way to introduce cathode feedback.
Cheers,
Ian

[loudthud]

Double yah tee eff colon left parenthesis.

If you think one side of the OT secondary doesn't need to be grounded, try touching one side of the speaker wire of an ungrounded secondary while the amp is playing. If you have a weak heart, connect a meter set to AC Volts, ground the other lead. It will shock the crap right out of you because there is interwinding capacitance between part of the primary (usually right at the plate lead) and the secondary. Besides the shock hazard, the amp may oscillate. Same reason you need a real or virtual center tap on heater windings. It doesn't take much capacitance to deliver significant current when the Voltage is very high.

[R.G.]

In that console schemtic moz posted it is wired that way to introduce cathode feedback

Just goes to show I didn't look closely enough at the schemo. I didn't completely follow that it's connected to both the cathode resistor/cap and then in series with the cathode. So yeah, you're right, it's feedback to the cathodes.

In my favor, at least I got the principle right. The secondary and speaker loop is connected at two places, therefore there is an opportunity to get the get the current to flow through a loop that affects things. In this case, the cathode current flows through the speaker winding and the voltage of its passing through affects the cathode to gate voltage, so yes, it's feedback.

That is, I got it just right enough to be completely wrong.

[Rhodesplyr]

because there is interwinding capacitance between part of the primary (usually right at the plate lead) and the secondary. Besides the shock hazard, the amp may oscillate.

This is what I remember having heard before. I'd just forgotten the specific cause.