Speaker Output Grounding Questions

[SerpentRuss]

If you're not going to have global negative feedback, the secondary of the OT does not need a ground reference, right? My speaker output jacks are not automatically grounded via the chassis connection.

Are there any other reasons to give the speaker a ground reference? If you're not using Global Negative Feedback is it better to let the OT secondary float? If you do ground the speaker side of the OT, where do you normally pick up that ground reference?

 

[chas.wahl]

Here's Valve Wizard Merlin Blencowe's take on it:

This is from his "Grounding" article, in which he advocates the unified bus from stem to stern of the circuit. Note that when trying to group grounds to approximate a star, the output is grounded to the negative pole of the filter cap feeding the output stage's input (on the "low" side of the first dropping resistor, along with the output tube's cathode ground.

If you prefer, like many, to have separate power and preamp chassis ground points, then I suppose the output ground would be grouped with that for the other power components.

Also note that the above example doesn't include NFB.

Blencowe's comments in the text of the article state: "A separate wire (which does not need to be heavy gauge) should then run from the negative connection of the speaker jack back to an appropriate star. If global feedback is not used then this speaker ground wire should be returned to the power-amp star. If global feedback is used then the speaker ground should be returned to the local star of whichever stage the feedback happens to be applied to, which is usually the phase inverter (e.g., fig. 15.14)."

Fig. 15.14 doesn't actually show the output, but a preamp with NFB, with a note showing where feedback is applied; take a look at the article itself.

 

[Phrygian77]

If the secondary if floating, there can be DC difference between the secondary and ground. There isn't any reason not to ground reference it. There's never going to be any significant current from the secondary to ground, assuming you don't actually allow the secondary to return current through the chassis.

 

[King Fan]

You can't beat Merlin, and I refer to his stuff all the time. Still, in the small amps I've built (biggest being a PR or 5G9) I don't actually seem to find an issue with grounding the speaker jack to the chassis -- that is, not using isolating washers on my Switchcraft jacks. They're all darn quiet. Even R.G. Keen, who's quite the purist in his grounding article, admits it may not matter in practice. His experience matches mine:

I tried the speaker jack connected/not connected to the chassis in an amp I was re-grounding, and for the life of me could not hear any difference. I don't know whether this was because I already had everything else off the chassis or whether it truly didn't matter at all. I guess I would take a somewhat rough and ready approach - if it can be isolated and star grounded with a reasonable amount of wiring, I'd do that. If not, I'd leave it to the end and see if the results were acceptable from a listening test.

 

[SerpentRuss]

Ok, good to know. So it seems Merlin thinks you should reference the speaker output to ground regardless. I have the darn book, I must have missed that. Some chapters I've read a dozen times, but not the grounding chapter.

I have at least two tiny amps breaking this rule right now but in the grand scheme of things it probably doesn't matter from a safety or sonic perspective. Those builds and my current build use the Amphenol switching jacks which are isolated by design.

 

[King Fan]

I think you're right. To clarify my Keen quote, I think he meant it should always be grounded, too, but he couldn't hear the difference between an explicit ground off the chassis and the standard trad Switchcraft ground via the case.

 

[2L man]

There flow very little current on NFB circuit BUT if loudspeaker jack is not insulated, its current path comes thru chassis to input jack to "0V/ground wire" and there could come something? When also instrument cable neutral connects there so why not do it better?

On few build NFB I have used a shielded cable and used the shield as a reference soldering its both ends, other end to output jack neutral and other end to "0V/ground" bus.

 

[andrewRneumann]

Would there be safety implications not having the secondary grounded? Imagine a fault voltage reaching the secondary from the primary. I’m not saying either way, but I would do what everyone does and ground that secondary.

 

[King Fan]

From another thread, I saw this by @Ten Over , whose knowledge is 10:1 over mine…

"Common" is frequently the same as "ground" in the circuits that we use, but not always. Some amplifiers do not connect the speaker output to circuit ground. In this case "common" is not the same as "ground", but the hookup with respect to the speaker jack is the same.

 

[2L man]

Common is very good term nowadays because Ground, as it was used when guitar tube amps were designed, should not exixts anymore in current builds. It was long gone on that era current electronics when I went to school to study Radio/TV/electronics 1978 here in Finland and I believe it has been all over the world very long time? Ground was teached because there were old radios, TVs and amps which we were teached to repair.

I've been working with electronics, which mostly has been quite modern, but there has been tubes as well until early 90s and CRT projectors did stay until about 2010. However there has never flown any current in chassis/frames so it was quite a surprice when I got inteest building tube amps that the use of chassis for operative current did and still exists so often.

 

[Jewellworks]

reviving this thread a bit...
ive been reading Merlin's Grounding page, and what about when using a universal OT? with several choices for taps and multiple outputs?
for instance, i have a Hammond 125D. it has several taps and depending on the load and impedance you want, you use different pairs. ive got a pair for 4, 8 and 16 ohm jacks in the same amp. if i tied them to a "common" i just upset the "pairs" on the 2ndary.
other transformers just give you 2 outputs and common is common, so i get that, but it seems like i should just float the outputs when using a universal OT, as there is no "common"

 

[andrewRneumann]

Assuming no NFB, you could ground any 1 of those taps and they will all work just fine. I assume most people would ground tap “1” even though it’s not labeled common.

 

[Jewellworks]

i should have been more clear...
ive got 3 DIFFERENT pairs for 4, 8 and 16 ohms, all in the same amp. there IS NO "common".
i think im answering my own question here... they'd have to float, and stay off the Chassis.

 

[2L man]

Separate transformer coils transfer AC power despite what potential they are. Direct heater tube rectifier filament is good example where filament elevate to what HV DC comes but there is still 5VAC.

So you can connect all three coil 0s together. Or to series like they are in common 4-8-16 setup and 0 to Common if you wan't. I don't know is there any harm if they are left "floating" but usually on schematics they are not even no NFB is needed.

There are OTs which secondarys are connected sometimes strange ways to different loadspesker impedances to use copper wire and space around iron core more efficiently. And to improve OT sometimes or usually coils are also interleaved in sections which possibly can be seen when looking coil stacks or bobbin "outlets which have more wire cinnections than there are actual outputs, when end bells are removed.

I just bought a PP HiFi OT which secondary seem to have three coils which are parallelled soldering them together on lugs just outside of plastic bobbin. It also has UL inputs on primary but also two additional lugs which are closer the B+1 input. That obviously is because if interveaving?

 

[Bitsleftover]

i should have been more clear...
ive got 3 DIFFERENT pairs for 4, 8 and 16 ohms, all in the same amp. there IS NO "common".
i think im answering my own question here... they'd have to float, and stay off the Chassis.

I tried using all 3 taps on that transformer exactly as you describe and I ended up with a potential difference between the speaker jack body and the amp chassis. Not huge, but enough to give a definite tingle if touched it with bare hands.
I ended up removing the 4 ohm and tying the wire now common to 8 and 16 to ground. It made 8 and 16 useable but I was disappointed I couldn’t use all 3 taps.
I’ve since read threads on here by other users puzzling over how to use all 3 to no avail.

 

[andrewRneumann]

i should have been more clear...
ive got 3 DIFFERENT pairs for 4, 8 and 16 ohms, all in the same amp. there IS NO "common".
i think im answering my own question here... they'd have to float, and stay off the Chassis.

Which taps do you have in use? I still don’t see the harm in using tap (1) to ground the secondary. Important to use isolated speaker jacks though.

 

[Jewellworks]

@andrewRneumann : that simple chart doesnt tell the whole story with this OT. see attached. different taps give you different loads, and impedance.
im using pairs :
3&5 @ 4ohms for a load of 8.8k
5&2 @ 8ohms for a load of 8.2k
6&2 @ 16 ohms for a load of 8k (chart says 15 ohms, but if i use it, ill be a 16ohm spkr, and slightly over 8k)

there IS NO common. so they'd all have to float, OR pick #5 and lose my 16ohm jack, OR pick #2 and lose my 4ohm jack. cant get all 3.
and they are lifted off the chassis.

 

[SerpentRuss]

@andrewRneumann : that simple chart doesnt tell the whole story with this OT. see attached. different taps give you different loads, and impedance.
im using pairs :
3&5 @ 4ohms for a load of 8.8k
5&2 @ 8ohms for a load of 8.2k
6&2 @ 16 ohms for a load of 8k (chart says 15 ohms, but if i use it, ill be a 16ohm spkr, and slightly over 8k)

there IS NO common. so they'd all have to float, OR pick #5 and lose my 16ohm jack, OR pick #2 and lose my 4ohm jack. cant get all 3.
and they are lifted off the chassis.

I would think ground whichever low number terminal you want for the matching negative feedback output, but let all the jacks float. Or, you could use one jack and switches to select the pair that goes to a single output jack and ground the negative of the jack. That's the most elegant solution.

I'm using two SPDT switches to choose between 3 impedances on my current build. Since I have one common for all three, I only need to switch three wires, you could use two DPDT and switch both wires of the pair.

 

[andrewRneumann]

@andrewRneumann : that simple chart doesnt tell the whole story with this OT. see attached. different taps give you different loads, and impedance.
im using pairs :
3&5 @ 4ohms for a load of 8.8k
5&2 @ 8ohms for a load of 8.2k
6&2 @ 16 ohms for a load of 8k (chart says 15 ohms, but if i use it, ill be a 16ohm spkr, and slightly over 8k)

there IS NO common. so they'd all have to float, OR pick #5 and lose my 16ohm jack, OR pick #2 and lose my 4ohm jack. cant get all 3.
and they are lifted off the chassis.

Was this determined experimentally or just theoretically?

I just don't see how you lose signal across any of those pairs just by connecting a single tap (any one of them) to ground. You are just tethering the whole secondary to 0VDC instead of letting it float wherever it wants to be. The appropriate AC signals still appear across all the taps. Floating to me means no reference to ground in the entire circuit. If you connect any one of those taps to ground, they all become tethered to 0VDC. Right?

 

[King Fan]

Good stuff, Andrew. I wonder if we tend to confuse 'grounding out' with 'linking to a reference voltage"...