Weird OT issue ?

[catalin gramada]

Hello,
I just noticed using one Hammond OT and trying to apply some nfb. Keeping common secondary lead to whatever ground point or keeping isolate didn't do any diference. I checked DC continuity in isolation condition and shows secondaries are isolated from ground, the output jack's are suspended from chassis, even the iron is physical dc isolate from its support. I injected 9db feedback in my amp and from an accident I disconnected the common ground lead from the ground. I didn't noticed any diference . I repeated more carefully, swapping signal from 100 to 10000 hz, no diferences,the feedback loop is pretty stable with, or without common lead grounded, it looks like OT was grounded by default, despite the fact it is completely isolated. In fact it not exist a physical nfb return.Now I wonder if my transformer could be as well AC coupled with my steel chassis to not see any diference in terms of grounding. It is normal , please? Thanks

[catalin gramada]

And the circuit:



How nfb loop can be functional without nfb return to common OT tap, please? Thanks

[Chuck H]

It looks to me like the global FB may be getting cancelled by the anti phase of the triodes at the shared cathode because both are elevated by a 47R resistor. What happens if lift the global FB loop? Any change? My thinking is that you may not be seeing that you still have a FB loop without common ground, but rather that you DON'T have a FB loop even with a common ground.

[catalin gramada]

Hey, thanks for you input.
The feedback is applied only over 47ohm resistor in cathode of first stage, the 100u over shared 820ohm resistor drain ac voltage to the ground, the 47 ohm over non inverted stage is there just to keep the bias as same and is not compulsory, so nothing is cancelled. The amount of feedback works exactly as should be. Lifting the nfb wire put the voltage gain of the stage with 9 db more.
I still not see it the issue in circuit, the global feedback works very well, except the fact lifting nfb return didn't do any diference. The OT is completely isolated by any grounding point. I don't get it....

[catalin gramada]

This is a pic at onset clipping with 9...or 10 db nfb applied ( not sure ...did a lot of changes) with nfb return lifted. It looks pretty decent, nothing to worry, ..but how the heck it works just with one wire...?



and with nfb lifted



Both pictured without any nfb return, without any common secondary tap physical tied to any ground. DC is infinite

[Chuck H]

Are you sure the load is floating also? Because that could provide a return path to 0V if it's weren't.

[catalin gramada]

Yes, I removed the output jack's from chassis to be sure and the power resistors are away on my table. It's woo-doo I tell Ya.
What it remains to do is to remove the OT bolts and to move it away to see if the loop is not closed by induced current in chassis - if it have any sense...

[catalin gramada]

Oh, wait a sec.... Can be the loop closed through the probes via wall outlet and far away through earth protection of the amp ? Should be a long return, but still...?

[Chuck H]

I'm interested. I'd be surprised if the OT EMF is coupling/damping at the chassis so completely that the NFB level doesn't change significantly from an actual common connection to 0V. Not much to work with. I'm out of ideas.

[Chuck H]

Oh, wait a sec.... Can be the loop closed through the probes via wall outlet and far away through earth protection of the amp ? Should be a long return, but still...?

I thought about this, but it seemed unlikely. The probes should have considerable impedance.

[catalin gramada]

....no, no...Bingo! ...I just checked ...between probes and amp chassis get 0.25 ohm resistance. It is a healthy path....Thank You.

The alligator clip which was connected to the common tap get contact direct with the chassis through 0.25 ohm resistance of earthing wiring.

[Chuck H]

....no, no...Bingo! ...I just checked ...between probes and amp chassis get 0.25 ohm resistance. It is a healthy path....Thank You.

The alligator clip which was connected to the common tap get contact direct with the chassis through 0.25 ohm resistance of earthing wiring.

Oh.!. You have the scope plugged in and obviously grounded. So because you're measuring AC at the load you have the scope - connected to one end of the load and the + connected to the other. Makes sense now. Glad you found it before unbolting the transformer

[g1]

Are you sure the load is floating also? Because that could provide a return path to 0V if it's weren't.

Yes, I removed the output jack's from chassis to be sure and the power resistors are away on my table.

....no, no...Bingo! ...I just checked ...between probes and amp chassis get 0.25 ohm resistance. It is a healthy path....Thank You.

The alligator clip which was connected to the common tap get contact direct with the chassis through 0.25 ohm resistance of earthing wiring.

'Floating' load means NO other earthed equipment (like scope) be connected. This is also very problematic when testing solid-state amps that use current feedback.
Best practice is to leave scope ground connected to amp chassis, or measure at load with a hand held multi-meter that has no earth connection.

[uneumann]

'Floating' load means NO other earthed equipment (like scope) be connected. This is also very problematic when testing solid-state amps that use current feedback.
Best practice is to leave scope ground connected to amp chassis, or measure at load with a hand held multi-meter that has no earth connection.

True - problematic, but many dual-channel scopes allow A-B display so they provide a solution. The A and B channel probes can be connected to the two floating outputs, with their ground clips both clipped to chassis. This safely avoids the ground problem described above, and it shows exactly what the output (spkr) signal looks like.

[J M Fahey]

The feedback is applied only over 47ohm resistor in cathode of first stage, the 100u over shared 820ohm resistor drain ac voltage to the ground, the 47 ohm over non inverted stage is there just to keep the bias as same and is not compulsory, so nothing is cancelled. The amount of feedback works exactly as should be. Lifting the nfb wire put the voltage gain of the stage with 9 db more.
I still not see it the issue in circuit, the global feedback works very well, except the fact lifting nfb return didn't do any diference. The OT is completely isolated by any grounding point. I don't get it....

Why do you say so?
The "common" end of the OT is connected DIRECTLY to ground, per your schematic .
And the 8 ohm tap is connected to first PI cathode through a 2k2 resistor and a series 100uF capacitor.

NFB loop, at audio frequencies, is 2k2/47 ohm