Cheaper T pad?

[abbey road d enfer]

I usually use this method

OK, each one his own. Do you think it has a definite advantage?

 

[Rob Flinn]

OK, each one his own. Do you think it has a definite advantage?

Not sure. However would I be right in saying it maintains the balancing better than the other method ? Not sure whether that makes much of a difference in most real world situations but ....

 

[abbey road d enfer]

Not sure. However would I be right in saying it maintains the balancing better than the other method ? Not sure whether that makes much of a difference in most real world situations but ....

Think of all the parasitic capacitances that exist in a xfmr, and how they couple with a (generally) grounded primary.
See attached pic.

Green is CM voltage (ref to 0dB input) of xfmr with only a load attached to it, no connection no cable.
Magenta is the same with attached 5m of cable (140pF/m between conductors 100pF between conductors and shield) and 20k load.Cyan is with your rheostat arrangement. Note that the signal is attenuated 6dB, so actually the CM voltage should be increased by 6dB in relative mode.
Red is the potentiometer arrangement.

One can see that the major parameter herre is the inter-winding capacitance. Of course, one may debate about the model I used for the xfmr, but these values are extracted for measurements I have done.
Other xfmrs will obviously result in some differences, but the trend is there.
If the xfmr was driven symmetrically, or the xfmr had an electrostatic shield, that would be another story, but in the specific case of the 312 and many others, that's a good picture.

 

[Rob Flinn]

Think of all the parasitic capacitances that exist in a xfmr, and how they couple with a (generally) grounded primary.
See attached pic.

Green is CM voltage (ref to 0dB input) of xfmr with only a load attached to it, no connection no cable.
Magenta is the same with attached 5m of cable (140pF/m between conductors 100pF between conductors and shield) and 20k load.Cyan is with your rheostat arrangement. Note that the signal is attenuated 6dB, so actually the CM voltage should be increased by 6dB in relative mode.
Red is the potentiometer arrangement.

One can see that the major parameter herre is the inter-winding capacitance. Of course, one may debate about the model I used for the xfmr, but these values are extracted for measurements I have done.
Other xfmrs will obviously result in some differences, but the trend is there.
If the xfmr was driven symmetrically, or the xfmr had an electrostatic shield, that would be another story, but in the specific case of the 312 and many others, that's a good picture.

That's interesting, thanks. My view was that if one was driving into a true balanced input then if one just uses a pot on it's own then it messes with the balancing. This is because it appears that if the pot is anything other than fully turned up it introduces a resistance into only the hot leg of the output, therefore the output is no longer properly balanced. Whereas if one uses the 2x 300R resistors at any position of the pot the added resistance is equal on both hot & cold therefore maintaining the balancing better. I have been mainly using this method on some home brew vari mu limiters & the drop of level due to the attenuator is not an issue because I'm trying to lose gain anyway. But as I said before, how much this benefits a real world situation is another whole debate.

 

[abbey road d enfer]

This is because it appears that if the pot is anything other than fully turned up it introduces a resistance into only the hot leg of the output, therefore the output is no longer properly balanced.

Wrong again. The secondary is floating (except for the parasitics), so there is nothing like a difference of impedance between legs. they both see each other through the same impedance.

 

[moamps]

....Of course, one may debate about the model I used for the xfmr, but these values are extracted for measurements I have done...

Could you please post the equivalent circuit you used in that simulation? Thanks.

 

[abbey road d enfer]

Yes.
I know I haven't pursued with the consequences on CMRR. It would take time I don't have right now. That was just to show that xfmrs should be taken with their numerous imperfections and variations, and the consequences on the circuits they are attached to..

 

[Rob Flinn]

I was wondering what the CMRR graph would would like if a Tpad attenuator was simulated.

 

[sonicmess]

Nice thread ! Need a pair of 600OHM T-Pad for my dual REDD47 preamp.

check white market