Dumb (?) question about the balanced dynalo

[00940]

It seems that anyone building a balanced "dynamid" is building two complete amplifiers by channel. Wouldn't it be more cost effective to wire a bridge-balanced dynalo as a fully differential amplifier ?

It's working in the simulator but there must be reasons it isn't done ?

 

[holland]

Because each channel has a positive and negative input (and a referenced ground point), meaning you need 4 for a stereo setup. That means 2 boards.

For mono, yes, what you depicted is fine.

 

[00940]

I know that Holland.

The configuration I suggested is using only the + and - inputs of one balanced bridge amp (with + and - outputs) for respectively + and - phases, with two different feedback path, going to - and + output (think of it as two closely linked inverting amplifiers). So with one board, you thus amplify both phases of the signal.

 

[amb]

If you look at the original Gilmore Dynamic amp project page in the headwize library, you'll find something in there that does essentially that (outputting both phases from the same amp to make a balanced amp). However, that solution only has one global feedback path for the non-inverting side. The end result is that there is nothing on the inverting side to correct for its output stage nonlinearities.

The design of the amp is such that it is not easily adaptable to having two feedback paths without major changes. What's shown in the OP above having two outputs and two feedback paths is actually not a simple amp block. Thus, a 4-channel dynalo (aka "dynamid") is thus a better solution.

I think a OPA1632 fully-differential opamp driving a pair of JISBOS buffers per channel might make a nice balanced amp.

 

[00940]

Quote:

Originally Posted by amb /img/forum/go_quote.gif If you look at the original Gilmore Dynamic amp project page in the headwize library, you'll find something in there that does essentially that (outputting both phases from the same amp to make a balanced amp). However, that solution only has one global feedback path for the non-inverting side. The end result is that there is nothing on the inverting side to correct for its output stage nonlinearities.

Guess why I asked my question ?

The "feedback for one phase only" thing wasn't attracting me much. I was originally playing around the dynalo in Ltspice to see if raising the voltage of the front end could increase its voltage swing (answer seems to be : not that much btw).

Quote:

The design of the amp is such that it is not easily adaptable to having two feedback paths without major changes. What's shown in the OP above having two outputs and two feedback paths is actually not a simple amp block. Thus, a 4-channel dynalo (aka "dynamid") is thus a better solution. I think a OPA1632 fully-differential opamp driving a pair of JISBOS buffers per channel might make a nice balanced amp.

I've figured

The more I look at it though, the more I think that the "BJTed" UGS found here :http://www.diyaudio.com/forums/showt...threadid=86300 could make a very nice balanced headphone amp, with parallel transistors at its output.