Transformer isolated input versus Capacitor isolated?

Can anyone discuss the relative advantages and disadvantages of transformer isolated inputs versus capacitor isolated?



My guess is the capacitor circuit is more linear (assuming high quality caps), while the transformer circuit has higher common mode rejection. Sound right?

Thanks,
Thomas

Great transformers are unrealistically expensive. Most people use them for the sound (distortion) anyway. I aways try to stay away from transformers in my designs and I get good sound and performance out of capacitors, just use good ones which are way less expensive than trannies.
http://www.jensen-transformers.com/an/ingenaes.pdf
If you want to improve CMRR.

Analag

Yeah, that's sort my take on it as well. I build monitor equipment and I'm really not interested in my input stage having a "sound".

The Whitlock design is pretty cool. But the patent licensing kind of sucks. I guess I could use the THAT-1200 but I like to have some gain available right at the input. So I actually use an instrumentation amplifier rather than the diff-amp I posted.

What about this arrangement?



It simply uses a traditional bootstrapping technique on both inputs. Seems a lot more straightforward than the Whitlock design.

Thomas

[quote author="barefoot"]What about this arrangement?
[/quote]
what is it???
looks like Zobel network at inputs, What kind of oscillations
are you taming ?
Bootstrap is feedback from output of amp to input, not between inputs.

BTW bootstrapping is unacceptable in low noise circuits like
preamps, because it have strange noise figure.
It is dual to paralel negative feedback, which I discussed in this thread:
The lab: Bizarre input Z behavior author: pstamler ¨


xvlk

I added some component numbers to the schematic to make it easier to discuss (refresh).

As long as R3 and R4 are much smaller than R1 and R2 (say 1/10) they form a stiff divider and the op amp output will set A approximately equal to the positive input .... ergo bootstrapping. No?

I'll read that thread.

I guess I don't really understand your point about "parallel negative feedback". Maybe you can describe how it would affect this circuit specifically?

Thanks,
Thomas


(edit)

I'm not 100% sure if this is what you're saying, but now I can see that just as A is modulated by the R3/R4 divider, the R1/R2 divider will conversely modulate the negative input terminal. The effect will be slight since R1, R2 >> R3, R4 but I imagine it is still a bad thing with respect to distortion.

Ok, I think I've got it. I just buffered the negative inputs from the bootstrapping capacitors.



I think this has a couple of significant advantages over the Whitlock circuit. 1. Since the bootstrapping is tied to the negative input, it is independent of the instrumentation amplifier's gain. So the gain can be variable in the normal fashion. In order to achieve gains above unity with the Whitlock circuit you would need to add a divider in front of the bootstrapping buffer. Possible, but that would probably mean a dual ganged pot, one for gain and one for bootstrapping, and the bootstrapping precision would likely suffer. 2. In my circuit since each input op amp has its own bootstrapping, the bootstrapping works in both common mode and differential mode, unlike the Whitlock circuit which only provides high input impedance in common mode. This allows the input coupling capacitors to be quite small and still yield very low frequency response.

How's it look? See and issues?

Thanks,
Thomas

Why bootstrap at all? Use TL072 amps and 10Meg input resistors. Source resistance effects vanish.

For an in-studio monitor, most any diff-input will do. A chippy will be cheaper and less colored than a core.

For a monitor that lives on the far end of a nasty cable, you can't beat a transformer for subduing volts of ground-crap and surviving many-volts across a ground fault.

[quote author="PRR"]Why bootstrap at all? Use TL072 amps and 10Meg input resistors. Source resistance effects vanish.[/quote]
Any ill effects from having the input so isolated from ground... noise, input offset, etc.?

For an in-studio monitor, most any diff-input will do.

Click to expand...

Maybe true, but I'm really aiming for the best of all worlds. And after all a power amplifier does have 30dB or so of gain. I don't see why we shouldn't pay the same considerations we would with any high linearity preamp.