simple line buffer

[fscarpa58]

Hi guys,

Look at the performance difference between
classic CF and its floating supply counterpart
in term of PSRR.

is it not dramatic?

comments are welcome

Federico

[dimitri]

Ciao, Federico, the idea is nice. I don’t see any necessity to connect both earths together – so you will get galvanic decoupling between the source (e.g. cd player) and the load (amp), not counting 1M resistor. That is good. I have several questions:
What if the output caps aren’t be equal?
As I presume you will connect the chassis to the input earth. How about simulation with two parasitic capacitance from both supply rail to that point?

[dimitri]

>I don’t see any necessity to connect both earths together – so
>you will get galvanic decoupling between the source (e.g.
>cd player) and the load (amp), not counting 1M resistor.

It wouldn’t work. Both earths should be connected! Sorry

[dimitri]

Federico, I’ve checked your simulations with ucap and sorry to discover that your PSRR data is doubtful. I got -45dB for your circuit values.

It is a bridge: left upper branch – internal triode resistance Ri; left bottom branch – 13.8kOhm. For the bridge null the capacitors must be in proportion Ri/13.8kOhm. Ri for 6DJ8 ~2.7kOhm. For the ucap model for 6DJ8 the max PSRR will be around 0.34u (upper cap value)

[fscarpa58]

Hi Dimitri, some interesting points here.

As you have underlined it is a bridge.
But, in your calculation you have probably forgot
the bias resistance whose value (300 Ohm)
has to be multiplied (1+mu) times
and added to the internal triode resistance.

In my example we have ri= 3254 ohm
mu= 34.285 that is
[ri + (1+mu)*300]= 13840 ohm.

Really, the exact value of the above parameters
are not so critical since, as you have suspected,
differences in the output caps will destroy the balance.

The Rk resistor (13.8k) has to be a pot. So,
if PSSR minimization is the target, one
can adjust its value as a function of the working
point ( on which ri and mu depends) and of
the caps value.

For instance if I change the value of the upper cap
to .47u ( instead of .5u) the PSSR shifts from –88db
to –61 db and I have to put Rk=13k ohm to restore
the original pssr value. On the other hand if I change
the lower cap to .47u I have to make Rk=14.7k
to balance the bridge.

The small resistance in series to the supply
is for lower PSSR at high freq.

Parasitic capacitance from both supply rails
to earth does not imbalance the bridge if they
are of the order of hundreds picoF.

Thanks for your comments Dimitri.

Bye
Federico