Quote:
Originally posted by kevin gilmore
Quote:
| If an amp design requires a regulated power supply to stop hum and PSRR problems, then that amp needs to be redesigned (paying better attention to trace widths, placement and routing more than anything else as it is difficult to make an amplifier so poor with modern op-amps). |
I'm sorry to say jefferyj is lacking in experience. It has nothing
to do with hum problems. It has everything to do with being
a low overall open loop gain and full dc coupling that requires
a power supply to be dual tracking to keep the dc offset at
the output as close to zero as possible without the use of
the servo.
|
Wow, Kevin. I didn't think you'd be one of those types so common at diyAudio where pride/ego compels you to be rude when you disagree with someone, but I guess you are...
Claiming that I am "lacking in experience" and "un-informed" only paints you as, well, bratty. Furthermore, you didn't even provide any factual evidence to back up your assertions that I am even wrong, much less lacking in experience and un-informed.
The actual truth of the matter here is that the closeness of tracking between the two rails has little to do with the PSRR of the amplifier itself. A properly designed gain stage will use current sources, not resistors, to set the collector/drain currents,
automatically providing a substantial amount of rejection to variations in supply voltage! A current regulator provides immunity to voltage changes just like a voltage regulator provides immunity to current changes. It's really that simple.
Next, the DC offset of the usual differential pair input is primarily determined by the
mismatch in beta between the two transistors, secondarily by mismatch between the Vbe's. Differences in rail voltages that don't result in asymmetrical clipping will have no effect on the DC offset! (I am assuming the use of bipolar transistors here both for simplicity reasons, and because bipolar always perform better than FETs if high input impedance isn't necessary. Always.)
Where is the fault in my reasoning? If you can illustrate it without resorting to name-calling/rudeness, I will gladly accept my error. Otherwise you are just being bratty.
Quote:
--------------------------------------------------------------------------------
Originally posted by Ben_Tech
...
The question is. . . would it 'sound' as good as the Gilmore designed supply.
--------------------------------------------------------------------------------
---------------------------------------
said jeffreyj in response
You are kidding, right?
---------------------------------------
jefferyj is once again un-informed. These other power supplys
will not sound as good, they will not be as stiff or low noise.
Whether or not you can hear the difference is another matter
but the difference certainly can be measured. |
How stiff is stiff enough for so wimpy a load as a headphone amplifier? 100 milliohms? 10 mlliohms? And how low of noise do you require for you amp design to sound good? I maintain my assertion that a properly designed amplifier
will not be at the mercy of it's power supply! If you amp design
is at the mercy of the power supply's quality, and you didn't put the power supply on the same board to ensure some level of control over its implementation, then you pretty much
relinquished the perfomance of your amplifier to somebody else.
Finally, I can't help but point out the irony of your comment that "these other power supplys [sic] will not sound as good..." but that one might very well not be able to hear the difference. Of course, you are implying that the builder of your amp must have tin ears, though, presumably, anyone participating in this forum must have some appreciation of fidelity else
why would they bother???
What amazes me about all of this, though, is that designing an amplifer to reject power supply PARD* is so much simpler to do than designing the power supply to exhibit vanishingly low levels of such.
* Periodic And Random Deviations.
