|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
73.225.74.165
In Reply to: RE: how calculated? posted by RayP on June 29, 2015 at 08:26:20
I looked at the data sheet (Mullard was handy) which has curves for g2 current and for plate current, both for 250v g2. Starting from your operating point of 320v plate to cathode, and assuming 35mA plate current to stay within the maximum dissipation, it looks like you can get 70mA at 50v as one end of the load line; that occurs with -4v of g1 bias. then at 4v g1 and 50v on the plate, the g2 current would be 21-22mA.
This is just the instantaneous peak current; the data sheet also gives quiescent current which is probably close to average current for music. If there is a big enough capacitor at g2, it could supply the peaks, but if it's a regulator at g2 then it has to pass the necessary peak current, at least momentarily.
Follow Ups:
"if it's a regulator at g2 then it has to pass the necessary peak current, at least momentarily."
The SS device being discussed below can drive more capacitance than gas tubes. That should relieve the regulator itself of some of the peak current.
--------------------------
Buy Chinese. Bury freedom.
Just the point I was hinting at!
This begs an interesting question - almost everybody (including me) says that regulators sound better. A regulator is equivalent to a very large capacitor in that both produce a very small variation in voltage when the current fluctuation has a shorter time scale than the time constant which is RC for a resistor-capacitor filter. Nobody has any idea what that minimum time constant is for an RC filter to produce the same audible result as a regulator. Experiments with huge capacitor banks were popular a few decades ago, but I don't recall any consensus agreed values for the effect.
"Nobody has any idea what that minimum time constant is for an RC filter to produce the same audible result as a regulator."
It's not possible to answer this with a single number. The necessary minimum time constant varies as a function of the load on the supply. The lower the load Z becomes in terms of its AC component, the shorter the time constant of the supply must be.
Personally, I think the sonic effectiveness of supply voltage can be easily observed and measured. If the amplifier at full output can't wiggle the supply to any significant extent, that aspect of the system is good.
--------------------------
Buy Chinese. Bury freedom.
"If the amplifier at full output can't wiggle the supply to any significant extent" - over what time scale? The biggest likely effect will be envelope modulation because the loud passages demand a bit more average current than the quiet passages. So, does the supply need a time constant equal to Beethoven's 9th? Will a minute do? A second?
It's at least a 2-dimensional problem - the depth of the effect (power supply ESR relative to load ESR) plus the speed or slowness with which the wiggling takes place. Jeff has (somewhat forcefully) advocated for a low power supply ESR; I am interested in the time dimension. A regulator addresses both.
" over what time scale? The biggest likely effect will be envelope modulation because the loud passages demand a bit more average current than the quiet passages. "
In order to have a meaningful discussion about this, we would first have to agree on the difference between A) power supply sag that occurs at less than 1Hz, and B) power supply output Z over the musical frequencies. I can only say at this point that I don't equate long-time-constant power supply sag with envelope modulation.
--------------------------
Buy Chinese. Bury freedom.
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: