teaser quote: ‘The program material might VERY SELDOM contain peaks up to 20dB in excess of the nominal level; we accept to take the risk these will clip the power amp but we DON’T want them to clip the preamp; the preamplification must have tons of leeway under the most tragic circumstances.’
a gift from Joe Roberts, SP editor: from the sound practices CD, here is the original article.
my take
It is really cool that Diego starts the article with investigating how much gain is really needed in a preamp, taking into account the big dynamics shifts that can be encountered on LPs. What I don’t find that useful is to put the full output of the power amp into the mix. The latter has everything to do with the sensitivity of the speakers used, the size of the room, the preferred maxed-out listening levels and the headroom designed into the power amplification. That is already enough of a puzzle in itself, no need for that to leak into the preamp gain discussion.
I think it is much more useful to define a line level standard—max level—for a system (say, 0.75Vrms, or, when a CD player is involved, 2Vrms) and design for the 10dB peaks of average records to appear on the output of the preamp at that level. It is up to the line amp—and its volume knob—to deal with the loud and the modest-level records and maintain the line level standard, which then refers to maxed-out listening levels. The line amp can be a separate unit, or integrated: in either the preamp (full-function preamp) or the power amplifier (integrated amplifier). But not in both: I envy you if you got an Audio Note (jp) M10 hooked up to an AN.jp Ongaku, but now you got one line amp too many.
Diego’s next point is headroom. It is fascinating that in a device that amplifies microvolts into part of a volt you will have to deal with ten(s of) Volts in the meanwhile. The point where the signal exits a gain stage and enters equalisation, the highs are living at a level that is either 10 times (3180µs / 318 µs) or 23.6 times (75µs / 3.18µs) higher than the lows. And all the gain is there for the lows. Thus it is easy to see that in preamps with passive equalisation, there is a point where this gets critical—the kathode of V3 in this case. I even bet that with full-feedback equalisation, under transient conditions, this headroom reduction point exists. Building a phono amp with op amps on +/-18V rails? Almost no headroom left, is there?
The differences of the standard and SL versions of the preamp show off two principles:
- the values of 13 out of 21 signal components need to be tuned to the type of tubes used (to their µ, gm, rp values);
- selecting component quality is a holistic process, not a per-component choice.
pop quiz
Is Diego running his heaters at 6.6V? That is not good for the life of your tubes. There are 2 diodes under the 7806 regulator, making it regulate to 7.2V. The pass transistor after it drops one diode back to 6.6V, or so it looks. Turns out the BDV65 is a darlington power transistor, dropping 2 diode drops, back to 6V.
bonus tracks
Check out that Audio Note (jp) M10; you can see how the familiar RIAA time constants are directly reflected by the choke values.
practical note: department of corrections, SP15: ‘B+ supply voltage should have been given as +350 V at the B+ symbol on the schematic at the top of page 30.’ It is a bit of a pain that the dot and star symbols are missing in the power supply component listing. It is easy (and safe) enough to guess that the 5W resistors are the ceramic ones. But the X2 surge cap? My best guess is C10, but proceed at your own risk.
Now go and read the article, see you next week.
No comments:
Post a Comment