Amp Gain Design Philosophy

[TheGimp]

Back in the day 1Vrms into 47K was a standard. Not necessarily THE STANDARD, but a standard.

From what I've been reading this has been upped to 2Vrms (what Z I don't know).

My current design is underpowered when played from a Computer CD player (yea, not exactly a standard but it is convenient in the shop).

I'll try my tape deck when I can remove it from upstairs without my wife noticing. (Black Friday might be a good opportunity, but expensive.) Till then, what is the recommended input for clipping to be compatible with most modern audio equipment without using a line amp.

Should I plan on building a line amp to take care of variations in output of various devices I'll be working with (phono will need an amp with equalization, tape, CD, ...)?

[DougL]

2 V rms is the redbook standard, and potentially allows "passive"preamps. As far as I can tell, output Z on CD was not specified. My Dinaco ST70 had a 250K input Z because of the 7199 pentode input.

I would think of it as a compromise. Too much gain is more noise and distortion. Too little gain makes for frustrating listening.

For my system, I am designing my amps to be about 1 Vrms for full output with as high an input resistor as I can manage without making the Miller capacitance significant at the input.With tubes, that's usually 47K or 100K. For Transistor amps, as little as 20K is normal, with lower not unheard of.

The really tough one is phono, because cartridge outputs vary so much. Either design the phono section for the cartridge you are using, or build in some "extra" gain to accommodate the least sensitive cartridge you might use. The alternitive is to use a line amp to supliment the gain of the various low output devices.

HTH

Doug

[revintage]

Hey Gimp,

This is how I would like thing to work as I run phono:

There is a semipro line standard of -10dBV/10k(just above 300mV). This is adequate for a typical phonoamp, tuner etc. albeit a higher Zin sometime could eventually be wished for. So a active preamp with 10-20dB gain is needed.

With a CD at 2V out we have a slight problem. It is not the best idea to damp the CD-signal resistive as dynamics seems to get lost. A 4:1 (-12dB) transformer could be the solution if we dump the resistive damping. Another solution is to bypass the gainstage when going CD.

[Mosquito]

Hi, I was wondering about pretty the same thing, but from another POV.
The standard for CD player outputs is said to be 2Vrms, are real CD players in the midclass Sony-Marantz -Denon range doing so?
And, what is the minimum gain You need for an integrated tube amp featuring a single gain splitter stage, if the output tubes need more than what a 6BQ5 type tube usually needs, say 25-30 Vrms for full swing...?
I mean an amplifier with a CDP directly attached to the imput and, say, a 100K pot as volume contrrol..

[kevinkr]

I design my sources (mostly DIY now) including phono pre-amplifiers for 1.5Vrms - 3Vrms output voltage and use a line stage with very low gain into a power amplifier that will typically have a sensitivity of 1Vrms for rated output power.

Putting all of the gain at the front end of the system is good for SNR and running higher signal levels everywhere else makes noise contributions elsewhere less of an issue.

[Eli Duttman]

1st, the IHF "standard" for line level loads is a NASTY 10 KOhms in parallel with 1000 pF. Good design deals with that horror.

IMO, power amps should be sensitive enough to reach full power when driven by a CDP/"unity" gain control center setup. The choice between an active or a passive control center is mostly a matter of personal preference.

[kevinkr]

Originally Posted by Eli Duttman

1st, the IHF "standard" for line level loads is a NASTY 10 KOhms in parallel with 1000 pF. Good design deals with that horror.

IMO, power amps should be sensitive enough to reach full power when driven by a CDP/"unity" gain control center setup. The choice between an active or a passive control center is mostly a matter of personal preference.

First point: Driving 10K//1KpF not much of a problem for most solid state sources, but a big problem for older designs not based on op-amps or low impedance discrete transistor circuitry. (i.e tubes) Need to pay attention to phase margin if feedback employed...

Second point: Absolutely in agreement here. Note that current practices in commercial HTIB products and car stereo is for sources and head units to put out much higher voltages than IHF standards call for as then less gain is required in the chain prior to the PA - this in practice usually results in less susceptibility to noise pick up with poorly shielded or long interconnects and lower PA amp gain which subjectively (and practically) results in a quieter system.

IMO in DIY systems where compatibility is not an issue it really makes sense to run the sources as hot as possible, and use as little gain as required subsequently to realize the maximum required amplifier output power. Obviously something like 6 - 10dB of additional gain beyond this is a good idea to make sure that the amplifier can be driven to full output on signals that are 6 - 10dB lower than the expected nominal.. (Like a CD recorded at maximum levels of -10dBFS for example.)

[Eli Duttman]

Driving 10K//1KpF not much of a problem for most solid state sources, but a big problem for older designs not based on op-amps or low impedance discrete transistor circuitry. (i.e tubes) Need to pay attention to phase margin if feedback employed.

No small part of my fondness for the ZVN0545A and IRFBC20 MOSFETs comes from being able to drive loads, via a source follower, without resorting to loop NFB. Setting up DC coupled source followers with quality enhancement mode MOSFETs is very easy.

Those who are adamant about no "sand" in the signal path can use 5687, ECC99, 6Н30П ... cathode followers.

The 12B4 is about the only tube I'll use with a line load connected to the plate.