Does really need a balanced preamp ?

[popilin]

Designing a phono preamp, in view of the current trend of making everything balanced, given my limited budget, I was forced to think a bit.

Thermal noise can be expressed in the form

e(Noise) = √ (4kTRB) ....(1)

where:
k is the Boltzmann constant
T is the absolute temperature
R is the resistance
B is the bandwidth

For n devices in series

[e(Noise)] ^ 2 = (4kTB) Σ Ri

If all n have the same R

e(Noise)s = √ n e(Noise) ....(2)

For n devices in parallel

[e(Noise)] ^ 2 = (4kTB) [Σ (1/Ri)] ^ (-1)

If all n have the same R

e(Noise)p = (1 / √ n) e(Noise) ....(3)

The stochastic nature of the noise is already covered by Eq. (1), for this reason, the thermal noise is not canceled on a balanced amp.

Conclusion: The intrinsic noise generated on a balanced amplifier is higher than on a SE amplifier by a factor √ 2

Due to I use short cables and I am very careful to shield the fields I'm happy to do my preamp SE !

 

[kevinkr]

I've never drank the balanced phono stage kool-aid and achieve snr results that seem to be relatively unimpeachable - balanced input (unbalanced gain path) makes the most sense when it is free as in the case of an MC step up provided that the wiring to the cartridge is suitable.

Doing it with an active differential input stage exacts a 3dB noise penalty, OTOH with careful design the noise floor can still be dominated by something ahead of the input stage.. Only correlated noise will cancel in a balanced stage and as you have pointed out gaussian noise sources will not cancel differentially although you can sum them. (another discussion)

Practically speaking I've never had problems keeping the hum out of my phono path so I don't see a good reason for the added complexity particularly with tubes.

(Edit for clarity)

 

[nattonrice]

I don't quite see how these eqn say anything about noise in a differential amplifier, how did you arrive at your definitive conclusion from the above?

How is (1) saying anything about a stochastic process?

Edit: Is not the point to "cancel" common mode noise?

 

[popilin]

I've never drank the balanced phono stage kool-aid and achieve snr results that seem to be relatively unimpeachable - balanced input (unbalanced gain path) makes the most sense when it is free as in the case of an MC step up provided that the wiring to the cartridge is suitable.

Doing it with an active differential input stage exacts a 3dB noise penalty, OTOH with careful design the noise floor can still be dominated by something ahead of the input stage.. Only correlated noise will cancel in a balanced stage and as you have pointed out gaussian noise sources will not cancel differentially although you can sum them. (another discussion)

Practically speaking I've never had problems keeping the hum out of my phono path so I don't see a good reason for the added complexity particularly with tubes.

(Edit for clarity)

Hi Kevin

I appreciate that you share your experiences, and I am pleased with this in particular.
Thanks !

I never did fully balanced active stage.

The last SE phono preamp I did for a client, has a gain of about 44 dB, is very quiet, even with speakers of more than 97 dB/W.
It was designed only for MM cartridges, but with the extra 22 dB from line preamp, my client use a MC cartridge without problems.

I must also confess that never use star ground, yet had no hum or buzz problems.

So why spend double the money ?

Not that I'm stingy, I am poor.

 

[popilin]

I don't quite see how these eqn say anything about noise in a differential amplifier, how did you arrive at your definitive conclusion from the above?

If you look at the data sheet of a low noise valve, Req (Equivalent noise resistance) gives the equivalent noise at the valve input, as if it were a resistor.
Then you can calculate the total noise with Eq.(1)
The rest is history, if you want to reduce noise, put two or more parallel devices (valves, JFETs, BJTs ...)
If you put two devices in series, the situation is analogous to a fully balanced active stage.
With valves, the two sections are usually a mirror reflection of the other.

How is (1) saying anything about a stochastic process?

Well, this I know since childhood, somewhere I have a more satisfactory mathematical proof, but this is an audio forum.

If it's any help, you may think that Eq. (1) comes from Thermodynamics and Statistical Mechanics.

You also can read

http://www.google.com.ar/url?sa=t&r...lm8ggY3rts641DLz4XOO1wA&bvm=bv.45960087,d.eWU

Edit: Is not the point to "cancel" common mode noise?

The point is cancel common mode noise, when noise is deterministic nature, e.g. induced hum, buzz, etc.

But I have no problem with external induced noise, and in general I get along with Mr. Maxwell.

 

[Vinylsavor]

Hi!

The last SE phono preamp I did for a client, has a gain of about 44 dB, is very quiet, even with speakers of more than 97 dB/W.

What does the quietness of a phono have to do with the speaker sensitivity?
People keep confusing that. Most people with a well chosen system will have power amps which match the sensitivity of the speakers, so lower power amps with sensitive speakers. Lower power amps typically have less gain, so they would amplifiy any noise from a phonostage less than higher powered amps

If the phonostage is very quiet, I would not be concerned about 3dB more noise when going to balanced

I must also confess that never use star ground, yet had no hum or buzz problems.

Low hum does not require star ground. Star grounding done wrong can cause many problems.

So why spend double the money ?

I mostly build SE preamps too. But I did also build fully differential phono stages. There can be certain advantages which are worth the effort. The 3dB additional noise have never been an issue for me

Best regards

Thomas

 

[popilin]

Hi Thomas

What does the quietness of a phono have to do with the speaker sensitivity?

Strictly nothing, but unless you have a spectrum analyzer (I haven't) a pair of sensitive speakers give a good idea of the noise in the audio band.

Low hum does not require star ground.

I can vouch for that, I am tempted to think that star ground is an adio myth.

There can be certain advantages which are worth the effort. The 3dB additional noise have never been an issue for me

Tell me please about that advantages which worth the effort to spend double money and tolerate 3dB additional noise.

 

[Vinylsavor]

Hi!

Strictly nothing, but unless you have a spectrum analyzer (I haven't) a pair of sensitive speakers give a good idea of the noise in the audio band.

All a sensitive speaker will tell you is about the noise of your power amp.
The audible noise of a phono stage will depend on the gain of the linestage and power amp. Only in conjunction with those gain figures it will tell you something.

Tell me please about that advantages which worth the effort to spend double money and tolerate 3dB additional noise.

It is just the use of different circuits and different parts in the signal path which a differential preamp enabled for me. If it's 'worth' the effort is another story. Some people who listened to my differential and single ended phonos preferred the differential, others single ended.

3dB is a very small increment in noise. If a phonostage is very quiet to begin with, 3dB is very tolerable. This argument of 3dB more noise in balanced circuits keeps coming up, which tells me that many people don't have a feel how much (or how little) 3dB is.

Best regards

Thomas

 

[popilin]

All a sensitive speaker will tell you is about the noise of your power amp.
The audible noise of a phono stage will depend on the gain of the linestage and power amp. Only in conjunction with those gain figures it will tell you something.

OK. In your opinion, when I connect my old Nakamichi, the hisss you hear is also from the power amp ?

How do you know the signal to noise ratio of my power amp ?

They sell good glass balls in Germany, right ?

It is said out there, that the largest contributor to noise is the phono preamp, but keep it quiet.

It is just the use of different circuits and different parts in the signal path which a differential preamp enabled for me.

The use of different circuits and different parts in the signal path implies Balanced amplification ?

If it's 'worth' the effort is another story. Some people who listened to my differential and single ended phonos preferred the differential, others single ended.

Ah, now I understand, pure fashion...

Let me guess, your differential phonos are more expensive, isn't it ?

3dB is a very small increment in noise. If a phonostage is very quiet to begin with, 3dB is very tolerable. This argument of 3dB more noise in balanced circuits keeps coming up, which tells me that many people don't have a feel how much (or how little) 3dB is.

Just these days I'm struggling to reduce 0.4 nV/√Hz in my new phono preamp, for me, every dB count.

 

[Vinylsavor]

Hi!

All I am saying is that the sensitivity of the speakers has little to do with the noise of the phono amp. People keep confusing that. Power amps can range in gain from 6dB for a low SE power amp (which is typically used with sensitive speakers) up to 26dB and even more with high power amps. That means the audible noise through the speaker from the phono can vary by as much as 20dB, depending on the gain of the amp. A low sensitivity speaker will make the noise just as audible as a high sensitivity speaker if it is used with an amp with more gain.

Just wanted to make that clear since this keeps being mixed up.

The rest of the comments in your last post I find rather polemic and not worth replying too.


Thomas