Balanced input stage beneficial for unbalanced turntables?

[Onvinyl]

Every real world phonoamp I have seen has more issues due to noise pick-up through the cartridge and the cabling then due to inherent noise of the input diff.

What I will try one day, when other projects are finished, is a p-jfet diff pair, ccs'd, and folded-cascoded with two tubes. These should allow quite some current, so one could probably parallel the fet's. The next stage shall change the signal to single-ended and have the riaa. The output stage should be again with tubes, for to take the huge signal swing when the vinyl pops due to dirt.

I disagree wrt the gain, I think the first stage should provide as miuch gain as possible both to feed the riaa network with enough current, good control of overload (due to popping, again) and good noise performance (the input stage should have the best input noise performance)

I never build with tubes so far, so it will take some time till I'll try.

Rüdiger

[mlloyd1]

[QUOTE]Originally posted by Onvinyl
Every real world phonoamp I have seen has more issues due to noise pick-up through the cartridge and the cabling than due to inherent noise of the input diff.

this has been my experience, too. consequently, i am less concerned by the 3dB noise penalty of diff input versus single ended.

mlloyd1

[GlidingDutchman]

You can do a balanced pre-amplification setup with two Single-Ended phono stages. Search the Van Den Hul site for the explination by mr AJ vd Hul himself... very innovative!

D

[EC8010]

If you're using a moving coil cartridge you can have your cake and eat it. You wire your arm wiring to be balanced (twisted pair, plus external screen) and take it via a balanced connector of your choice into a transformer that has been wired to be balanced (no input terminals connected to earth). Even better, you use a transformer that was intended to be used this way and that has balanced impedances to earth, such as the Sowter 8055 or (I believe) the Jensen JT-346. The output can then go to a conventional single-ended amplifier and avoid the 3dB differential pair noise issue. This is how microphone amplifiers in professional mixing desks work.

[forr]

Hi EC8010,

===quote:
Originally posted by Conrad Hoffman
I do have a suspicion, however, that a differential amplifier will never quite equal the noise performance that can be achieved with a single ended amplifier because you need twice the active devices.===

---That's right, the noise goes up by 3dB---


I am in difficulty here.
3 dB more noise, yes but 6 dB more amplification, I think.
So maybe less noise in the end. Am I wrong ?
Some good and very silent mic preamps are buit with a differential input pair, aren't they ?

[Conrad Hoffman]

forr, that's why I still have some doubts, but in a real world design I did (not a phono preamp) that was near theoretical noise limits, going to a differential input always cost me a noise penalty. OTOH, I absolutely agree that hum is a bigger problem in phono preamps than noise, so I'm just bringing up the point, not suggesting that differential is a bad thing to do. Obviously signal to noise ratio is set in the first gain stage, and nothing that comes after can improve it, so you have to look at gain distribution (most of it up front) and the performance of that first stage.

[SY]

I think the first stage should provide as miuch gain as possible both to feed the riaa network with enough current, good control of overload (due to popping, again) and good noise performance (the input stage should have the best input noise performance)

Fully agree. In my phono stage, the first stage has close to 60dB gain, just to get stuff out of the muck. The Sowter 8055 that EC8010 mentioned is responsible for 20 of those dB, and its input balance is so good that hum pickup really is much lower than stage noise. It's a remarkable transformer.

[EC8010]

Originally posted by forr
3 dB more noise, yes but 6 dB more amplification, I think.
So maybe less noise in the end. Am I wrong?

I was assuming that the gain had remained the same. You'd expect the noise to double because each input has only half of the input signal, but when the noise from the two sources (each input) is summed, it only goes up by 3dB because the noise sources are uncorrelated, whereas the signal goes up by 6dB (because those signals are correlated), so the total degradation is 3dB.