Sutherland KC Vibe MK2 Phono Stage Review

[Michael Fidler]

Getting solid-state, no op amps is a challenge; I can only come up with the Bryston BP-2 with that feature.

@Michael Fidler makes great designs at different price points that combine op amps, in the right places, with bipolar transistors.

Welcome

Classic Audio ltd by Michael Fidler

michaelfidler.com

If you have SMT at your disposal then you can throw quite a lot of transistors in there to make remarkably linear amplifier stages, although the utility of it is somewhat questionable as I can't see much reason for it other than to increase the rails and headroom, but by that point you're exceeding 10V RMS so it's doubtful whatever lies downstream will be able to handle all the output.

The video posted about increasing power supply voltage to combat insertion loss in passive networks would not be my approach though. It would be much better to use an active filter IMO...

 

[LTig]

If you have SMT at your disposal then you can throw quite a lot of transistors in there to make remarkably linear amplifier stages, although the utility of it is somewhat questionable as I can't see much reason for it other than to increase the rails and headroom, but by that point you're exceeding 10V RMS so it's doubtful whatever lies downstream will be able to handle all the output.

IMV it makes no sense to throw in lots of discrete transistors to circumvent opamps since they already have lots of transistors inside with much better tracking and identical behaviour.

It makes sense to use several ultra low noise transistors in parallel together with a low noise opamp for the first gain stage of an MC phono pre (or just get the rather expensive AD797 and call it a day). All other gain and filter stages can be designed easily for very low noise and distortion using proper selected opamps.

The video posted about increasing power supply voltage to combat insertion loss in passive networks would not be my approach though. It would be much better to use an active filter IMO...

Yep. Adding lots of gain to a signal and then attenuate it again through a passive filter makes no sense regarding noise and headroom - one or both of them will suffer.

 

[Angsty]

IMV it makes no sense to throw in lots of discrete transistors to circumvent opamps since they already have lots of transistors inside with much better tracking and identical behaviour.

It makes sense to use several ultra low noise transistors in parallel together with a low noise opamp for the first gain stage of an MC phono pre (or just get the rather expensive AD797 and call it a day). All other gain and filter stages can be designed easily for very low noise and distortion using proper selected opamps.

Yep. Adding lots of gain to a signal and then attenuate it again through a passive filter makes no sense regarding noise and headroom - one or both of them will suffer.

The way I recall the PS designs, Sutherland uses high voltage switching power supplies then uses cascaded LCR filters to smooth out ripples, with resulting voltage loss. For the 20/20, he claims to get the power supply ripples below 1 microvolt. His older designs used DC (battery) power sources.

 

[Michael Fidler]

IMV it makes no sense to throw in lots of discrete transistors to circumvent opamps since they already have lots of transistors inside with much better tracking and identical behaviour.

It makes sense to use several ultra low noise transistors in parallel together with a low noise opamp for the first gain stage of an MC phono pre (or just get the rather expensive AD797 and call it a day). All other gain and filter stages can be designed easily for very low noise and distortion using proper selected opamps.

Yep. Adding lots of gain to a signal and then attenuate it again through a passive filter makes no sense regarding noise and headroom - one or both of them will suffer.

My thoughts exactly, although there is some joy to be had in designing very linear discrete circuitry. One valid application I can think of is a mini class B amplifier for headphones that runs of pre-regulated linear rails of 25V or so for 15V RMS output.

Making a high-voltage linear discrete piece of circuitry without a lot of complication and attenuating the result in a filter seems to be asking for trouble.

 

[Michael Fidler]

The way I recall the PS designs, Sutherland uses high voltage switching power supplies then uses cascaded LCR filters to smooth out ripples, with resulting voltage loss. For the 20/20, he claims to get the power supply ripples below 1 microvolt. His older designs used DC (battery) power sources.

So many filters is rather wacky - the sort of thing Douglas Self would probably refer to as 'clumsy/inelegant', but it certainly fills the board out with expensive parts to justify the cost without much R&D effort. I can't for the life of me see why it's necessary as op-amps have plenty of PSRR to handle the millivolt or so of noise/ripple that you get out of conventional 78/8XX regulators.

Usually when we need such heavy power supply filtering it's for the kind of very low noise discrete circuitry we might want to use for a phono input. In my experience, this can be sucessfully accomplished with a time constant of 0.5 to 1 seconds in a single RC filter to remove all power supply hum and noise from the output. If you need more current (80mA in the case of a fully discrete simple MC head for a mid-priced product I'm currently working on), replacing the resistor with a CCS (pair of BJTs) and using 1000uF of capacitance also does the job perfectly.