Jfet/opamp "hybrid" phono stage

[00940]

I've promised a friend to build him a phono preamp (for MM cartdriges). It has to be decent but no more (he's using an AT95e on a stock thorens td166). A strong output (about 2VRMS) is a plus, to match volume levels from typical CD players.

What I've got on hand : a few opa134, some matched quads of 2sj74/2sk170 (at 8ma), a 24VDC supply (schematic attached).

I'm considering building the following : the input stage is a classic cascoded 2SK170 (idea coming from the "simplistic riaa" by Salas). It's followed by an opamp filter/final gain stage (the active filter values come from the Grado PH1). The opamp itself is buffered by a 2sj74 to drive the filter network. The 10uF feedback cap is slightly undersized, to provide some attenuation below 10hz.

I'm about to order the caps and resistors, so I would really appreciate any comments/suggestions/criticisms.

[stratus46]

2V RMS? That's 5.6 V p-p which is unusually high. 2V p-p would be more in line with CD players and cassette decks. Is there some reason you like a single power supply rather than balanced? Balanced supplies get rid of turn on thumps and some of the capacitors can go away.

G²

[00940]

- 2Vrms is definitely standard with today's CD players and DAC.

- The advantage of a single supply is that I only need one regulator (already built and gathering dust) and that I can power it from a 30VDC printer wallwart.

- Turn on thump can be dealt with by using a relay and a simple RC delay. Turn off thump by using a 3-way switch for power on/off: you cut the outputs when you cut the DC.

[Werner]

1) Heed your circuit's power supply rejection!

2) There is no need for that much gain. CD is 2Vrms ... at 0dBFS.
Average levels on a decent recording are at -12dB or lower, so that is
less than 500mVrms. Cartridges are specified at 5cm/s, which is at their 0dB or average level. Peaks are 12-15dB above this.

Of course there is always the loudness race but that is not a reason to build excessive gain into a phonostage.

3) Why waste these FETs on a cheap preamp? Keep your pairs for something worthier.

4) A lone SJ74 on a resistor isn't worth much as a buffer.

5) If you move the RIAA parts around you can use C10 to close a loop around the output cap.

[00940]

Thank you, everything makes sense but the point 5... what do you refer to by "C10" ?

If I follow you, I might as well just use the lone opamp, using a virtual ground supply to address stratus46's points ?

[Werner]

Sorry, was working from the top of my (poor) head. I meant C7.

When you wrap the other (RIAA) cap around it you have a bit of feedback reducing the sonic impact of C7, so much so that you can perhaps mount a decent elcap. But the whole frequency response has to be simulated in detail, as these things tend to do freaky things in the bass.

--

A totally different path you could take, given that 24V supply, is to re-invent the classic Naim phono circuit, but then for MM only and with a FET as front-end. That's without opamp and with a handful of BJTs.

[00940]

All done

I finally went for something very simple. Virtual ground and the opa134 for the whole thing. Filter values are similar to those used by Grado. A listening test show no problems and neither does the scope. If I can force myself to it, I'll do a frequency response test.

Filter caps are fkp/fkc @2%, resistors are better matched.

The regulator is under the phono board.

[mlloyd1]

even though you have C7, you might want to check that with DC offset, you're not running out of headroom earlier than you planned.

also, for improved HF riaa accuracy, you may want to use an R-C filter at the output since the HF gain flattens.

i like the 3D style construction :-)

mlloyd1

[00940]

Actually, dc offset is surprisingly low. I measured less than 100mv in between the input of C7 and the virtual ground when connected to the turntable and nothing playing. With the input disconnected, it goes up to 1V.

I'm currently testing on an old pionneer turntable with 681eee, a Sony 222ESR amplifier and some good kef columns. It's a tad bright but so was it with the preamp integrated into the sony. I'll tweak the filter if necessary once I get it installed on my friends' system. I might add some resistance on the output anyway, to relieve the opamp driving the cables.

Deadbug is fun.