Balanced MM phono preamp with SSL9k Front End

[JohnRoberts]

... this is 2011 and were still debating phono preamps? The end must be near.  8)

Phono afficionados:
"Do not go gentle into that good night..."

Gold, if you decide to try John's suggestion of some 2SK389's for the match and noise, they're still findable for fairly reasonable prices (not so the P compliment 2SJ109 which is mega ££'s)  but, if you send me a PM, I have a few and could send you a couple to try. 

Jean - who quite likes the passive eq approach to phono eq - Vive la différence!  8)

I have messed with variants on passive EQ also... The RIAA EQ in the P-10 above follows the time constants all the way up above bat frequency, and my later P-100 uses a real pole for the 75uSec time constant, so likewise accurate way above 20 kHz. Perhaps more important to use good dielectric caps (I used polystyrene). 

I am guilty myself, but lots of effort has gone into over engineering phono preamps that approach perfection, while cartridges and the rest of the early path can vary far more (not to mention room and speakers).  I recall one situation where two different magazine reviewers performed listening tests on the exact same unit and gave it very different reviews. Clearly one was hearing deficiencies in his system, at least that's my story and I'm sticking to it.

Designing phono preamps was fun for me, and I had further tweaks that I never pursued when it became apparent to me that there was no connection between sound quality (linearity, precision, low noise) and love in the audiophool market. Not to mention the emergence of CDs. 

No regrets, but perhaps a few sour grapes.  8)

JR
 

 

[Gold]

This has turned into a nice discussion. I'm not in immediate need for a preamp. I am using this as an opportunity to learn something about high gain circuits. I don't use microphones so I never payed much attention to mic pre circuits. The only high gain circuits in my studio are a phono pre and the feedback amp for the cutting rack.

I guess my goal is to design a preamp that would be useful for cutters. I don't see this happening soon.

The Neumann preamp in the SAL74 rack is nice feature wise but I think they may have gotten the approach wrong. They did this fancy driven shield design to lower cable capacitance because they knew the run would be long. I think noise is the greater problem.

I thought I would try Belden Mediatwist bonded twisted pair CAT6 cable. Steve Lampen from Belden touts a shield as unnecessary because of the cable geometry. No shield would help with cable capacitance. I envision the cable being part of the system. I could always add a braided shield. Belden doesn't make an STP cable with bonded twisted pair.

I think cable capacitance could be taken care of with active EQ if necessary. I would probably include a nice EQ to flatten out cartridge response. That's a really nice feature of the Neumann pre amp.

 

[JohnRoberts]

This has turned into a nice discussion. I'm not in immediate need for a preamp. I am using this as an opportunity to learn something about high gain circuits. I don't use microphones so I never payed much attention to mic pre circuits. The only high gain circuits in my studio are a phono pre and the feedback amp for the cutting rack.

As PRR alluded to the gain requirements of the preamp is not all that hard thanks to the RIAA curve. While circa 60dB at 20 Hz it drops to only 20 dB at 20 klHz so even modest opamps can often deliver serviceable results.

I guess my goal is to design a preamp that would be useful for cutters. I don't see this happening soon.

The Neumann preamp in the SAL74 rack is nice feature wise but I think they may have gotten the approach wrong. They did this fancy driven shield design to lower cable capacitance because they knew the run would be long. I think noise is the greater problem.

It seems both could be an issue and driving the shield is interesting, while adding extra complexity.

We are not home free yet on the noise.. turntable/tone arm wiring can be unshielded twisted pair, relying upon the turntable metal for shielding. You may need to experiment some to get a good result.

I thought I would try Belden Mediatwist bonded twisted pair CAT6 cable. Steve Lampen from Belden touts a shield as unnecessary because of the cable geometry. No shield would help with cable capacitance. I envision the cable being part of the system. I could always add a braided shield. Belden doesn't make an STP cable with bonded twisted pair.

?? Unnecessary. We are not dealing with very low source impedance, then applying gains in the 50-60 dB range at 60 Hz.. I would advise shielding.

I think cable capacitance could be taken care of with active EQ if necessary. I would probably include a nice EQ to flatten out cartridge response. That's a really nice feature of the Neumann pre amp.

Active EQ to correct for cable loading is surely practical... Being a purist not my first choice.

If you are ready for an adventure let me suggest another possibility...

it is not out of the question to build some simple discrete JFET buffers, perhaps phantom powered over the shielded cable used. The modern JFETs are quiet enough that you could probably get away with running the buffers at unity gain. placing this small buffer PCB near the cartridge/tonearm would make capacitance a non-issue and the lower source impedance output would be less susceptible to noise pickup..

The mV level cart signals would not be very distorted by simple  JFET followers and you could improve linearity by increasing the current density, perhaps also supplied through the cable. This is mostly a thought exercise, but not crazy... My later P-100 ran the input JFET effectively open loop and exhibited respectably low distortion, while i didn't locate the input JFET tens of feet away from the rest of the preamp. So I wouldn't just spread out that design since drain current could get corrupted over a long path, but using the input JFET as a low impedance source follower should work.

Simpler yet using SMD opamp buffer at the cart would only be slightly larger and modern uber opamps could deliver respectable noise, and perhaps even lower distortion than a bare JFET follower..

Looks like several choices that should make an improvement. Simple is good.


JR

 

[Gold]

?? Unnecessary. We are not dealing with very low source impedance, then applying gains in the 50-60 dB range at 60 Hz.. I would advise shielding.

I guess I didn't understand the whole picture. He was talking about line level signals.

Active EQ to correct for cable loading is surely practical... Being a purist not my first choice.

Agreed.

it is not out of the question to build some simple discrete JFET buffers, perhaps phantom powered over the shielded cable used. The modern JFETs are quiet enough that you could probably get away with running the buffers at unity gain. placing this small buffer PCB near the cartridge/tonearm would make capacitance a non-issue and the lower source impedance output would be less susceptible to noise pickup..

I think this is what they did at Masterphonics. They built the transistor front end into the head shell and used phantom power. That would probably be the ultimate solution quality wise. A little box that could be easily mounted to the lathe would work. Good idea.

 

[tv]

In regard what you're after, a 47kOhm phantom-powered "headamp" would be the way to go imo. But unity-gain buffers are a not-so-good idea imo  - unless you want to reuse some existing phono preamp. Personally I would use some differential fet stage (with gain), feeding a differentially connected passive RIAA - and you're off to races...

Here are some "complicated" schematics you could use as a startpoint. (MKIII and MKIIIb)
http://www.klaus-boening.de/html/schematics.html

He uses a discrete "headamp" feeding a passive RIAA netwerk, and you can see this circuit could be adapted to be used remotely as a 47kOhm phantom-powered "headamp".

You could strip away the overengineered input circuit, and use that as a startpoint for further tinkering.

You could do a functionally similar thing with opamps...

If you are after a very clean, noise-free signal, a "micpre" chips a-la INA and THAT would have higher noise-floor because their input stages would inject ahigher amount of current-noise.. but the simplicity of it does have a certain appeal.

 

[tv]

here's a opamp concept.. fill-in your values, but keep the resistances within circuit fairly low to keep the self-noise down. use some fet super-opamp..

 

[JohnRoberts]

While you could literally build a buffer into the head shell I wouldn't want to add more mass, and perhaps wires to the tone arm.

If you can find a really nice (low noise) JFET dual, perhaps take just the left input side of TV's schematic, perhaps without the gain, and 470 ohm build outs.. 

Unity gain and 50 ohm build outs works for me, and using SMD this could be tiny.. I would throw in at least some .1uFs on the power supply rails near the opamp, since they will now be driving the long cable capacitance.  KISS

JR 

 

[Gold]

Thanks TV. I'm liking the remote phantom head amp concept more and more. I guess I would have to run the opamps on a single ended supply via phantom power or use multi conductor cable and send +/- 15v separately.

I need to look into the discrete transistor front end a little more. Powering it seems like it would be easier. Is the advantage of using a unity gain buffer that the output impedance is lowered to a microphone like source impedance and easier to send across long lines?

 

[tv]

http://sound.westhost.com/project35.htm

scroll down to see a simple implementation for a phantom-powered opamp..

..or dissect some other gear to see how it's done..

If you are not-so-experienced with electronics, build a opamp version first - these tend to "just work". With discretes, you can develop a headache before you "do it right"..

 

[Gold]

If you are not-so-experienced with electronics, build a opamp version first - these tend to "just work". With discretes, you can develop a headache before you "do it right"..

Right. The discretes scare me. I don't even know the basic configurations well. I might very well start with a mic pre chip. It's always nice to get something that works rather than start way above my pay grade.

 

[JohnRoberts]

Thanks TV. I'm liking the remote phantom head amp concept more and more. I guess I would have to run the opamps on a single ended supply via phantom power or use multi conductor cable and send +/- 15v separately.

Phantom power might be useful if you want to mount the buffer inside the head shell and not add wiring. I am not up to speed on the SOTA for very low noise Pch JFETs. I lean away from using bipolar devices, while a couple PNP bipolar emitter followers could be shoe horned into the head shell if you share the cart lows with follower grounds.

I need to look into the discrete transistor front end a little more. Powering it seems like it would be easier.

Could be similar but if you add one wire it's not much more effort to add 3.

Is the advantage of using a unity gain buffer that the output impedance is lowered to a microphone like source impedance and easier to send across long lines?

I was suggesting unity gain for simplicity... A 1 nV rt/Hz opamp should not increase noise floor over the cart self noise, and could plug directly into your existing phono pre.

If you add gain, you will want to pad that down, or build the rest of a preamp with RIAA EQ.

1- simple dual opamp buffer requires supplying +/- power to distant PCB.

2- simple Nch JFET source follower, needs simple + supply., and signal current will flow in ground so perhaps differential that. 

3-phantom powered P-ch JFET possible?  Phantom powered PNP emitter follower easier but use a high beta low noise device so it doesn't dump base current into cart... (this is probably manageable as low noise devices are usually high beta, but i just prefer JFETs here).

There are probably multiple good solutions. I still like keeping it simple.. i.e. unity opamp buffer built into the base of the turntable, or near the  tone arm, not inside it..

JR

 

[Gold]

I was suggesting unity gain for simplicity... A 1 nV rt/Hz opamp should not increase noise floor over the cart self noise, and could plug directly into your existing phono pre.

Ahh, I didn't get the last part of that. I think I'd like to go for the whole shebang. Remote head amp with RIAA and any other features in another box sounds like a plan.

 

[Jean Clochet]

Post removed.

 

[Jean Clochet]

Post removed.

 

[JohnRoberts]

No thanx (on Pch JFETs)  for me..  I have no desire to reinvent any more wheels.

=======

As I have openly offered that my p-10 and later P-100 are probably over-engineered for the task at hand.

While the p-10 upgraded with more modern JFETs will work well and meets several of your requirements...

=======
re: merit of passive EQ...  perhaps Gold can share what the -3dB point of his cutting lathe is. In practice I think the divergence from true RIAA caused by the simple one opamp phono preamp topology occurs up around 200 kHz (gain plateaus at unity instead of still falling further. Ironically this zero in the playback response may actually be more accurate than tweaky passive EQ designs (including mine).

I expect a one opamp topology using a modern low noise FET input opamp would sound respectable.

The two opamp split/passive EQ would also work fine.

FWIW the good RIAA EQ caps that I like are probably several times larger than good SMD opamps.  ;D

I Predict whatever you choose will make a good improvement.  8)

JR

 

[Gold]

While the p-10 upgraded with more modern JFETs will work well and meets several of your requirements...

Yes they would and I may very well try it out. I would like to take this opportunity to tiptoe into "design". Well, maybe advanced copy and paste would be a better description.

perhaps Gold can share what the -3dB point of his cutting lathe is.

I don't know but I could scan and post the RIAA encode circuit for the cutting rack if you are interested.

 

[JohnRoberts]

perhaps Gold can share what the -3dB point of his cutting lathe is.

I don't know but I could scan and post the RIAA encode circuit for the cutting rack if you are interested.

I have an academic interest in this but don't need this to get you squared away. It might be nice to prove or disprove my old ASSumptions (that 200 kHz response doesn't conform to RIAA curve). 

IIRC I asked a friend of mine who had a working  cutting lathe in his recording studio complex, back when I still designing preamps but don't recall his specific answer. While I dismissed the need for accuracy up at 200 khz there was another old quad system (CD-4?) that used a carrier and sideband up above the normal RIAA so extended up to 50kHz or so...  Not sure these were mastered on any old lathe tho...

My judgement is some extra rolloff at 200kHz is more good than bad....  so true RIAA up high could be good for other reasons than playback accuracy (for bats).

JR

 

[Gold]

I have an academic interest in this but don't need this to get you squared away. It might be nice to prove or disprove my old ASSumptions (that 200 kHz response doesn't conform to RIAA curve). 

I have enough to chew on for a while. I'm not particularly interested in arguing about what's going on out of the audio band. And the audio band for records is more like 15k than 20k

While I dismissed the need for accuracy up at 200 khz there was another old quad system (CD-4?) that used a carrier and sideband up above the normal RIAA so extended up to 50kHz or so...  Not sure these were mastered on any old lathe tho...

There were two competing quad systems but that was before my time. I know the JVC system was proprietary. You had to have the disk mastered at their facility. The other system used a specially designed Ortofon cutting system.

My lathe rack is the older Neumann VG66. I don't think anyone was too worried about those poles in the mid 1960's. I'll post he schematic. I think the later SAL74 did address this. I don't have those schematics as readily available. Sorry for the image split. I don't know how to stitch them together

 

[Gold]

I have an academic interest in this but don't need this to get you squared away. It might be nice to prove or disprove my old ASSumptions (that 200 kHz response doesn't conform to RIAA curve). 

I have enough to chew on for a while. I'm not particularly interested in arguing about what's going on out of the audio band. And the audio band for records is more like 15k than 20k

While I dismissed the need for accuracy up at 200 khz there was another old quad system (CD-4?) that used a carrier and sideband up above the normal RIAA so extended up to 50kHz or so...  Not sure these were mastered on any old lathe tho...

There were two competing quad systems but that was before my time. I know the JVC system was proprietary. You had to have the disk mastered at their facility. The other system used a specially designed Ortofon cutting system.

My lathe rack is the older Neumann VG66. I don't think anyone was too worried about those poles in the mid 1960's. I'll post he schematic. I think the later SAL74 did address this. I don't have those schematics as readily available.

 

[JohnRoberts]

Kool... I'll have to wait until I'm sober to follow that...

Later..
.
cheers

JR

=======
OK, it's later. Upon sober reflection, it's still not that easy to follow that schmo but I see what looks like a two pole LPF at just over 30 kHz on the very output so my assumption that it wasn't tracking true RIAA up to 200 kHz seems pretty secure. Also safe that they weren't cutting 45kHz sidebands for CD-4 quad records either.
------
I'm not sure i want to open this can of worms but using a ua709 opamp in a critical audio path looks like there should be some room for improvement there.  I would approach this with extreme caution since they have probably invested a great deal of effort into getting this right. So who knows what you would get from just dropping a better part in that very old socket? Life isn't usually that simple.

709s are probably hard to buy these days so I'd be careful about even probing around. As i recall 709s were easy to kill. So first do no harm.

JR