Question on passive summing and phono preamps

[Benjaminbear]

Has anyone ever tried to make the stereo output of a passive summing box to be phono level?  It seems like it would be convenient to just attenuate the signal and add the RIAA pre-emphasis...


Thoughts?

[JohnRoberts]

How is that convenient?? RIAA EQ changes some 40dB (100:1) from 20-20kHz

Basically you are adding one potential frequency response error, before sending it though a path with another.

I vote nahhh

JR

[Benjaminbear]

well my reasoning was the inherent inexpensiveness of phono preamps, but I can see how that wouldn't really work that well.

call it one of those decisions "made by accountants." 

[abbey road d enfer]

well my reasoning was the inherent inexpensiveness of phono preamps, but I can see how that wouldn't really work that well.

call it one of those decisions "made by accountants." 

In addition to what JR says, most "passive summing boxes" have low output impedance, which requires a make-up amplifier optimised for that low impedance; typically a microphone preamp, optimised for 150-200 ohms is adequate. OTOH, a phono pre is optimised for 47kohms impedance.

[Benjaminbear]

i'm sure this question has been asked a gorillion times, but what would be the recommended solution for makeup gain on these things?

i did use the search function but i couldn't find anything with a schematic attached to it

[abbey road d enfer]

i'm sure this question has been asked a gorillion times, but what would be the recommended solution for makeup gain on these things?

A good mic pre.

[JohnRoberts]

i'm sure this question has been asked a gorillion times, but what would be the recommended solution for makeup gain on these things?

i did use the search function but i couldn't find anything with a schematic attached to it

As has been answered a gazillion times, this is a bit of a boondoggle, since the make up gain required for a passive sum is indistinguishable from the noise gain of a virtual earth sum amp.

The impedance of the passive sum is arbitrary and for a modest sum of only a few stems, there is no need for a mic preamp (gain or impedance).

It i s exactly like designing a summing amp for the same number of stems.  There are many modern opamps that would drop in and do well for most cases.

JR

[Benjaminbear]

ok i got it now.  i was overthinking the problem, when all i had to do was actually measure the output.


thanks guys

[pstamler]

In addition to what JR says, most "passive summing boxes" have low output impedance, which requires a make-up amplifier optimised for that low impedance; typically a microphone preamp, optimised for 150-200 ohms is adequate. OTOH, a phono pre is optimised for 47kohms impedance.

Er, not quite. A phono pre is optimized for the impedance of the cartridge (about 500-1000 ohms at DC, plus more from the inductance at high frequencies), in parallel with the input resistor of 47k, which loads the cartridge.

Whether the inductance contributes to the noise depends on the amplifier circuit. The DC resistance contributes Johnson noise, and if you're using a tube or FET circuit that's it (other than the active device's own noise, of course, plus unbypassed cathode resistances and the like). If the amplifier is bipolar, however, there's also current noise to contend with, and the cartridge's inductance will react with that.

So, for example, my turntable for playing 78s has two possible cartridges I can use, a Grado (low-inductance) and a Stanton (high-inductance); they have about the same DC resistance. My preamp is bipolar-based (NE5534a). When I switch from the Grado to the Stanton I see the noise level on the DAW jump up about 3dB.

Peace,
Paul

[abbey road d enfer]

In addition to what JR says, most "passive summing boxes" have low output impedance, which requires a make-up amplifier optimised for that low impedance; typically a microphone preamp, optimised for 150-200 ohms is adequate. OTOH, a phono pre is optimised for 47kohms impedance.

Er, not quite. A phono pre is optimized for the impedance of the cartridge (about 500-1000 ohms at DC, plus more from the inductance at high frequencies), in parallel with the input resistor of 47k, which loads the cartridge.

Whether the inductance contributes to the noise depends on the amplifier circuit. The DC resistance contributes Johnson noise, and if you're using a tube or FET circuit that's it (other than the active device's own noise, of course, plus unbypassed cathode resistances and the like). If the amplifier is bipolar, however, there's also current noise to contend with, and the cartridge's inductance will react with that.

So, for example, my turntable for playing 78s has two possible cartridges I can use, a Grado (low-inductance) and a Stanton (high-inductance); they have about the same DC resistance. My preamp is bipolar-based (NE5534a). When I switch from the Grado to the Stanton I see the noise level on the DAW jump up about 3dB.

Peace,
Paul

Well, I didn't want to go into details of noise generation in RIAA amplifiers. Considering the application (summing box), it is very likely a solid-state solution that was considered. In that case, the increasing impedance of a phono pick-up, combined with the low pass effect of the RIAA equalization make noise proeminent in the high midrange. As a consequence, most of phono preamps are optimized for  an impedance of about 10k, (which happens to be the domain of most LN opamps, such as the archetypical 5534), which makes them less than optimum for the typical 500 ohms of a summing box. The Grado pick-up (45mH) is an oddball; most Shure, Stanton, Ortofon, AudioTechnica et al are high-inductance (450-1000mH)

[dcollins]

I was wondering if anyone had experimented with the "synthesized load" circuit as seen in the Self book? The 47k resistor is replaced with the same V/I parameters yet lower Johnson noise? 

Said to give a dB or two lower noise at the expense of a dual opamp and a few R's and C's.


DC

[JohnRoberts]

I was wondering if anyone had experimented with the "synthesized load" circuit as seen in the Self book? The 47k resistor is replaced with the same V/I parameters yet lower Johnson noise? 

Said to give a dB or two lower noise at the expense of a dual opamp and a few R's and C's.


DC

While I haven't read Self's book, I wasted too many years designing preamps to expect much practical noise benefit from such a scheme.

The 47k input termination is already in parallel with the MM cartridge source impedance for noise analysis, so we are talking circa 1.5k || 47k. In other words we are not starting out with a simple 47k for johnson noise analysis but already more than a decade lower. At HF the cartridge impedance starts rising due to inductance, but there is also a cap termination in parallel with the 47k and the gain of the RIAA EQ is falling 6dB/oct above 2k. 

This is all ancient history by now, or should be.   

JR

[dcollins]

While I haven't read Self's book, I wasted too many years designing preamps to expect much practical noise benefit from such a scheme.

The 47k input termination is already in parallel with the MM cartridge source impedance for noise analysis, so we are talking circa 1.5k || 47k. In other words we are not starting out with a simple 47k for johnson noise analysis but already more than a decade lower. At HF the cartridge impedance starts rising due to inductance, but there is also a cap termination in parallel with the 47k and the gain of the RIAA EQ is falling 6dB/oct above 2k. 

This is all ancient history by now, or should be.   

I just looked at the chapter and in his example of the Shure M75ED cartridge (610R + 470mH) the improvement is 1.5dB over the 47k resistor, fwiw.

There is a table of calculated comparisons of different opamps and discrete (note spelling) transistors showing theoretical improvements from 2.3 to a paltry 0.2 dB with this technique.

DC

[ricardo]

I just looked at the chapter and in his example of the Shure M75ED cartridge (610R + 470mH) the improvement is 1.5dB over the 47k resistor, fwiw.

There is a table of calculated comparisons of different opamps and discrete (note spelling) transistors showing theoretical improvements from 2.3 to a paltry 0.2 dB with this technique.

Could you oblige the impecunious beach bums among us with a scan or 2 of the relevant pages please?

Theoretical max. advantage is 3dB but subject to all the stuff that JR lists and measured by Self.  I'm surprised he gets 2.3dB even.

Anyone have clean enough records to take advantage of this?

IMHO, the electrical noise spectrum will still be whitish so more obnoxious than that of a good MC cartridge into my 0.28nV/rtHz preamp which will be quieter and pinkish.

[dcollins]

Could you oblige the impecunious beach bums among us with a scan or 2 of the relevant pages please?

Theoretical max. advantage is 3dB but subject to all the stuff that JR lists and measured by Self.  I'm surprised he gets 2.3dB even.

Anyone have clean enough records to take advantage of this?

IMHO, the electrical noise spectrum will still be whitish so more obnoxious than that of a good MC cartridge into my 0.28nV/rtHz preamp which will be quieter and pinkish.

I don't know how clean my record is, I may still be on probation with JR for accidentally putting a +N in a spec, so we'll see how it goes............

DC

[JohnRoberts]

Could you oblige the impecunious beach bums among us with a scan or 2 of the relevant pages please?

Theoretical max. advantage is 3dB but subject to all the stuff that JR lists and measured by Self.  I'm surprised he gets 2.3dB even.

Anyone have clean enough records to take advantage of this?

IMHO, the electrical noise spectrum will still be whitish so more obnoxious than that of a good MC cartridge into my 0.28nV/rtHz preamp which will be quieter and pinkish.

I don't know how clean my record is, I may still be on probation with JR for accidentally putting a +N in a spec, so we'll see how it goes............

DC

I don't do probation or police anything more than bad internet behavior... while some people may take my posted opinions as criticisms.. My +n comment was wrt a post that was left hanging for a while in need of clarification.

While I consider phono preamps ancient history I still have a passing interest in the technology as I sold several different preamp designs between the late '70s and mid '80s, and killed many brain cells in that pursuit.

I am familiar with the concept of using active feedback to game input terminations but never seriously pursued it for mic or phono stages.

I can't tell from that chart, is the load synth and 47k noise measurements with a MM cart in parallel or not? For bench noise measurements a dummy cart load would be nice.

FWiW I actually used 2sb737s input devices in a MC phono preamp I did (P-10) back then, but they didn't strike me as optimal for MM, so i used JFET input devices for the MM version. IIRC the 737/786 complementary pair were actually engineered originally for making MC head amps.

JR

 

[ricardo]

Thanks for this Dave.  But like JR, I'm not sure these measurements are with a MM source.

What it does remind me is how good 5534/2 is for MM preamps.

Have you a scan of the topology that Self uses for his virtual input Z?
Here's a 3rd 8ve noise plot of my MC headamp, the quietest in the known universe.

The B&K 2307 chart recorder has slipped by 1 tooth so 50Hz (hum) appears at 40Hz. 1dB/division

The MM is an Ortofon VMS20E. The Ortofon MC, either MC20-2 or MC10. The MC preamp is set for very high gain 28.8dB so if adjusted for equal loudness from the 2 cartridges, the noise advantage of my little circuit is even better. With RIAA, MM noise is whitish while MC is redder than pink.

At the time, every other MC step up device was at least 6 - 10 dB worse (usually much worse) depending on how you measured the noise. So you normally don't see this cos the noise of the step up device.

If you can match a MC properly, you'll get better S/N than MM. MCs have quite high power output even if voltage is low. MM is inherently inefficient in comparison. Also the inductance doesn't let you get close to true efficiency except for 1 frequency.

Does ultra low noise on vinyl playback matter? I think so. From analysing my listening test results and ignoring any obvious Golden Pinnae raving, there is a sense of less grittiness and also the system seems to disassociate clicks, pops & record noise from the music. Maybe a noise modulation effect. But I'm probably pontificating from the wrong orifice and it was 30 yrs ago.   

The loading capacitor also affects the noise, improving it for the usual MMs.

[dcollins]

I don't do probation or police anything more than bad internet behavior... while some people may take my posted opinions as criticisms.. My +n comment was wrt a post that was left hanging for a while in need of clarification.

And I never use smileys.


DC

[ricardo]

Self2.jpg

Thanks for this Dave.  It's a bit late for me to do a detailed noise analysis but ...

Are the numbers in Self's table actually measured or from a simulation?  'Magnoise' and the 0.1dB accuracy suggests simulation.

I'm a rather dubious of his circuit from the noise viewpoint.

[JohnRoberts]

I don't do probation or police anything more than bad internet behavior... while some people may take my posted opinions as criticisms.. My +n comment was wrt a post that was left hanging for a while in need of clarification.

And I never use smileys.


DC

Emoticons are a good concept to improve upon the incomplete communication afforded by basic prose. How well they work is another matter. 
=======

Has anybody built and checked out that proposed preamp? It is not what I expected.

JR