A Low Noise Balanced Input Moving Coil Preamp Using the ZTX851

[JR.]

I would say, measure where possible to find all possible design flaws first, but at the end listening to music can only be the final test.

Nonsense!

The most important thing is the Price Tag.

Secondly, you have to deface any 5532/4s so they can't be identified and print "hand carved from solid Unobtainium by Virgins" label on each OPA. This can be a retrospective mod which will improve the sound to no end.

BTW, I've (really) tested the first extensively in DBLTs. When you reveal to the subject what he was listening too and its price tag, his opinion changes dramatically. Applies to everyone, even me. My most successful design was inexpensive and looks it. I ALWAYS mark it higher when tested blind. At one time it was one of the 3 best small speakers in the world.

I am weary of this thread veer, but have an anecdote to add about your "perception bias" wrt price note.

Years ago Peavey (AMR division) has a series of studio monitors that didn't suck (accident I guess ). A recording magazine assembled a panel of recording studio professionals from the SF bay area and performed listening tests between competitive sized studio monitors. The AMR monitors had a two position switch (with different voicings) so were tested as two different speakers in the comparison. I suspect the test was only single blind (?), but after the test results showed that the AMR 8" 3 way won both 1st and 3rd place overall, and the results were shared, a few of the studio professionals involved withdrew from the test and refused to allow their names to be made public.

Brand perception is a powerful thing, makes some people ignore their own ears.

JR

______________________________

Hans, are you going to show us the circuit for which you posted those measurements?
______________________________

We also eagerly await your DBLTs on whether a good digital codec audibly affects the sound of analogue LP. I can advise on the conduct of DBLTs as I was a true DBLT guru in my previous life.

For a start, test 3 at a time, ie A, B & C of which 2 are the same chain. ABC testing gives you statistical significance a LOT faster than ABX.

You test each person individually. Let each person choose his own music and listening level

Frequency Response and Level for the 3 presentations need to be matched to 0.1dB 20 - 20k Hz

There's a LOT more .. enough for a book(s).

[*} Unobtainium, as you know, is in very short supply. A much more readily available substitute is solid BS. But I advise you to say 'Unobtainium' on the label.

[Hans]

Hans, are you going to show us the circuit for which you posted those measurements?

No problem, to which one are you referring: the graph with the 3 different noise measurements, the measurement of the CMRR or the "measurement" of the signal with/without A/D/A.

We also eagerly await your DBLTs on whether a good digital codec audibly affects the sound of analogue LP. I can advise on the conduct of DBLTs as I was a true DBLT guru in my previous life.
For a start, test 3 at a time, ie A, B & C of which 2 are the same chain. ABC testing gives you statistical significance a LOT faster than ABX.
You test each person individually. Let each person choose his own music and listening level

Good points, valid advice.
I have planned to use two different D/A converters, one with a conventional FIR filter as used by most suppliers of D/A converter chips and one with an apodizing filter, having no pre-ringing before the signal starts.
Listening will then be a possible choice from 3 different options over two sessions doing an A/B/C.
Session 1: direct from pick up to amplifier / with a conventional A/D/A in between
Session 2: direct from pick up to amplifier / with an apodizing A/D/A in between

Hans

[Hans]

I'm coming back to an earlier discussion on how how to cope with the current to flow into a cart.
Using an electrolytic capacitor to block too much cart current, may result in a serious worsening of the signal.
Cyril Bateman did extensive investigation on the distortion caused by capacitors.
He did the testing for several capacity groups. The link provided here is for the group "Electrolytics from 10uF to 100uF."
https://linearaudio.nl/sites/linearaudi ... ents_0.pdf

This test shows how disastrous distortion can be with various types of electrolytic caps.
Best results were obtained by connecting 2 non-polar electrolytic caps, having twice the needed capacity, in series.

So almost by definition, a phono-preamp having an polarized electrolytic in its input, cannot be regarded as a top performer and should not in any way be compared to the circuit in this thread.

Hans

[ricardo]

No problem, to which one are you referring: the graph with the 3 different noise measurements, the measurement of the CMRR or the "measurement" of the signal with/without A/D/A.

The circuit for which you posted 10Hz_2Khz_1.jpg

A recording magazine assembled a panel of recording studio professionals from the SF bay area and performed listening tests between competitive sized studio monitors. The AMR monitors had a two position switch (with different voicings) so were tested as two different speakers in the comparison. I suspect the test was only single blind (?), but after the test results showed that the AMR 8" 3 way won both 1st and 3rd place overall, and the results were shared, a few of the studio professionals involved withdrew from the test and refused to allow their names to be made public.

I've had exactly the same thing happen to me except mine were UK HiFi Reviewers.

After nearly 2 decades of DBLTs, I'm prepared to stick my neck out and say HiFi reviewers are generally worse (less consistent) in DBLT than the Man in the Street who is worse than the Woman in the Street who usually has no interest in HiFi. So pay attention if your girlfriend, wife, mother says of your new supa dupa system, "I like the old one better".

Also any one claiming to be a Golden Pinnae (chalk & cheese differences between cables bla bla) is almost certainly deaf (random results in DBLTs). As I said, your measuring instrument is your DBLT panel. It has an accuracy that you have to check and calibrate regularly ... which is VERY expensive.

There are exceptions. One of the best ears to grace my panel was a UK reviewer who is now prominent in the US. But even the best ears are prejudiced. I have great respect for his comments blind .. but take with a large pinch of salt what he says sighted. But he is one of the few honest enough to change his opinion based on his own DBLT results.

[JR.]

I'm coming back to an earlier discussion on how how to cope with the current to flow into a cart.
Using an electrolytic capacitor to block too much cart current, may result in a serious worsening of the signal.
Cyril Bateman did extensive investigation on the distortion caused by capacitors.
He did the testing for several capacity groups. The link provided here is for the group "Electrolytics from 10uF to 100uF."
https://linearaudio.nl/sites/linearaudi ... ents_0.pdf

This test shows how disastrous distortion can be with various types of electrolytic caps.
Best results were obtained by connecting 2 non-polar electrolytic caps, having twice the needed capacity, in series.

So almost by definition, a phono-preamp having an polarized electrolytic in its input, cannot be regarded as a top performer and should not in any way be compared to the circuit in this thread.

Hans

I am not arguing for using electrolytic caps in the MC preamp input. I advocate designing such caps out of audio paths when possible (see Wayne's floating mic preamp to eliminate input caps), but it is worth trying to understand the application before making such a broad sweeping condemnation.

Bateman's excellent series on capacitors is stressing the electrolytic caps at high enough current to generate significant terminal voltage across the capacitors. Capacitor non-linearities express only when there is terminal voltage due to significant charging and discharging current. In a simple DC blocking application the capacitor is passing very little current, and developing almost no terminal voltage at any audio frequency.

Bateman's advice is perhaps more useful when using the electrolytic caps to form filter poles in a passive loudspeaker crossover where they are seeing significant current and terminal voltage at audio frequencies. For DC blocking in active electronic circuits the current, terminal voltage, and therefore distortion will be significantly lower.

Note: electrolytic caps used in gain legs or in series with gain pots and low impedance for mic preamps can be stressed harder than just basic DC blocking.

JR

[JR.]

I've had exactly the same thing happen to me except mine were UK HiFi Reviewers.

After nearly 2 decades of DBLTs, I'm prepared to stick my neck out and say HiFi reviewers are generally worse (less consistent) in DBLT than the Man in the Street who is worse than the Woman in the Street who usually has no interest in HiFi. So pay attention if your girlfriend, wife, mother says of your new supa dupa system, "I like the old one better".

Here is another amusing anecdote told to me by a publisher of a long defunct audiophile magazine. He was in a restaraunt having dinner with the owner of some audiophile product manufacturer, and the guy's wife starts complaining about how unrealistic the restaurant's background music sounds. Except the music was coming from a piano being played in the next room.

JR

[Hans]

No problem, to which one are you referring: the graph with the 3 different noise measurements, the measurement of the CMRR or the "measurement" of the signal with/without A/D/A.

The circuit for which you posted 10Hz_2Khz_1.jpg

O.K., that's one step in the direction, but what do you mean with "circuit".
You probably don't ask for the very obvious things like having used a Pick-Up, a test record with an 1kHz signal and a blank trace, a phono-preamp, a preamp and a digital oscilloscope, so what else do you want to know exactly ?

Hans

[JR.]

No problem, to which one are you referring: the graph with the 3 different noise measurements, the measurement of the CMRR or the "measurement" of the signal with/without A/D/A.

The circuit for which you posted 10Hz_2Khz_1.jpg

O.K., that's one step in the direction, but what do you mean with "circuit".
You probably don't ask for the very obvious things like having used a Pick-Up, a test record with an 1kHz signal and a blank trace, a phono-preamp, a preamp and a digital oscilloscope, so what else do you want to know exactly ?

Hans

I believe he is asking for the phono preamp schematic as that will dominate noise performance.

Many of us here are circuit designers and viewing a schematic is required before we can intelligently discuss the finer points of a design (details matter).

If you are unwilling to share I am perfectly content with not seeing it.

JR

[Hans]

The single reason for posting the noise graph was to show that there is a limit to were it still makes sense lower the SNR of the Phone-Preamp, because of LP's surface noise.
The used preamp for making the graph has a SNR-A of 75dB ref 0.5mV@1kHz which seems to be more than adequate when driven by a Benz LP Cart, having a relatively high Ri of 38 Ohm.
For Carts having 6dB less output which also have a much lower Ri, like the SL Carts from Lyra and the Ortophon Anna, there remains still a sufficient margin of 10dB at 1kHz

A SNR of 75dB-A, input shorted, boils down to roughly 1.4nV/rtHz@1kHz at the preamp's input.
With a 38 Ohm Benz LP connected this will become 1.6nV/rtHz, resulting as shown in the graph, in a margin of 15dB below surface noise .
With a 3.3 Ohm Cart connected instead, input noise will be 1 dB lower with 1.4nV/rtHz@1kHz.
So even with 6 dB less output, the low output and low Ri Cart will be 10 dB (=15-6+1) below surface noise at 1kHz, which still seems to be adequate.

This is not meant to criticize the nice circuit design in this thread, having only 0.33nV/rtHz or 12.5dB less noise.
It is only to make aware that in case of using Fet's instead of bipolars, generating slightly more noise, does not at all render the circuit into a less valuable design.
There is still enough margin to negotiate.

Hans

[Hans]

Many of us here are circuit designers and viewing a schematic is required before we can intelligently discuss the finer points of a design (details matter).
If you are unwilling to share I am perfectly content with not seeing it.

The design will be published in full detail in the next issue Vol12 of Linear Audio.