A Low Noise Balanced Input Moving Coil Preamp Using the ZTX851

[ricardo]

The tested very low noise preamp might be just a better design with a higher overload margin and maybe even a balanced design with a high CMRR compared to the other one in the test.

Apart from two ZTX bipolars, being steadily improved with modern production techniques, all active components in Wayne´s circuit can be modern components from today´s industry, not comparable what was available some 30 years ago.
The circuit is balanced, I agree with Wayne the "magic" advantages of that topology, and CMRR will be very high too.
There is just one thing of concern. The overload margin being only ca. 16dB ref 5mV@20kHz, exactly in the high frequency region where scratches causing sharp clicks, may cause a serious overload.
Reducing the gain of the head amp quite a bit, would inversely increase the overload margin at the higher frequencies.

Last but not least, I'm amazed by the ease you accept positive listening results from an ultra low noise preamp, yet having great difficulties in accepting that A/D/A in the signal line of a Pick-Up may change the listening experience.

I didn't "accept" this. I TESTED it. In 1980, there was at least one codec which was completely transparent to the best ears in the business. I repeat, a DBLT is a measurement.. You instrument is your DBLT panel. It has an accuracy that you can determine & should check regularly.

Here's more from DBLT101.

For something like the effect of a codec on your vinyl playback chain, you do ABC tests. 2 of these 'presentations' are the same but you NEVER tell your panel this or say they are to find the odd man out. You just ask them to rank the 3 presentations and give them a score without telling them what the test is about or what they are listening too. The true golden pinnae will rank the odd man either top or bottom and say the differences are small for the other 2.

You repeat the test 3 times. Each test may take up to 1/2 a day and only 1 person is tested at a time. Anyone who gets the same odd man in the same position 3 times is not deaf. There is a 1:9 chance of him guessing. Apologies to the statisticians reading this but this is probably the quickest way to get useful statistical significance.

You IGNORE the results of anyone who puts the odd man in the centre. They are deaf for the purpose of this test and not worth testing the 2nd & 3rd time.

THEN you ask which of the 'presentations' is preferred ... and sometimes this gives the most surprising results of all. Only the opinion of true golden pinnae counts. For da wannabe Golden Pinnae, you just tell them LOUDLY AND REPEATEDLY, your stuff is Hand Carved from solid Unobtainium by Virgins.

It can only be regarded as astounding how music still sounds after all this "deformation" of the original sound, especially because of all the steep digital filters.

I tested the effect of brickwall filters in those days and found that off those who could tell the difference, ALL preferred the bandlimited signal. I was surprised to find all other properly conducted tests also came up with the same result.

Some 30 yrs later, I was amused to find that ALL Bandwidth Limitation tests conducted since then came up with the same results. (including one with a 17kHz bandwidth)

Maybe I should market a black box which I'll guarantee to improve the performance of any chain in a DBLT. Nay. Only a very small population can actually hear the difference and practically none are audiophools. The Hand Carved from solid Unobtainium by Virgins label is a lot cheaper and lends itself to more marketing BS.
_______________

BTW, I did say I thought the good performance of my device could be due to other stuff besides its ultra low noise .. eg its overload performance.

Remember, a vinyl playback system has HUGE distortion which goes up with level & frequency. The HF distortion goes up at 18dB/8ve. In this situation it is highly likely that a device with complementary characteristics might result in better sound.
_________________

Hans, I'm interested to know which OPAs you feel are superior to 5532/4 and for what applications.

And you are certainly wrong in thinking good 1980's MC cartridges like the EPC100, AT37, Koetsu, the better Ortofons are worse than present day Golden Pinnae product.
________________

Paul, you are right about RFI/EMI etc.

Based on my listening tests on EVIL digital systems, I'd say 20dB below the 'theoretical' (?!) noise floor for everything else is worth aiming for. I've dabbled with phono cartridge design but with much less success than speakers & microphones. I worked with a couple of guys who had designed MC cartridges for LEAK in pre-Jurassic times.

They had quite a large collection of lacquers, some done direct to disc .. all stored (like Test Records) with detailed notes of how many times each was played and with what equipment. The noise level on these was certainly a lot quieter than on commercial records. No. I didn't use these for my listening tests as one of its precepts is the listener ALWAYS chooses his own music.

[Hans]

Here's more from DBLT101.

For something like the effect of a codec on your vinyl playback chain, you do ABC tests. 2 of these 'presentations' are the same but you NEVER tell your panel this or say they are to find the odd man out. You just ask them to rank the 3 presentations and give them a score without telling them what the test is about or what they are listening too. The true golden pinnae will rank the odd man either top or bottom and say the differences are small for the other 2.

Your information on how to perform a dblt is very informative and constructive .
A few point are not completely clear to me, so if you allow me I have a few questions

You mention that a test person should rank 3 music fragment of which 2 are the same. Later in your posting I read that the test person should tell his preference.
So therefore my question: what criterium was than used for ranking ?

The music played, should be music of his choice, but how long is he allowed listen or can he decide himself to go from A to B to C and can he repeat or switch randomly between the three?

You repeat the test 3 times. Each test may take up to 1/2 a day and only 1 person is tested at a time. Anyone who gets the same odd man in the same position 3 times is not deaf. There is a 1:9 chance of him guessing.

I assume that for every new session out of three, A and B are randomly assigned to A,B and C , so the above text must probably be read as "Anyone who gets the same odd man in the right position(s)....." ??

Some 30 yrs later, I was amused to find that ALL Bandwidth Limitation tests conducted since then came up with the same results. (including one with a 17kHz bandwidth)

Funny enough, a similar test was performed recently in the Netherlands with exactly the opposite outcome.
Since everybody was very upset with the outcome, the test will be redone with a new set of test candidates before publishing. It just shows that test results are not always what you expect.
Maybe you still have a link to one of these test you are referring to.

Hans, I'm interested to know which OPAs you feel are superior to 5532/4 and for what applications.

You asked me which opa's I prefer above the 5532, which is quite a tricky issue, since some almost see it as the best opa ever, but nevertheless, I won't hide for my opinion.
From a price performance point of view, the 5532 is hard to beat and the measured results are very good.
But you are right in guessing that I don't see it as the best sounding amp.
Just looking at distortion does not guarantee the best sound.
Pickups, tube amps and loudspeakers can produce excellent sound having a whopping 1% distortion, so what exactly is the meaning of an opa producing only 0,0001% distortion.

In all situation where I replaced the 5532, recontruction filters in Dac's and preamps, I preferred the sound of several other opa's.
For use in preamps, I very much like the opa627 being ridiculously expensive these days. But also the lm4970, the opa1612 and the opa1642 are excellent opa's, the latter probably the best opa to be used for an MM phono preamp.
The AD797 with 0.9 nV/rtHz and a GBW of 100MHz is also a top performer, but only for very low level signals in certain situations and not at all as a replacement for the 5532.

Hans

[JR.]

Here's more from DBLT101.

For something like the effect of a codec on your vinyl playback chain, you do ABC tests. 2 of these 'presentations' are the same but you NEVER tell your panel this or say they are to find the odd man out. You just ask them to rank the 3 presentations and give them a score without telling them what the test is about or what they are listening too. The true golden pinnae will rank the odd man either top or bottom and say the differences are small for the other 2.

Your information on how to perform a dblt is very informative and constructive .
A few point are not completely clear to me, so if you allow me I have a few questions

You mention that a test person should rank 3 music fragment of which 2 are the same. Later in your posting I read that the test person should tell his preference.
So therefore my question: what criterium was than used for ranking ?

I'm no Ricardo... But I believe the point of most DBLT is to determine "if" a difference is reliably heard, only after that would you try to make a "what" determination. The successful golden ear ranks the same samples as the same, better or worse than the other doesn't matter yet. If they rank the same sample differently they just voted themselves off the panel.

Another unfortunate truism about audiophool products is to actually sound different from others, the transfer function needs to be different. I recall a magazine review of a high end (Mark Levinson) phono preamp and the frequency response had a fraction of a dB deviation in the middle of the audio band. Not a big difference, but when listeners hear any difference between a $5000 preamp and their $200 Sony they ASSume the expensive one is better (expectation bias).

The music played, should be music of his choice, but how long is he allowed listen or can he decide himself to go from A to B to C and can he repeat or switch randomly between the three?

You repeat the test 3 times. Each test may take up to 1/2 a day and only 1 person is tested at a time. Anyone who gets the same odd man in the same position 3 times is not deaf. There is a 1:9 chance of him guessing.

I assume that for every new session out of three, A and B are randomly assigned to A,B and C , so the above text must probably be read as "Anyone who gets the same odd man in the right position(s)....." ??

Some 30 yrs later, I was amused to find that ALL Bandwidth Limitation tests conducted since then came up with the same results. (including one with a 17kHz bandwidth)

Funny enough, a similar test was performed recently in the Netherlands with exactly the opposite outcome.
Since everybody was very upset with the outcome, the test will be redone with a new set of test candidates before publishing. It just shows that test results are not always what you expect.
Maybe you still have a link to one of these test you are referring to.

Yup bandwidth is another holy grail argument, but more argued about digital paths. Luckily not much of a point with RIAA that is only defined 20 to 20kHz, while there is much hand wringing among RIAA circuit tweakers about how to handle the octaves above 20 kHz.

JR

Hans, I'm interested to know which OPAs you feel are superior to 5532/4 and for what applications.

You asked me which opa's I prefer above the 5532, which is quite a tricky issue, since some almost see it as the best opa ever, but nevertheless, I won't hide for my opinion.
From a price performance point of view, the 5532 is hard to beat and the measured results are very good.
But you are right in guessing that I don't see it as the best sounding amp.
Just looking at distortion does not guarantee the best sound.
Pickups, tube amps and loudspeakers can produce excellent sound having a whopping 1% distortion, so what exactly is the meaning of an opa producing only 0,0001% distortion.

In all situation where I replaced the 5532, recontruction filters in Dac's and preamps, I preferred the sound of several other opa's.
For use in preamps, I very much like the opa627 being ridiculously expensive these days. But also the lm4970, the opa1612 and the opa1642 are excellent opa's, the latter probably the best opa to be used for an MM phono preamp.
The AD797 with 0.9 nV/rtHz and a GBW of 100MHz is also a top performer, but only for very low level signals in certain situations and not at all as a replacement for the 5532.

Hans

[ricardo]

You mention that a test person should rank 3 music fragment of which 2 are the same. Later in your posting I read that the test person should tell his preference.
So therefore my question: what criterium was than used for ranking ?

This is entirely up to the listener. In earlier attempts at formal testing, we used a huge list of characteristics but this was at best confusing and didn't produce any more info than asking for comments on the presentations. http://www.aes.org/e-lib/browse.cfm?elib=3798

He is asked to rank the presentations to help him make up his mind.

At the end of the test, we get him to add up the 'scores' for each 'presentation'. Then he is asked for an overall score for each presentation which doesn't have to tally with the 'total' scores. It's very rare for anyone to give an 'overall' ranking which is different from the 'totals' but he has that choice. He is of course still 'blind' at this stage.

The music played, should be music of his choice, but how long is he allowed listen or can he decide himself to go from A to B to C and can he repeat or switch randomly between the three?

He can take as long as he wants. That's why some people may take half a day to do a 'test'. The 'half day' is simply because I have to pay my staff.

But in nearly 2 decades, we have not found anyone who gives consistent results in DBLTs (ie not deaf) who wants more than half a day.

Yes. The switching is completely under the control of the test subject.

I assume that for every new session out of three, A and B are randomly assigned to A,B and C , so the above text must probably be read as "Anyone who gets the same odd man in the right position(s)....." ??

Yes. Thanks for the correction

Some 30 yrs later, I was amused to find that ALL Bandwidth Limitation tests conducted since then came up with the same results. (including one with a 17kHz bandwidth)

Funny enough, a similar test was performed recently in the Netherlands with exactly the opposite outcome.
Since everybody was very upset with the outcome, the test will be redone with a new set of test candidates before publishing. It just shows that test results are not always what you expect.
Maybe you still have a link to one of these test you are referring to.

Well I'll be VERY interested in a link to your Dutch test.

The 21st century tests were listed in a thread by one of DIYaudio.com or GroupDIY moderators. IIRC, he participated in the 17kHz test which was done by one of the European broadcast organisations.

Let's be very clear. We are actually testing 2 propositions. ONE : can xxx be reliably detected on music?

If and only if, the answer to this is YES, do we ask TWO: is XXX preferred?

There are some people who can pick up a 20kHz brickwall filter on MUSIC. The number is MUCH smaller than da deaf audiophools believe. (There are test signals which make things easier but they are so unpleasant that I discount them entirely)

Of those, ie the ones who aren't deaf, the preference is ALWAYS for the band limited signal.

I should also point out that the speakers (I'm a speaker designer with some electronics too but my real expertise is using DBLTs for design) which do well in DBLTs are liked by the Pop fans, da Heavy Metal Brigade, the refined Classical Chamber music crowd, the Recording Engineer who specialises in large orchestral recordings bla bla ... as long as the choice is done blind ... so true golden pinnae aren't restricted to one class ... but its certainly VERY rare, perhaps non-existent, in Audiophools.

Just looking at distortion does not guarantee the best sound.
....
For use in preamps, I very much like the opa627 being ridiculously expensive these days. But also the lm4970, the opa1612 and the opa1642 are excellent opa's, the latter probably the best opa to be used for an MM phono preamp.
The AD797 with 0.9 nV/rtHz and a GBW of 100MHz is also a top performer, but only for very low level signals in certain situations and not at all as a replacement for the 5532.

Err.rh! LM4970 is a LED driver.

If you mean LM49710, you might like to check out a couple of threads on this website on da beast.

As for the others, AD797 is a rotten OPA for either MM or MC though it has applications elsewhere as Guru Scott Wurcer's masterpiece. OPA1612 is similar. Both have EVIL Ib cancelling

I rather think you are just looking at distortion to choose your OPA

http://www.groupdiy.com/index.php?topic=37307.80 is worth a look.

Many true gurus chime in. It proves how OPA rolling takes a VERY poor second place to correct earthing, layout & decoupling. It’s a long thread but read the whole thing from #41 to find pearls of wisdom.

There's more rationale for zillion V/us OPAs in DAC reconstruction filters but it usually means the filter topology & operating point could be improved.
_______________________

Yup bandwidth is another holy grail argument, but more argued about digital paths.

I'm actually doing matched digital filters for a fancy microphone. For various reasons, some microphones may have up to 0.5dB response differences in a sensitive audio rang between 44.1kHz, 48kHz, 88.2kHz and 96kHz versions. ie ju.uust audible to a true golden pinnae

Instead of rejecting or reworking these, we sell them to those who profess to hear differences between plain 44.1kHz and zillion GHz sampling Everyone else gets boring units with no audible differences with sampling rate. Da customer is always right
________________________

Wayne, my apologies for dragging this thread so far off topic. I think from the original point of view, we have to wait for Hans' Linear Audio article ... and hope he will deign to show us key bits of the circuit (Hint, hint)

[Hans]

-70dBm is the number I have seen in the literature most as the theoretical noise floor.
All of it is a bit of a moving target. Pops and clicks wouldn't count I guess but crackle would. I think A weighted from 20Hz - 20K Hz is a good way to measure. That is probably the most often used way. The measurement would be after the RIAA decode.

Hi Gold,

I transferred the noise from a blank trace, as already shown in a graph some postings before, to an excel spreadsheet.
In this spreadsheet I did an inverse Riaa, showing the noise as coming from the LP to be flat above 350 Hz up to 20Khz.
Also did I apply A-weighting to the Riaa processed noise curve and consequently calculated the noise from 20Hz to 20kHz.
This noise turned out to be -48dBr
With the 0dB/1kHz signal at +15dBr,
SNR for this recording is therefore -63dB, very close to what you predicted.

LP Noise.jpg (78.46 KiB) Viewed 11991 times

In the same recording, the used phono preamp with a SNR of 75dB-A ref 0.5mv@1kHz, showed a noise level of at least -15dB below surface noise.
Maybe a super quiet LP can still approach the theoretical level even better with -66db.
But even than, a 75dB-A phono-preamp, equivalent to 1.4nV/rtHz at the input, is still -12dB below surface noise.

Wayne, just having 0.33nV/rtHz at the input, produces even 12.5dB less noise, which is a fantastic result, but a bit more could easily be tolerated as proven with this measurement.

Hans

[Hans]

AD797 is a rotten OPA for either MM or MC though it has applications elsewhere as Guru Scott Wurcer's masterpiece. OPA1612 is similar. Both have EVIL Ib cancelling.
I rather think you are just looking at distortion to choose your OPA

Ha, ha, what a cheap and predictable reaction.
I agree, opa's are a disaster, and Analog Devices and Texas Instruments have no idea what they are doing.

Hans

[Gold]

At one time it was one of the 3 best small speakers in the world.

I'll be in the market for small speakers for my cabin. If your's are still made I'd like to check them out. At the NY Hi-Fi show a few years ago I heard the NSMT PSM Super Monitor and loved the way it sounded. https://nsmt-loudspeakers.com/PSM-Super-MonitorNot robust enough for daily mastering use though. I thought I'd check out his small speaker the 10S.

Any suggestions for a little speaker that is priced for mortals? I think I'm going to get this to power them. http://www.psaudio.com/products/sprout/

[JR.]

At one time it was one of the 3 best small speakers in the world.

I'll be in the market for small speakers for my cabin. If your's are still made I'd like to check them out. At the NY Hi-Fi show a few years ago I heard the NSMT PSM Super Monitor and loved the way it sounded. https://nsmt-loudspeakers.com/PSM-Super-MonitorNot robust enough for daily mastering use though. I thought I'd check out his small speaker the 10S.

Any suggestions for a little speaker that is priced for mortals? I think I'm going to get this to power them. http://www.psaudio.com/products/sprout/

again I am not Ricardo,

For making modest SPL, in a modest sized space there are lots of decent choices. The hard part is making speakers play loud and still sound good. I don't how it is these days but some studios ran their monitors way too loud.

I have a friend who does industrial design for NHT and their speakers are pretty well regarded (but not cheap). Surely with all the studios that have gone out of business there are actual studio monitors to be found on the cheap used.

JR

[Gold]

I'm not sure what Ricardo meant by small but I mean a 3" or 4" woofer. Smaller than a studio monitor. I'd be interested to check out his speaker design. The Soundfield is no slouch.

There are a lot of choices. Including the ADS speakers I've had for 25 years and like. I've heard a ton of speakers but mostly studio varieties. The PSM M10 looks interesting to me because I liked the designer's larger speaker so much.

NHT was an early adopter of the satellite and sub arrangement which in general I don't like.

[ricardo]

At one time it was one of the 3 best small speakers in the world.

I'll be in the market for small speakers for my cabin. If your's are still made I'd like to check them out.
....
Any suggestions for a little speaker that is priced for mortals? I think I'm going to get this to power them. http://www.psaudio.com/products/sprout/

Sadly this is 3 decades ago and the company has changed hands twice in that period. Wharfedale no longer design & make their own drive units and are now what we used to call "box stuffers". I'm wary of recommending 30 yr old speakers as spares are no longer available. There is one old design which some friends are trying to corner the eBay market on. This small speaker came up top in 12 DBLT series .. usually against MUCH larger & expensive speakers. No other speaker in my 2 decades of testing comes anywhere near this.

I think I can replicate some aspects of its performance if I was again R&D Manager of a large speaker maker .. but though I'm told it has me written all over it, I can't put my hand on heart and say this is why it does so well.

This Millenium, the best KEF speaker at ANY price I have heard is the KEF 'Egg' which is a re-badged version of the satellites in their KHT2005.2 AV system. I was pleased to discover they were designed by Phil Anthony who I sponsored through college. Alas, he is now Tech Manager at Martin Audio doing only PA stuff.

http://kef.com/html/en/explore/about_ke ... index.html

The new KEF LS50 at $1500 is likely to be good. A bit larger & more $$ than you specified.
http://www.stereophile.com/content/kef- ... KUbUC1O.97

John Atkinson is one of the VERY few reviewers who's ears I trust. (He's appeared on my DBLT panel)

Spares for both these should still be available for at least another decade.

I should add that my advice is ALWAYS to go and listen to speakers with your own music. Buy them if you like what you hear and ignore the Reviewers and so-called Experts including myself. It's exactly the instructions that are given to our DBLT participants.

Alas its practically impossible to do this today with shops.

[*] original KHT 2005 was bla. No idea about 2005.3 & others.