Darlington Labs MM-5 Review (Phono Stage)

[welsh]

Interesting. No doubt some people will enjoy that distortion, and that it should be fairly benign. It should sound like some of the common tube based circuits. If it had some pretty glowing tubes on top, it would be more interesting to me. Not sure I would want the noise with sensitive speakers though.

Thanks for chiming in and welcome to ASR!

@AdamG247 can someone get this guy a badge?

When exactly did ‘feedback-free’ become an audiophile shibboleth? I assume we are talking about global feedback?

 

[cgallery]

When exactly did ‘feedback-free’ become an audiophile shibboleth? I assume we are talking about global feedback?

Silently marks "shibboleth" on his audiosciencereview.com bingo card.

 

[krichard2496]

When exactly did ‘feedback-free’ become an audiophile shibboleth? I assume we are talking about global feedback?

From a post yesterday:
>>What we do not have is "loop feedback" whereby a portion of the output signal is returned to a summing node and there is a phase inversion.
Since this is a passive RIAA product with two gain blocks, true global feedback over the entire circuit wouldn't be appropriate anyway.
Each gain block is inverting; 2 in series preserves absolute polarity; and no loop feedback occurs in either stage, only moderate local degeneration to set gain and optimize bandwidth.

 

[KSTR]

I would think this is an excellent phono stage, after all. It measures quite well (putting the hum issue aside for a moment). Yes, it has more distortion than many would like to see but I'd say -60dB H2, effectively absent higher harmonics, a very controlled rise vs. level plus a stable distortion vs. frequency profile is really as benign a distortion mechanism as it ever gets, basically equivalent to static transfer characteristic distortion introduced digitally. 0.1% pure H2 is irrelevant.

I'm fully speculating now and might be fully wrong here... but I would assume the hum could be reduced some, probably sufficiently, with a redesign of the internal construction wrt electrostatic shielding and minimized magnet coupling. I know at these gain levels at mains frequencies it is very hard to get below the noise floor of a large FFT, PCB layout and wiring is extremely critical also... definitely not a walk in the park ;-)

 

[krichard2496]

I would think this is an excellent phono stage, after all. It measures quite well (putting the hum issue aside for a moment). Yes, it has more distortion than many would like to see but I'd say -60dB H2, effectively absent higher harmonics, a very controlled rise vs. level plus a stable distortion vs. frequency profile is really as benign a distortion mechanism as it ever gets, basically equivalent to static transfer characteristic distortion introduced digitally. 0.1% pure H2 is irrelevant.

I'm fully speculating now and might be fully wrong here... but I would assume the hum could be reduced some, probably sufficiently, with a redesign of the internal construction wrt electrostatic shielding and minimized magnet coupling. I know at these gain levels at mains frequencies it is very hard to get below the noise floor of a large FFT, PCB layout and wiring is extremely critical also... definitely not a walk in the park ;-)

We're finishing up our SU-7 (MC head amp) and optimizing these same issues. Any learnings we take away will be incorporated into future production. Electrostatic shielding is costly and the current PCBs haven already been laid out scrupulously (in-house) in attempts to minimize the noise floor. Future higher end modells, if introduced above the 7 series, may go to a larger enclosure. Everything is a cost/benefit analysis.

 

[Lttlwing16]

I have a Darlington Labs MM-6 on it's way. I also happen to have a Art DJ PREII. I'll do some A/B recordings and post them here (or a separate thread since this is a mm-5 review) in a week or so since phono preamp reviews don't include a subjective review.

 

[CausticStorm]

I have a Darlington Labs MM-6 on it's way. I also happen to have a Art DJ PREII. I'll do some A/B recordings and post them here (or a separate thread since this is a mm-5 review) in a week or so since phono preamp reviews don't include a subjective review.

Send the MM-6 to Amir for testing

 

[Lttlwing16]

Send the MM-6 to Amir for testing

Sounds like @krichard2496 probably has all the same measurements @amirm would do. Perhaps he'd be kind enough to share those in the separate thread when I post the A/B recordings. I have a Pro Tools setup and BlueCat frequency analyser plugin, so maybe I'll play around and see if there is any objective data I can glean between the two as well. Stay tuned..

 

[krichard2496]

Sounds like @krichard2496 probably has all the same measurements @amirm would do. Perhaps he'd be kind enough to share those in the separate thread when I post the A/B recordings. I have a Pro Tools setup and BlueCat frequency analyser plugin, so maybe I'll play around and see if there is any objective data I can glean between the two as well. Stay tuned..

The conventionally measured specs of the MM-6 will be fairly similar to the MM-5.
The differences between our products are potentially measureable, we don't claim they are not, but we don't know of any practical way to show them in a fully completed product. We estimate that some of the capacitor and resistor differences are on the order of -140dB to -170dB, as shown in the late Cyril Bateman's excellent set of articles in Electronics World in 2002 and 2003. Jan Didden of Linear Audio has an excellent repository.
https://linearaudio.nl/cyril-batemans-capacitor-sound-articles

As you can imagine, these differences are throughly submerged in noise. We believe the human ear is so discerning that it can hear far into the noise floor. Therefore, practical listening tests and iterative changes in the product development cycle are critical to the final product.

Our internal testing regimen is varied including a large assortment of analog test gear. We don't have a clean way to visually present it. However, as Amir noted, his measurements (basic though they are) do largely agree with our quoted specifications.

We would challenge Amir to develop a SINAD test that weights the harmonics at N^4, that is, each harmonic is raised in the summation to the 4th power. This was proposed by Shorter in 1949 and is commented on in RDH4 (1953). We believe that this would go a long way to unifying the subjectivist and objectivist assessments.

As 72 years have passed, we won't be holding our breath, however.

 

[AnalogSteph]

The mains harmonics observed here appear to be dominated by 120 Hz which generally is power supply related, so I'll be going out on a limb and suggest that ultimately both them and the relatively high distortion found are in keeping with the low NFB circuitry used. Whenever you want to use simple amplification circuitry with low PSRR, you have to ramp up your efforts on the power supply instead.

I'm guessing we're talking something similar to a Boozhound Labs JFET phonopre? I may be inclined to splurge on a cheap small-signal pnp BJT and one or two passives for each stage to get loop GBW up a bit (i.e. make a pnp "VAS" and get some voltage feedback going). Gives the JFET some input bootstrapping, improves PSRR, decreases output impedance and improves load driving.

This whole "negative feedback baaad" nonsense can just die as far as I'm concerned... it has more to do with religion than science. I am firmly with Bruno Putzeys here, who has shown fairly clearly that while a little NFB can be more problematic than none, this trend reverses above ca. 20 dB and from that point on it's the more the merrier, especially past 40 dB. As far as stability will allow, of course, which is why higher-order compensation schemes tend to be handy if bandwidth is limited. Guess how he got his Hypex and later Purifi Class D amp designs to perform as well as they are. Likewise, we wouldn't be seeing DAC output stages able to support -120 or -130 dB of distortion without tons of feedback either. Nobody says it's necessarily easy... but it's worth it.

 

[krichard2496]

The mains harmonics observed here appear to be dominated by 120 Hz which generally is power supply related, so I'll be going out on a limb and suggest that ultimately both them and the relatively high distortion found are in keeping with the low NFB circuitry used. Whenever you want to use simple amplification circuitry with low PSRR, you have to ramp up your efforts on the power supply instead.

I'm guessing we're talking something similar to a Boozhound Labs JFET phonopre? I may be inclined to splurge on a cheap small-signal pnp BJT and one or two passives for each stage to get loop GBW up a bit (i.e. make a pnp "VAS" and get some voltage feedback going). Gives the JFET some input bootstrapping, improves PSRR, decreases output impedance and improves load driving.

This whole "negative feedback baaad" nonsense can just die as far as I'm concerned... it has more to do with religion than science. I am firmly with Bruno Putzeys here, who has shown fairly clearly that while a little NFB can be more problematic than none, this trend reverses above ca. 20 dB and from that point on it's the more the merrier, especially past 40 dB. As far as stability will allow, of course, which is why higher-order compensation schemes tend to be handy if bandwidth is limited. Guess how he got his Hypex and later Purifi Class D amp designs to perform as well as they are. Likewise, we wouldn't be seeing DAC output stages able to support -120 or -130 dB of distortion without tons of feedback either. Nobody says it's necessarily easy... but it's worth it.

We have very little in common with the Boozehound design.
Also, we use a large portion of our power supply voltage as active gain range, in order to optimize headroom.
If you think about how the IDss vs VGS vary significantly across J-FETs from the same batch, even after careful selection, you may find it difficult to adjust the DC bias without changing the AC gain, and without using source bypass capacitors. That is some of what our technology is able to do. We have a different set of design goals and tradeoffs than many kit makers.

 

[Francis Vaughan]

We estimate that some of the capacitor and resistor differences are on the order of -140dB to -170dB, as shown in the late Cyril Bateman's excellent set of articles in Electronics World in 2002 and 2003.
...
As you can imagine, these differences are throughly submerged in noise. We believe the human ear is so discerning that it can hear far into the noise floor.

The trouble here is that this is mathematically impossible. Not just a matter of physics or human physiology, but from a fundamental information theory point of view this cannot be done. Ability to resolve information below the noise floor is limited by Shannon, and whilst Shannon's theory provides for arbitrary resolution below the noise, it requires corresponding time to do so. Real life music does not afford that time. Nor does the physiology of the ear.

As to addressing the subjective impost of distortion, there are some very interesting modern metrics that use models of the ear's perception that are significantly more interesting. However we need the transfer function, for which we need the complex component of the harmonic distortion. So far there seems no easy way of getting that. But given the transfer fuction we can determine the ear's ability to perceive distortion components taking into account things like masking effects, and thus give a much more accurate and meaningful weight to higher harmonics. Advancing the art really needs this step.

 

[elvisizer]

I would think this is an excellent phono stage, after all. It measures quite well (putting the hum issue aside for a moment). Yes, it has more distortion than many would like to see but I'd say -60dB H2, effectively absent higher harmonics, a very controlled rise vs. level plus a stable distortion vs. frequency profile is really as benign a distortion mechanism as it ever gets, basically equivalent to static transfer characteristic distortion introduced digitally. 0.1% pure H2 is irrelevant.

I have been using a darlington labs mp-7 for a couple of months now and haven't had any issues with mains hum coming through to the output on mine- I'm running it into a benchmark system- HPA-4-> dual AHB2's, so everything else in the chain is quiet as heck.
I'm happy with it!

 

[krichard2496]

The trouble here is that this is mathematically impossible. Not just a matter of physics or human physiology, but from a fundamental information theory point of view this cannot be done. Ability to resolve information below the noise floor is limited by Shannon, and whilst Shannon's theory provides for arbitrary resolution below the noise, it requires corresponding time to do so. Real life music does not afford that time. Nor does the physiology of the ear.

As to addressing the subjective impost of distortion, there are some very interesting modern metrics that use models of the ear's perception that are significantly more interesting. However we need the transfer function, for which we need the complex component of the harmonic distortion. So far there seems no easy way of getting that. But given the transfer fuction we can determine the ear's ability to perceive distortion components taking into account things like masking effects, and thus give a much more accurate and meaningful weight to higher harmonics. Advancing the art really needs this step.

Perfect Sound Forever was achieved with 16 bits and 44.1kHz sampling by Philips in 1982.

 

[Postkrunk]

This review is a pleasant surprise. I expect this preamp has a very short impulse and step responses (without the equalization circuit, of course). It would be interesting also to look into IMD distortion spectrum.

A few years ago, I built for myself a couple of copies of Creek OBH-8SE preamp, which also has zero global NFB. I greatly prefer its sound to basic opamp-based designs. I was wondering, whether it due to pleasant 2nd harmonic, or because of another intrinsic qualities of non-NFB circuit. Now I tend to the second version.

It is nice to see that there are people in the industry with their own creative approach, and that they also understand the importance of overload capability, unlike the others. *cough* Mani *cough*

 

[Doodski]

has zero global NFB. I greatly prefer its sound to basic opamp-based designs. I was wondering, whether it due to pleasant 2nd harmonic, or because of another intrinsic qualities of non-NFB circuit. Now I tend to the second version.

When I studied op amp circuitry we where shown that op amp negative feedback in audio circuitry is not a bad thing with op amp circuitry. Is there something I missed regarding op amps' negative feedback being a detriment to audio fidelity? I figured that we covered op amps pretty well.

 

[krichard2496]

When I studied op amp circuitry we where shown that op amp negative feedback in audio circuitry is not a bad thing with op amp circuitry. Is there something I missed regarding op amps' negative feedback being a detriment to audio fidelity? I figured that we covered op amps pretty well.

From our FAQ:
Barrie Gilbert, a (recently passed) top-flight engineer with more than 60 years’ experience including at Tektronix and Analog Devices, pointed out that:

“…there is significant odd harmonic distortion even at frequencies well below the unity-gain crossover; and that the phase angle is not only a function of frequency, but also of amplitude -- that is, the op amp generates a peculiar kind of amplitude-to-phase modulation, something which one should not expect of a linear system.

But, then, an op amp is by no means as pristine in this respect as the textbooks suggest, and that gm cell has very strong open-loop distortion for even quite small inputs.” (Barrie Gilbert, "Are Op-amps Really Linear?", Electronic Design magazine, June 10, 1998, available online at https://linearaudio.nl/sites/linearaudio.net/files/Are Op Amps Really Linear.pdf

 

[Francis Vaughan]

Perfect Sound Forever was achieved with 16 bits and 44.1kHz sampling by Philips in 1982.

No is wasn't. Really, quoting this demeans you and your arguments and achievements. Shannon's theory underpins the entire communications industry, and is pivotal to the entire arena of information theory. Trying to say you somehow know better is little different to the arguments of flat earthers.

Have a read: These are some of my favourite papers of all time. Anyone who wants to have any clue about the fundamental limits of communication should be familiar with them.

https://webusers.imj-prg.fr/~antoine.chambert-loir/enseignement/2020-21/shannon/shannon1949.pdf
http://people.math.harvard.edu/~ctm/home/text/others/shannon/entropy/entropy.pdf

If there was a Nobel Prize for engineering or mathematics, Shannon would have got it.

 

[krichard2496]

No is wasn't. Really, quoting this demeans you and your arguments and achievements. Shannon's theory underpins the entire communications industry, and is pivotal to the entire arena of information theory. Trying to say you somehow know better is little different to the arguments of flat earthers.

Have a read: These are some of my favourite papers of all time. Anyone who wants to have any clue about the fundamental limits of communication should be familiar with them.

https://webusers.imj-prg.fr/~antoine.chambert-loir/enseignement/2020-21/shannon/shannon1949.pdf
http://people.math.harvard.edu/~ctm/home/text/others/shannon/entropy/entropy.pdf

If there was a Nobel Prize for engineering or mathematics, Shannon would have got it.

We're not in the business of building electronics for the telecommunications industry.
Why would you want to bother with reproduction from a medium with THD and IMD in the many % to begin with? Dragging a rock through vinyl? Seems distinctly old-fashioned. Along with all that vacuum tube gear.
I'm sure the marketplace will have it's say.
Huh.
I wonder what the resale value is of a Dynaco Stereo 70 tube amp (1959) vs. it's replacement, the Dynaco Stereo 120 (1966) transistorized amp?

 

[Bob from Florida]

No is wasn't. Really, quoting this demeans you and your arguments and achievements. Shannon's theory underpins the entire communications industry, and is pivotal to the entire arena of information theory. Trying to say you somehow know better is little different to the arguments of flat earthers.

Have a read: These are some of my favourite papers of all time. Anyone who wants to have any clue about the fundamental limits of communication should be familiar with them.

https://webusers.imj-prg.fr/~antoine.chambert-loir/enseignement/2020-21/shannon/shannon1949.pdf
http://people.math.harvard.edu/~ctm/home/text/others/shannon/entropy/entropy.pdf

If there was a Nobel Prize for engineering or mathematics, Shannon would have got it.

I took a quick look at the 2 pdf links posted. I am puzzled by the choice of topic. Please explain how this relates to a phono preamplifier.