Musetec (LKS) MH-DA005 DAC


Some history: I was the OP on a four year old thread about the Chinese LKS MH-DA004 DAC. It achieved an underground buzz. The open architecture of its predecessor MH-DA003 made it the object of a lot of user mods, usually to its analog section, rolling op amps or replacing with discrete. The MH-DA004 with its new ESS chips and JFET analog section was called better then the modified older units. It has two ES9038pro DAC chips deliberately run warm, massive power supply, powered Amanero USB board, JFET section, 3 Crystek femtosecond clocks, Mundorf caps, Cardas connectors, etc., for about $1500. For this vinyl guy any reservation about ESS chips was resolved by the LKS implimentaion, but their revelation of detail was preserved, something that a listener to classic music especially appreciated. I made a list of DACs (many far more expensive) it was compared favorably to in forums. Modifications continued, now to clocks and caps. Components built to a price can be improved by costlier parts and the modifiers wrote glowingly of the SQ they achieved.

Meanwhile, during the 4 years after release of the MH-DA004, LKS (now Musetec) worked on the new MH-DA005 design, also with a pair of ES9038pro chips. This time he used more of the best components available. One torroidal transformer has silver plated copper. Also banks of super capacitors that act like batteries, solid silver hookup wire, 4 femtoclocks each costing multiples of the Crysteks, a revised Amanero board, more of the best European caps and a new partitioned case. I can't say cost NO object, but costs well beyond. A higher price, of course. Details at http://www.mu-sound.com/DA005-detail.html

The question, surely, is: How does it sound? I'm only going to answer indirectly for the moment. I thought that the MH-DA004 was to be my last DAC, or at least for a very long time. I was persuaded to part with my $$ by research, and by satisfaction with the MH-DA004. Frankly, I have been overwhelmed by the improvement; just didn't think it was possible. Fluidity, clarity, bass extension. A post to another board summed it up better than I can after listening to piano trios: "I have probably attended hundreds of classical concerts (both orchestral and chamber) in my life. I know what live sounds like in a good and bad seat and in a good and mediocre hall. All I can say is HOLY CRAP, this sounds like the real thing from a good seat in a good hall. Not an approximation of reality, but reality."

melm

Showing 14 responses by batvac2

| He must be well aware that the 005 takes 6-8 weeks to fully burn in, maybe more.

The major issues found with distortion and jitter in this unit will not be impacted by burn in..

| This is a peculiarity of all ESS chips, which cannot be heard

At minimum the continued presence of the hump represents inadequate engineering skills and testing equipment. Audio companies using this chip - including Weiss - have engineered it completely away.

There are SO many well-engineered DACs with vanishing levels of distortion and jitter and close to textbook linearity- you could spend a lifetime auditioning these. If you have bat ears that go beyond 20-bits of resolution, and available funds, then the May, Weiss, Mola Mola, Meitner and others await your appraisal.

There may be a small group of truly gifted designers who can build certain components largely by ear, but even they typically have years of previous experience with measurements and circuit design. In general, building by ear is NOT a good strategy for DACs.

 

 

 

 

 

 

 

 

| Despite Daniel Weiss' white paper regarding his digital volume control being as good or better than the best pre-amps, it has been proven to be grossly inaccurate.

Do you mean proven by experimental data or another paper, or does this refer to an uncontrolled case report of a person comparing these in their individual system?

 

So what it your hypothesis here- that a lower signal to noise ratio yields a more "musical" sound? That deviations from linearity produce a more three-dimensional soundstage? Topping probably uses a lot of negative feedback in their analog output stage, which can sometimes result in a "flatter" presentation if the "euphonic" lower order distortions are preferentially suppressed. I don't understand the point of celebrating poor measurements as some sort of badge of honor. 

My Topping D90 (original with AKM chip) sounds fantastic receiving upsampled DSD256 data from HQ player. It also sounds really good receiving upsampled 768hz data, which I believe disables the internal filter of the DAC.

 

 

True. This one bugs me because it does not appear to be a purposeful design strategy, such as when Nelson Pass or Ralph Karsten add a little more second order distortion, describe their rationale, and could re-design the component any way they want. Here, the designer admitted not having the requisite equipment to test the DAC or full awareness of the IMD issue with the ESS chip as a start. Associations that are supported by plausibility are more likely to be valid than those that are less plausible. I'm not calling for blinded listening tests all the time, but in this case would personally need to see objective listening tests or do them myself before believing that a component with accidental errors in measurement sounds better than one without these errors. 

 

 

| you seem to be wasting this time writing when you could be listening…

That was a very long post - you could have been listening to your musetec… I compared the topping d90 to my previous DAC - an Eximus DP1 - and liked the d90 more. It is going into an integrated tube amp. Of course other dac designs may sound different. Like most people I don’t have endless time to compare dacs - when the time comes, I will pick from different designs that have been reported to sound great and measure well, such as Holo, T and A, Weiss, and Meitner. And no - early solid state did NOT measure well. I don’t view the measurements of the musetec as the cause of how the unit sounds but more as a general indication of the engineering involved. 

| Chinese manufacturers produce DACs that respond perfectly to poor Amir's tests

By "Amir's tests", are you referring to fundamental tests of linearity, noise, and distortion that are staples of audio engineering? Do you think Mola Mola and Holo specifically produce DACs that "respond to Amir's tests" (they measured great)?

| In the engineering field, the analysis of the variables involved in sound reproduction is very similar to those that occur in the presence of tubulence

Are you trying to argue that perceived sound quality represents a complex interaction among many variables. Fine. In this case, a couple of those variables were found to be dysfunctional in single variable testing, Why combine them?

| Who has Topping in his audio chain is probably a person who only pays attention savings so I don't expect it to have a high-end system alas ...

Mine takes upsampled DSD256 from HQ player into an Ars Sonum integrated tube amp and Merlin speakers. I recognize this is not the highest end gear. Alas...

 

| I didn't talk about brands that produce serious DACs, which in fact are not reviewed by poor Amir, just in order not to generalize about Chinese production.

False - he has reviewed DACs from Mola Mola, Holo Audio, Chord...

| Measurements are based exclusively on a few and insufficient parameters able to describe the physical phenomenon.

For DACs, these measurements describe whether the digital signal is correctly converted into the analog waveform that it represents.

| The ONLY low cost DAC that musically produces comparable results with DACs that cost 10 times more is the Musetec MH-DA005.

OK - you win. Enjoy the piece.

 

 

 

You are taking me away from listening to music - I will respond to you but not americanspirit as he/she makes no logical sense. I have listened to enough gear to know that I like sources and speakers that are "straight up" but amplifiers that add a little cream to the coffee - either big class A or tubes. 

You're right - there were no measurements of the Ars when I purchased it - I relied on the speaker designer himself using this amp to voice the speakers and present them at shows. I am fairly confident that in my nearfield setup, moderately sized room, and easy load speakers, this 30W amp pushes out the requite 5-10 watts needed for 75-85 db listening with low distortion and a little room left over for dynamics. I recently brought home a lovely sold state Luxman integrated for comparison and got to experience all I am missing and gaining from this tube amp.

IF measurements showed this amp to have an obvious engineering flaw - frequency modulations, gross distortion at moderate power, excessive power supply noise - then I would be pissed. It bothers me that expensive gear may come with "scratches and dents." A $10,000 amp recently measured in Stereophile had a 2 db channel imbalance and other measured anomalies that were linked to a cold solder joint. 

Can audio components be designed entirely by ear- exchanging capacitors in the circuit based on the sustain of a piano in one recording or the input jacks based on one singer's voice? I don't think so. Maybe there is a rare unicorn with the requisite skills, absolute perfect pitch, and access to reference recordings. Otherwise,  engineering and measurements are important pieces to this puzzle. 

 

 

 

 

 

 

 

 

 

 

 

I spoke with Bobby a few times about his speakers - super nice guy. Apparently, massed choir was his passion. Which brings me to your point about voicing. I recently decided to learn some basics about audio, purchased a UMIK, and started making measurements- including simulated anechoic, on axis, off axis... My Merlin VSM-MXM have a clear 4-7db hump from about 700-1000 hz. With a lot of experimentation, I was able to create a reliable convolution filter that yields a smooth harman-like frequency response at the listening position (and emulates the BAM digitally so I could remove this). The filter - to my ears - is a real improvement - but the amount of improvement varies by music type. And one type of music actually seems to sound best (to me) without any correction: massed voices. As I hear it, goosing up the 700-1000hz region gives multiple singers a more "spread out" presentation. 

For this, and other reasons, I more skeptical than you about audio designers being able to reliably pull off the trick of voicing.

 

 

 

 

 

 

 

 

 

 

| my friend, why do you insist on wanting to see the measurements too? don't you trust your ears? 

We've been through this - I do research for a living. Case reports are the lowest form of evidence - they often provide false information about causal relationships. I know enough to know that my own ears have provided both reliable and unreliable evidence depending on the type of music, equipment, mood, and time of day,

On occasion, I will consider individual listening reports - always with some skepticism - IF the person listens to music for a living and is able to intelligently describe differences in sound- that is a rare individual. For the general population, I am confident playing Vegas odds that you could not reliably distinguish the Musetec in any sort of believable controlled experiment, and that any possible differences would be removed by changing the music or other gear. 

This was never about the false idea that minute differences in distortion profiles, signal-to-noise ratio, linearity, and other measurable characteristics can somehow be amalgamated into a predictable sound signature. These characteristics - to me - provide a broad sense of the fitness of the engineering and industrial design. For an amplifier, a 2 db difference in channels or separation less than 70 db may not even be audible to most people. But - for me - these signal other problems that exclude the component from audition.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Yes, but I think a little more because it is an engineered device rather than a painting. It has to produce an analog signal that bears resemblance to its input in terms of frequency, phase, linearity, timing... even if the source material is not ideally recorded. The "art" is in passing as much of the signal as possible, including any warts and blemishes. Still, no device can do this perfectly, and there seems to be no way - other than listening - to predict how small differences in each characteristic of reproduction and their interaction will combine to form perceived sound. So I am comfortable letting personal taste dictate the choice of component among those with the highest degree of fidelity. 

 

 

 

I don't see how that analogy applies. In audio, the finished recoding is all we get - a product of the musicians and engineers who put the thing together. If they were highly capable and a good match then we get lucky. If an inexperienced tech boosted the loudness by 20 db and applied a smiley faced EQ curve then we are irreversibly screwed. Do the two different strikes of the etching represent two different components trying to reproduce the identical recording? If so, then of course the "unique voicing" of one may form a particularly nice synergy with specific recordings. But I hold out that closer to reality will be more enjoyable over an entire music collection.

 

 

 

Flat earther? No- academic physician, lead a large research group, teach, and have hundreds of published papers. Afraid I have drank the Kool Aid of reality.

I agree completely with sns - tiny differences in signal to noise ratio or jitter are unlikely to be audible; there is absolutely NO reason to choose a DAC based on a sinad of 130 versus 127. That was never my point. The point was that gross errors in basic measurements - whether audible or not - are worrisome for larger design and engineering flaws. Plus it feels like buying a brand new car with a ding on it.

Yes Melm that was a good review and agree that classical is probably the best for auditioning because it is typically natural instruments in space. I am generally unmoved by “quality of the parts” arguments except maybe for the analog attenuator- would be more excited if Nelson Pass or Air Tight designed the output stage.