April 2006


Mhdt Laboratory
Paradisea and Constantine DACs
A tale of two DAC units.
Review By Joe Audiophile (Scott Faller)
Click here to e-mail reviewer
I've got no doubt that you are asking yourself, who the
heck is Mhdt Labs? Well, lets see if I can help fill in that gap.
Not too long ago one of our readers dropped me a note asking about some
different DACs that were available on the market. He was in the market for a
good sounding, affordable tubed DAC (aren't we all). We talked about the usual
suspects. I brought up the Hagerman Chime, the Scott Nixon tubed DAC, the Audio
Note DAC Kit 1.1, the Musical Fidelity TriVista and a few others. I even brought
up taking a vintage California Audio Labs DAC and sending it to somebody like
John Hillig at Musical Concepts and having it modified. One of the local guys
uses a CAL Alpha modified by John and it's stunning compared to my stock CAL
DAC.
At any rate, the reader wasn't overly interested in a kit which is by far the
most affordable. The other DACs we talked about were getting little on the
pricey end of the spectrum. As I thought and searched, I didn't come up with a
thing to help him out. Then one Saturday I called my mate Geoff over in France.
Geoff knows all things cheap and great sounding. He told me to do an eBay search
for a DAC. He mentioned that when I got all the hits, take a look at the ones
from Taiwan. He mentioned a certain builder made one heck of a good sounding DAC
for not much money (Geoff couldn't remember the exact name).
Sure
enough, I did a search at eBay and found a guy offering several different
designs, including a tubed DAC for not much money, relatively speaking. I looked
at his feedback and noticed he had hundreds of sales without a single neutral or
negative feedback. The manufacturer was Mhdt Labs. So I decided to drop them a
note and see if he was up for a review. Of course, they jumped at the chance.
Mhdt Labs offers four different offerings in their line of DAC's. Two of the
DAC's are standard solid state and the other DAC's utilizes a tube as a buffered
output stage. I asked Jiun-Hsien Wu, Chief Designer for Mhdt Labs if he would be
interested in sending one of his solid state DAC's along with the tubed DAC.
Sure enough he agreed and sent the Constantine (solid state) and the Paradisea
(tubed).
I've stated over and over previously, that I don't know anything about DAC
designs, so don't get the idea that I'm going to try to impress you with digital
drivel. One thing I do know about is the quality of parts and the
importance of a good layout. That is what initially drew me to these units. I'll
pepper the design section with bits I've read on the net but I can't attest to
their accuracy. In the end, I can tell you exactly how these two little gems
sound. That you can take to the bank.
Basic Design
Let me start by recapping what Jiun-Hsien Wu of Mhdt Labs states about their
DAC units. Both of the units on hand utilize the Philips TDA 1545A
non-oversampling (henceforth know as NOS) DAC chip. The TDA 1545A has 16 bit
resolution. The units have a dual voltage power supply transformer that can
accept input voltages or 110 or 220 at either 60 or 50 Hertz. Both DAC's can
handle Toslink, RCA coax or BNC digital inputs. The printed circuit boards are
fiberglass FR4 94V grade, which meet the UL-94V-0 standard (as I'm told). This
is more than ample considering the B+ rails of the digital section is running
under 15 Volts (see later text for the tube buffer operating voltages).
The
cases are constructed from a 5mm smoked, transparent acrylic. The cases are
etched with the Mhdt Labs logo and the name of the DAC. In the case of the
Constantine (the tubed DAC), Jiun-Hsien Wu has laser cut ventilation holes in
the top of the casing to allow the heat buildup from the tube to escape. On both
units, the left and right sides are open to allow ‘flow through' ventilation.
You might be concerned about the fact that these are ‘handmade' pieces of
gear. I'm here to tell you that the fit and finish on both of these DAC's are
stellar. You'd swear that these were production pieces rather than something
hand cut, drilled and assembled. They are beautifully made.
Mhdt Labs has their transformers custom wound for both DAC's. From top to
bottom, these DAC's are using quality passive parts. You'll find cottage
industry names like Rubycon (ZA and ZL series) capacitors, Sanyo OSCON, Elna and
Silmic capacitors used throughout. Mhdt private labels their own metalized
polypropylene capacitors used in the signal path. There is also a combination of
Dale RN series, carbon composition and other metal film resistors used
throughout the DAC's.
The active components used in both Mhdt DAC's are the NOS 1545 DAC. The
Constantine uses the Analog Devices 826 opamp and the Paradisea uses the Burr
Brown OPA2604.
As I understand from talking to someone that knows more about digital design
that I do, Mhdt Labs has chosen the same route that Audio Note does with their
DAC's, they don't incorporate the ‘brick wall filter' in their designs. There
are people within the industry who feel that this is ‘improper' design, but
then there are people who design by listening and deciding what sounds best.
Mhdt is in the latter camp. I'm not saying who is correct, just letting you know
there are two different philosophies out there.
The Paradisea's Tube Buffer
The
Paradisea DAC uses a single 5670 vacuum tube. This tube is directly coupled to
the output of the Burr Brown OPA2604 opamp. Mhdt Labs is not using a coupling
cap, resistor or transformer between the opamp and the grid of the 5670 vacuum
tube. In other words, the 5670 tube is directly coupled to the opamp using
(essentially) a modified Loftin-White design. Mhdt has chosen a slightly larger
output coupling cap for the Paradisea, a 1.5uF as opposed to the 1.0uF used in
the Constantine.
The Paradisea can take various equivalents like the 2C51, Western Electric
396A, 6N3Pi (6H3pi), or the 6CC42. The great part, the most expensive of these
tubes is the Western Electric. At last check on eBay, the WE396 was selling for
under $20 bucks. The vast majority of these tubes will run you between $4 and
$6. That means you can pick up scads of different flavor tubes and still have
plenty of money left over for pizza and beers.
Jiun-Hsien Wu has added a separate transformer and Pi filter to handle the
load and regulation of the tubed buffer stage. Mhdt Labs has chosen to run the
tube at a very low voltage. Jiun-Hsien Wu picked 19 Volts as the B+ operating
point for the tube buffer. This does a couple of things. First it extends
(dramatically) the life of your tube. Second, it makes the design and size of
the DAC much more manageable since you don't need a big honkin' transformer and
high voltage caps for the Pi filters power supply.
Just a quick word about the buffer stage for those not familiar. A pair of
nice sounding opamps, the Burr Brown OPA2604's, handles the gain stage on the
downstream side of the internal Philips 1545 DAC. Just downstream of the output
opamps resides the 5670 vacuum tube. This tube adds a nominal amount of gain to
the output stage. The tube acts as a buffer, which gives the DAC the tube
qualities that many of us have come to know and love. In other words, the design
of the Paradisea is not a true tubed gain stage, it is merely a tube buffered
output with the critical gain stage being handled by a pair of quality opamps.
A Non-Statement About Digital Sampling Rates
In the realm of digital to analog conversion, there are hugely differing
opinions as to which method of conversion sounds the best. We have
non-oversampling (as is the case with the Mhdt Labs DAC's), we have oversampling
and we have upsampling (notice I left out SACD as I'm only speaking of Redbook).
The camps out there will each defend their chosen sampling procedures to the
death (if need be). I've found that each of them can sound quite good when done
well or quite bad when done poorly. In my case, I've got the AH! Njoe Tjoeb
which allows me to use the stock output of the Philips TDA 1545 along with the
OPA2604 (or the AD 826) opamp or I can drop in the 24/192 upsampling board if I
so desire.
As for my preference in sampling rates, I like them both, but I vacillate. I'll
go for a while listening to the stock, NOS chip and then I'll get a wild hair
and install the upsampling board and listen to it for a few months. Then I get
bored and yank the board back out for something different. As you can tell, I
really don't have a preference. I sort of sway with the prevailing winds. I just
like the breeze.
When it comes to the sonic differences between the two sampling rates,
hopefully I can explain them without causing some sort of flame war or getting
too many email bombs. The upsampled units that I've heard tend to exude a large
amount of (apparent) detail. I said ‘apparent' because of the mathematic
interpolation an upsampler does. The upsampler samples the information on a CD
and then ‘approximates' the additional detail through a complicated algorithm.
Upsampling (as I understand it) stretches the data points apart (of sorts) and
fills in the gaps with approximated data, smoothing the data stream curve. The
end result is, the upsampled signal that comes out of your speakers supposedly
has higher resolution. You hear more detail, there is a greater breathiness to
the sound on your CD. As I stated earlier, done well, upsampling can sound very
good. Done poorly, it can sound like cats mating. Upsampling can also bring the
soundstage quite a bit more forward into your room. And yes, just ‘average'
upsampling can be quite harsh, fatiguing and can have a definite digital sheen
to it.
In a back to back comparison between the two, non-oversampling can sound a
bit dull and lacking in detail. But, after your ears become accustomed to the
non-oversampled chip, you begin to understand that the presentation becomes far
more relaxed and less aggressive. I hate to use this term but it sounds less
forced. In essence, less digital sounding. Even though it isn't vinyl, it
contains a few more of its qualities (IMO) than its higher resolution cousin,
upsampling.
I guess what I'm getting at is the differing sampling rates are ultimately a
personal preference. Each can sound quite good when done properly. If you want
some seriously technical information regarding sampling rates, do a
Google search using "upsampling vs. oversampling" as the search parameter. You'll
get enough reading material to keep you going for weeks on end.
So Lets Talk About How They Sound
The vast majority of my review of the MHdt Labs DAC's were done using the
Bolder Cable modified Squeezebox and the Rev 1 Power Supply. I used my DeZorel
mains conditioners in front of all of my gear. Since the Constantine and the
Paradisea are essentially the same design with the only major difference being
the tubed buffer stage of the Paradisea, they sound quite similar. The
difference being the restored dynamics and the naturalness the tube provides.
When directly comparing the two, the Constantine comes across with a light,
delicate touch to the music providing loads for detail without coming across as
being aggressive or overbearing. The Constantine is quite dynamic and paints a
sharper sonic picture. The Constantine provides slightly better imaging. The
sonic musicians are just a tad better defined in the holographic space we all
love so much. As I listen I find the bass and midbass response of the
Constantine slightly less pronounced when compared to the Paradisea. I suspect
it has to due with the tube buffer stage of the Paradisea. In turn, the bass of
the Paradisea has more weight and impact which may or may not be what you are
liking for in your DAC. The bass of the Paradisea is quite firm and
authoritative when you pick the right tube (more on tube choices later).
Comparing the two (at least in my system and with my heavy preferences for
tubes), the Constantine can seem a bit thin in the bass and midrange. Don't read
that as being a negative attribute because it isn't, it's just my preferences
coming out. In fact, another hardcore SET and Lowther I know (and you all know
too) and respect fan thinks the Constantine sounds more ‘real' than the
Paradisea. Again, it all comes down to personal preference.
Keeping my preferences in mind, I find the Paradisea has a more accurate
timbre than the Constantine. The tube restores some of the dynamics to the
original compositions. Though I'd prefer a true tubed gain stage as the output
of a DAC, the combination of a quality opamp and the tubed buffer stage works in
concert extremely well. The Paradisea gives up a little definition and polish
for more authority and dynamics. At the same time the Paradisea gives you a very
slightly wider and deeper soundstage.
Guessing at your personal tastes here, if you are a detail freak, the
Constantine is the logical choice for you. It does a great job. Its very well
suited to you solid state guys out there. The Paradisea falls right into my
musical sweet spot. The authority and weight it projects is a perfect match to
my system. It isn't the least bit fatiguing. To me in my system, the Paradisea
sounds more like real life. Don't misunderstand, they both have their faults but
in the same vein, they both sound extremely good.
The great part is neither of these units show any hints of veiling the music
in the least. They both are some of the cleanest and most accurate NOS DAC's I've
had the pleasure of listening to. Granted, pick the wrong tube for the Paradisea
and you'll hear some sonic goo, but the tubes are cheap and very readily
available. It's more than obvious to me that Mhdt Labs listened first and then
looked at the bottom line last.
Roll Your Own
As I've mentioned, the 5670 tube and its variants are cheap and readily
available. To date I've found 6 different variations of the 2C51 tube. Each one
has its own unique flavor. The Paradisea ships with a new old stock JAN GE tube.
It sounds just fine. As I started collecting different brands I found that in my
system the black plate Bendix sounded the best, even better than the Tungsol's I
picked up. The bass was nice and firm and the treble was crisp, clean and
extended.
I haven't gone to the point of searching out every variant out there and
giving them a listen. The owner of Mhdt Labs seems to like the sound Western
Electric 396A. One of these days I'll run across one other than on eBay. Last
thing I want is to get into a bidding war and pay more than the tube is worth.
Lets Talk About Some Music
In my time I've spent with both of these fine sounding DAC units, one of the
songs I was most taken by was Eva Cassidy's "Fields of Gold" (the
Sting classic) from the Songbird CD. Every time I put this CD on a good CD
player, I usually end up having to listen at lower levels because Eva's voice
gets butchered by jitter and God knows what else. I have to say that the
combination of my Bolder Modified Squeezebox and the MHdt DAC has made this CD
sound better than I've ever experienced before.
Eva actually takes on weight and proportions. The notes from her guitar leap
from my Lowthers. I'm hearing layer upon layer of detail just as it was
recorded. Her voice has that chestiness that drew me to her initially. There is
ZERO hint of grain or edginess coming the Paradisea. The highs are nice and
extended. I can also plainly hear the gentle tape hiss from the analog masters.
Listening to Lyle Lovett and his Large Band doing I Know You Know, these guys
absolutely come to life with the Paradisea. From the gentle touch of the brushes
swirling around the drums to the ever present strumming of the hollow bodied
rhythm guitar coming from the left channel, the sounds are deeply dynamic and
extremely well defined. Lyles voice carries the presence not often heard from
lesser gear. He comes to life with the help of the Mhdt Labs DAC units.
Moving onto something a bit livelier, Robben Ford's Up The Line has
Robben matching lead guitar notes with a baritone sax in portions of the song.
The sax has this big phat sound that is just nearly perfect. The overlaid
instruments are easily distinguishable. The Paradisea and the Constantine don't
cover up any of these complex music passages.
Next up is easily one of the most dynamic CD's that I own. Ronnie Earl's Still
River on the Audioquest label. It's now long out of print. Listening
specifically for that ultra cool Hammond B-3's intonation, each of the DAC units
do a fabulous job reproducing its sound. The B-3 comes through on both units as
the warm, wonderful organ that it truly is.
Side Story
Funny, not long I was in an antique shop and I was asking the owner if she
had any old radios or vacuum tubes in the shop. She said no but then stopped me
and said she had an old organ in the back room. She mentioned she couldn't
figure out how to turn it on. She said it made noise like it was running but
they couldn't get any sound out of it.
When I looked at the organ I saw it was a Hammond. I spun it around to see
which model it was I saw the tag that stated it was a M-3. Not quite the caliper
of the B-3 but a great sounding organ nonetheless. I reached under the keyboard
and flipped the on switch and then pushed the start motor for the speaker. Once
the oscillator got up to speed I engaged the drive and she was playin' music.
Not anything recognizable since I was at the helm but some really cool sounds
were coming out of this vintage organ. I showed the owner how to get the organ
running and talked to her for a little while her mom played some gospel music
and hymns and then I set myself down the road again.
Sorry, I got sidetracked
In The End...
The sounds the Paradisea and Constantine produce are nothing short of lovely.
So few non-oversampling DAC's get the music right top to bottom and this happens
to be one of those occasions. Its not very often I get a piece of gear in the
old homestead that can make my nipples stand up at attention but this one does
it (I know, that was probably too much information but you guys know me by now).
Truly, the MHdt Labs Constantine and Paradisea DACs have made me forget that I'm
listening to gear. When I listen, all I hear is music, magnificent and pure just
as it should be.
When it comes to which one is best for you, that could be a really tough
choice. Both are truly excellent sounding units that will absolutely hold
their own with the heavy hitting non-oversampling DACs out there. If you are a
solid-state guy looking to see what tubes might do for your system, the
Paradisea is a great choice. The tubes are cheap and easy to find. I really
think you'll like what you hear. If you are a tube guy at heart, the choice is
obvious, you'll love the Paradisea.
If you are a detail freak or don't care for the sound of tubes, the
Constantine should be best suited for your system. It's clean, accurate and
doesn't come close to hinting at listeners fatigue. If you're a tube guy and
your system is just a bit too ‘tubey,' take a hard look at the Constantine. It
does a great job and should tighten up your digital front end quite nicely.
Both of these DAC units are truly soul stirring... and the best part,
they're cheap compared to much of the competition. Are you guys looking for a
giant killer? These little David's are hiding in the bushes if you are willing
to look outside the traditional purchasing paths. Will either of these
miraculously transform a crappy recording and make it sound like something from
Sheffield Labs? Not hardly. Will this little pair of DACs make those marginal
recording sound considerably better? Absolutely. They will also take all of your
favorite audiophile recordings and it will breathe new life into them.
Keep in mind, as it is with all small manufacturers, Mhdt Labs crank gear out
as quickly as time affords. They don't have huge factories with dozens of
assemblers and automated soldering lines. The small manufacturers usually have a
backlog that may take a few weeks (or longer) to plow through before the piece
you ordered gets built so try to be patient.
Both of these pieces come with my highest recommendation based upon their
tremendous sound and their miniscule price tag. And don't think for a minute
that these can't compete with the big boys out there, because they can, easily.
My Ratings
Please keep in mind this rating system is used to compare the Mhdt Labs DAC
units against absolute perfection, or cost-no-object DAC design. If you see what
you think may be a low(ish) score, it's because there are DAC designs that are
even more refined but consequently cost considerably more. To top that off, if I
assign 5's across the board, I've just painted myself into a corner leaving no
room for that ‘ultimate' DAC. You won't see me handing out many 5's. In turn,
I feel I need to leave room in the ratings system to accommodate those DAC units.