DIYHiFi.org

Diyparadise.com's Monica2 NOS DAC, shown via schematic above, has gained somewhat of a cult following. It seems to be popular with some DIYers and has received praise from apparently "respectable" sources:

http://www.tnt-audio.com/sorgenti/monica2_e.html
http://www.positive-feedback.com/Issue27/monica_dac.htm
http://www.6moons.com/audioreviews/road ... our3b.html
…etc.

In late December of last year, a time when I was very new to DIY DACs (and probably didn't know any better), I decided to try out the Monica2. It was ~$90USD + $17USD for S/H from Malaysia. Since roughly 60% of my listening time is on-the-trail, the Monica's small, portable size seemed very appealing. Plus, if this little DAC was at least half as good as users and reviewers claimed, then I'd have a nice little home DAC, too. Or so I thought.

I built the unit, per kit instructions, and added several Wima bypass caps (they were not included but the PCB had holes for them). In addition, I created a hefty 12VDC battery PS, with capacitors to overcome some of battery-based PSU's current-supplying limitations. I also have a well-filtered and well-regulated STEPS PSU, based on tangentsoft.net's design.

I let the Monica2 unit break in for 100+ hours.

Folks…I don't know what the reviewers and DIYers, mentioned above, are listening to, but the Monica2 sounds *very unimpressive*. Grainy treble, weak bass, lack of macro-/micro-dynamics. Could it be that that's all one should expect from a $100 device? I don't think so!

Shortly before experimenting with the Monica, I modded a Toshiba $40USD SD-3990 CD/DVD player with another $40 upgrade parts (rope caulk for mech damping xport, PSU parts, AD8620 op-amp; the Toshiba's have been popular on diyaudio , head-fi, etc. as units that once properly modded, sound excellent wrt price/performance). It uses an all-in-one audio/video Zoran DSP which features a 24bit/192 2-channel audio DAC. This modded $80USD player sounds very good (= *way* better than the Monica2). In fact, the Toshiba sounds better than the stock Musical Fidelity A324 I've posted on elsewhere in this and other forums. The noted quality is all-the-more dramatic as I pair the MF A324 with the Toshiba as a transport. And most of the Toshiba mods -- mech damping, PS upgrades -- affect its transport qualities.

Maybe the average budget-conscious DIYer doesn't have access to several modern DACs for cross-comparative evaluations. So they may not know any better. But surely, the folks at positive-feedback, tnt-audio and 6moons have had some decent-sounding designs go thru their door. So why the glowing reviews for Monica???

Questions/Comments:

Did I build the Monica2 incorrectly? Were one/several parts "defective"?
I doubt it. I double-checked most component values and connections with a Fluke 87V Industrial True RMS Multimeter (that all I have -- no 'scope). I have no way of truly gauging whether the asynch reclocker or CS8412 are defective. I doubt there's anything wrong with them as their voltages check out OK. As for the TDA1545A…I ordered six extra pieces with Monica2 ( I was going to parallel them). Anyway, thinking one may have been defective, I swapped one with another and another. No difference. I am almost 100% certain that I have a normally-functioning stock Monica2. And the sound it is reproducing -- which I don't like -- is normal for this unit.

Based on my comments and schematic topology below, what do you think may be causing its "characteristic" sound?

Could it be…

(1) Lack of output buffer (e.g. I/V, either opamp-based or discrete)? Stock unit has none.
(2) 80Mhz oscillator. Yeo, the proprietor of diyparadise, claims he dropped in several oscillators, evaluating by ear. And 80Mhz was best. His former fave was 50MHz.
(3) Philips TDA1545A. Not much about this DAC on any forum. Not sure that's a good or bad thing.
(4) Use of asynch reclocker.
(5) other...

If this DAC can be made to sound better, it'd be nice to have the whole package retain its portable dimensions. Here's my current setup, using a small Serpac enclosure:



---NOS rant below-----

A note on NOS's "characteristic" (?) sound (my opinions based on NOS DACs I've heard at audio club meetings, audio "shows", and Monica2): there is something "holographic" about NOS sound. It does make certain vocal sonics sound "special". I'll give NOSers the benefit of the doubt and say NOS has an "edge" here -- but it's a nebulous, blunt edge IMO. IAC, all NOS DACs, in my limited experience anyway, failed to sound good in every other sonic "parameter": micro/macro dynamics, bass, treble, tight imaging, grain-free sound, etc.

NOS and "over-built" back ends:

I know I'll get slammed for this by NOS proponents but having a SUPERB back end (I/V; analog section) on a NOS DAC doesn't count!

But that's just what diyparadise seems to have done with their Grounded Grid back end for Monica2. To me, this is a ridiculous "solution" if *only* judged based on the size, functionality and environment-used differential: a small, "portable", battery-powered DAC paired up with a boat-anchor back end!

Someone from LessLoss.com once noted: "It is *harder* to get rid of all jitter [and other digital-end artifacts] than to put on some tubes and smear the digital grain out of the signal. It is true that digital sounds harsh when there is jitter [and other digital-end artifacts]." Is this what diyparadise has done? I personally don't care for vacuum-tube sound in and of itself, but one can also smear with solid-state.

As noted here I, TTBOMK, believe that having a good DAC chip (1704, TDA1541A double crown, etc.), well-executed oversampling (my preference), low-jitter clock, well-designed SPDIF input, etc. are only pieces of a larger, aggregate and symbiotic whole. So while (IMO) NOS is mathematically-compromised topology, a designer can "get away" with it by "souping-up"/tweaking/optimizing these other pieces.

Why I feel oversampling is best-sounding/superior:

Empirical, statistical evidence ("data") based on…

- core-component trends/product offerings (DAC ICs, DSPs, etc.) manufacturer product offerings (most modern DACs and DSPs have multi-bit OS filters, although some are delta-sigma)
- audiophile component-manufacturer product offerings (majority offer/prefer multibit OS)
- DIY build preference (NOS definitely have strong-voiced supporters but more DIYers build/prefer(?) multibit OS than NOS)

Hollow,

Got me... But really the TDA1545 is not capable of being used with passive I/V. This is one thing I found out looking for a real nice small current dac. I was trying to use it with a 0.5VA 6V battery regulated to 5v. But I just could not get the thing to really sing. I added an active opamp i/v and it really took off. But not being into opamps it then fell apart.

I am also not really keen on async reclocking. In my studies using the Dr. Jordan Analyzer, I found the jitter, noise and distortion to increase using this method.

I have not been doing any SYNC reclocking but I see Guido and Jos do this to lower jitter before the dac. Anyone have any comments on that?

Thanks
Gordon

Hollow... man...
Jut looking at that schematic got me scared.
How can that thing even aspire to sound decent?
The most serious mistake there is the (unbuffered) passive I/V, with that dac chip.
On the whole, that thing will sound like a $1 walkman.
I don't get surprized anymore from reading rave reviews from stuff like this. You can find everything on the net...
Btw in the right context it can sound half decent (suppose you have it VERY near an integrated amp or preamp, with an input buffer). But it will always suffer from more than desirable treble rolloff.

Yeah, it has all the errors one can make :

- passive unbufferred IV on a DAC chip that hates that
- passive filter using crap little inductors
- electrolytics in the signal path
- crappy power supplies (7805 ?? WTF ?)
- split ground plane introducing massive noise on clock and data lines
- unspecified oscillator
- etc

I don't need to hear it to tell you that practically anything included a $20 korean divx player would sound better than Monica. Monica's problem is (like very often) designer stupidity.

TDA1545A isn't the best DAC there is, but when properly used (like in my own DAC) it is still capable of sounding very good.

You were gouged.

But at least, I'd say this was't that expensive for a learning experience... You could have bought a much more expensive kit and ended up with the same impressions.

> If this DAC can be made to sound better

NO. You can salvage the TDA1545A and use it in another project, but that's about it.

> Grounded Grid back end for Monica2

This is called polishing a turd. Much better to have something good in the first place !

I used TDA1545 NOS, but now I use 2x oversampling (done in the PC) and it does sound better. So much for NOS.
Many people on this forum (including me) can take a piece of paper and draw you a DAC that will sound very good from the first try, it's just about using the known good practices and avoiding the common errors. Now, going from very good to excellent needs a lot more work than just using recipes !

If you want portable goodness, here is what I'd advise :

- take a discman with an anti-shock buffer that does not use lossy compression to extend its little RAM
- rip it open, add a clock input and I2S output
- create a DAC board with master clock, reclocking, a good DAC chip (TDA1545 is OK, a lot of others are better), folded cascode IV (ala Jocko) followed by filtering, buffer, LDR volume control, and a good output buffer suitable for headphones.

If you use SMD's, it can be smaller than Monica, and sound excellent.

I wonder if the lack of a decoupling cap at pin 22 (VA+) on the CS8412 is a schematic error, or if it's really missing.
Also, those standard datasheet component values on the FILT pin are easily improvable.

Hi peufeu,
how do you do 2 x oversampling in the PC?
Something like libsamplerate for oversampling and then brutefir for the interpolating filter?
And, if you did so, which kind of filter?

maybe too much questions?

ciao
andrea

It's rather brutal, I pipe the output of Audacious into sox (via an ALSA FIFO).
Some day I'll try something cleaner, but eh, it works.

Very interesting! Any further thoughts?
I see there´s a thread at the other place about this.

Regards!
Alexandre

peufeu: thx for your excellent, detailed and EDUCATIONAL feedback...

Two questions:

What is (are) your favorite DAC(s) chip(s)?

Where is your own DAC -- plans, schematics, etc? Are they online somewhere?

At the risk of putting you on-the-spot -- ;) -- and since I'm fairly new at this game, can you point to something "VG" or "excellent"? Perhaps it's your own DAC you were referring to above.


What you describe seems do'able, once I have mastered (and have gained experience from) some of the "lower-level" projects I'm working on.

I have a few a late-model Panasonic and Philips PCDPs; the Philips model has defeatable anti-shock and it seems to have better-quality stock parts, so that's a model I'll probably use.

Actually, for portable use, I'm currently using a slightly-modded portable Philips PET724 CD/DVD player with a DIY PIMETA headphone amp. I'm really pleased with the how well the PET724 works in this application using a backpack to carry it -- doesn't skip and seems to be holding up (durability-wise) after over 6mos of daily 4-mile walks/hikes.


How would you improve the FLIT values?

Pin 22 is tied to pin 23. One side of both pins go to 1mH and then LM7805; the other side goes to (+) side of 47uF cap, the (-) side of which goes to gnd.

BTW: Looking at the schematic above, I am just noticing some of it does not correlate to the actual PCB/circuit or is just plain sloppily-rendered (and, hence, hard/confusing to read). Arrg! I hacked that schematic together from two disparate schematics on diyparadise.com thinking I had assembled the whole circuit. The "designer" from diyparadise initially told me he only has Mac and MSWord to work with, so schematics are crude (true!) and hard to draw and that's why no up-to-date version exists. Recently, the designer was informed of the free version of Eagle (see: http://diyparadise.com/yabb/YaBB.pl?num=1179438189). Now the designer's excuse is that too many people are "cloning" Monica (ripping off the design), so no schematic will be provided to anyone. Sigh. Anyway, here is the front and back of the complete DAC board:

In my AA dac I'm using 390R + 1.5uF MKT, and a small 3.3nF multilayer ceramic directly soldered across FILT and AGND pins.
Btw check that Monica schematic above again... the filter must go to AGND, not DGND!




Really... what's in there that is worth to be copied?
I only see bad 'ideas'...

I haven't compared really... I plan to do a shootout some day.



It's in bits and pieces on my old and new site, it's pretty old.

The very old pages :
Reclocking : http://peufeu.free.fr/audio/extremist_d ... ent_3.html
Supplies : http://peufeu.free.fr/audio/extremist_d ... ent_2.html
Current IV : http://audio.peufeu.com/files/IVsch-02.png

New ethernet interface : http://audio.peufeu.com/



I wouldn't say it's excellent (I've heard better) but it's quite good (better than my speakers anyway, and better than most DACs I heard). I suspect the TDA1545 to be the first thing I should change.

I am not going to "point out", merely say that simply applying the known-good recipes that many members here know will provide you a very good starting base with much potential (and also, very important : no wasted potential). Also since you say you don't have experience it would be safer and provide you more "security" and chance to end up with simething your like. And satisfaction !

I mean, you can :

* pick a DAC chip, there I have no advice to give. Mister Cuffioli did some pretty extensive tests I think, seems OK. Trusting comments on DAC chips from people who have tried to implement them (esp. on "the other site") is a bad idea, since most will build something like your Monica, and then conclude they don't like the chip, while in fact their implementation sucks.
Hints to detect untrustable sources :
- no ground plane
- no mention of jitter
- no reclocking
- uses passive IV without doing distortion tests to check if the dac's output overloads or not
- really crummy layout
- sending high speed digital signals through flying wires
- split ground plane like in yout Monica (split planes can help, but NOT THIS WAY)
- listens to the thing on a fullranger from 1912 treated with C37 lacquer and only plays Pierre Boulez on it.
- the guy raves on something that you know sucks (example : CD723, CD63SE man did that one make my ears bleed, etc)

Now you have picked your DAC chip (always a shot in the dark !), use a version of Jocko's IV (like I did) ; I liked it a lot better than all the other variants I tried (opamps... see my old website, link above). Folded cascode IV sounds nice, easy to design.

- steal my reclocking (I don't remember who I stole it from hehe)

- use a low jitter clock and slave the source, be it a CD player, a soundcard, or an ethernet embedded computer (tried all of that, it works). For bonus points, add a SPDIF output to your DAC (in sync with your master clock) and use that to slave any soundcard (did that : it's good). Use your 8412 receiver in slave mode.
You know, asynchronous reclocking and Guido Tent style PLLs are only needed if you want to be compatible with existing gear in order to sell. If you are ready to add a clock input to your CDP, this does not matter, and it will save you LOTs of complications.

- and very important, read on layout and EMC, like papers by Guido Tent, or books, so you can handle those high-speed signals.

On neither the Monica or my other DAC, the MF A324 noted elsewhere on this forum, is there a filter between AGND and ground. Here's the MFA324 schematic once again (the CS8420 is on the left sice of schematic):





Not sure the about the intentions. Maybe it is to sell a product after all. Unfortunately, it doesn't seem that many of diyparadise's customers/DIYers dig deeper than the sites own jaded forum. Hopefully, though, with this thread, the search engines pick up on keywords and, hence, warn current and future customers interested in this DAC.

AGND is ground.
You should read the datasheet.

I think you guys went astray with these comments... There is no breakthrough here, it is rather simple DAC a la Kusunoki so I don’t expect it to be the best but it has to sound decent at least. If it doesn’t, something is IMO screwed up.

That's exactly the problem.
The Kusunoki approach is flawed too.

Hey, it's a good tool for testing interconnect cables...

I don't like the sound of uncompensated NOS dacs (it's not neutral, too much treble rolloff), and even less with unbuffered passive I/V.
Been there, done that.
It bugs me if there is someone around there building speakers, using as a reference source an uncompensated NOS dac.
Anyway, hollow_man has posted his impressions, which IMO and IME are accurate.
It sounds muted and undynamic, because it is NOS and not compensated (naturally...) but even more important, the passive output will never be happy driving an interconnect and a normal input impedance on the next stage (preamp).
It is a flawed design and it's sound will be very depended of what you have after it.

I'm not bashing anyone. Just giving my unbiased, sincere oppinion.
And I'm sorry if it is not nice for some commercial interests out there. I hope we can discuss freely without thinking about that.