[Haden2866]

I wonder if the holographic imaging of the SPS is due to a better implementation, or some possible advantage in clarity from a single chip, or maybe just the 1543 itself versus 1387?

Sorry to be OT but I thought this may be of interest to you NOS fans...a DIYAudio member has just freely released a RPi HAT PCB design using 8x TDA1387. With some DIY effort this could yield a nice NOS DAC that can be slaved to a Kali, for minimal cost. I'll start another thread for further comments about it

IMO 1543 is just slightly colored in some funny way that makes it sound holographic.

Yes, the SPS DAC3 is undoubtedly coloured; whether that is down to the single TDA1543 or the DAC implementation or, more likely, both, I am unclear on as this is my only experience with this chip.
There was some good fun way back in this thread discussing objectivity vs. subjectivity, or science vs. magic as I put it. I don't expect this DAC will measure well, far from it... and yet, it has its appeal and brings something special IMO. Richard agrees, certainly.
Malefoda (who is from France), over on diyAudio, had this to say when comparing it to his own (now mine) heavily-modified 8x (now 4x) TDA1387 DAC. Bear in mind this is based only on his assessment of the specs as published on the SPS website; he hasn't heard the DAC:
"About the DAC3, I must say I've looked for more and found the eBay auctions, I guess one can get one cheap if auctions are very low! BTW, his desriptions shows mistakes and "sell myst", maybe to sell more but here are some facts:
- the TDA1543 is not the last Philips Multibit, the TDA1387 is the last, up to 24bit/384kHz, even if useless that tells about internals manufacturing level, when the older TDA1543 is 16/192,
- these are multibit but not R2R! Few ICs are true R2R, like the BB PCM1704 which cost and arm. These Philips are segmented current sources, including patents and such, no R2R at all. But multibit yes, they share a lot sound-wise.
-the jitter-killer is sales pitch for me, the SPDIF DIR9001 does correct the jitter in some amount and is one of the best in this area even if not being an ASRC (which draw current and won't be great for batteries), your DAC is with CS8416 which does nothing in this area, so I guess that's what he calls some "reclocking jitter killer", and the USB card may be a classic daughter one which uses at least one oscillator to reclock, maybe 2 (your DAC does reclock USB like the Allo DigiOne with 44.1 and 48 base freq). So yes, it does lower some heavy jitter. And keeps some, hence soft sound and loss of some HF details.
- he does not correct the NOS, which is a problem always sold as a "feature": it's not accurate and roll off the highs, non-compensated raw NOS does that, and instead of correcting it sellers pretend it's the greatness of the NOS where it damage the original record... I understand some like it, it's a equalizer with less highs, not a good DAC... yours correct the problem, and show the real raw highs. Yes, that's not as fine that good DS, no wonder why the go one keep that "bug"...
- his output caps are too low, people use the smaller cap possible for the right frequency cut, to much low freq = unhappy amp ! And his caps are electrolytic...
his output caps are cheap lytics, not the best, worst, passive I/V is not right, cheap yes but not right! Some low impedance next stage won't cope with that (the DAC won't cope, the preamp does not mind). I've had a passive I/V on another TDA1387 (https://www.ebay.fr/itm/L1387DAC-4X-Quad-TDA1387-fever-USB-decoder-beyond-the-TDA1543/272418310441) and no way to get it work on my 5k preamp (I admit this is very low).
And no need for power supply design, he solved the problem with batteries. Clean power! Mine relies on mains quality, which nowadays is very polluted! Once I had a loud 12kHz pitch coming each time someone took the elevator, no way to filter it out... luckily they fixed the elevator (faulty power card)! With shunt supply no way for batteries, but I wanted life and slam and current is needed!
But again, maybe his very simple design (and risky/poor choices (passive I/V), output caps, my opinion no offence to him) does music well in most case (again, low Z preamp no way!), my choices are opposites, technically done first and it were done for my personal use, not for sell and profit. I never wanted to "smooth" the music but to hear it in its truth, even if then bad records sound terrible =) You can oversample on the RPi, that may change the SQ in some ways... on highs I guess.
You'll tell, maybe you'll hate it and then me... but I guaranty it's more properly engineered that the simple TDA1543, even if he managed to get less jitter on jittery sources and noise-free supply =) Voilà !"

 

[richard51]

Thanks for all this very interesting technical precisions.... I understand all that but I am in no way able to compare or judge your remarks.... But that were food for thinking and i trust more technical mind than me.... My own experience is that this SPS dac is very sensible to the noise on the electrical grid, and to the power supply...I use battery lithium, and all my electrical grid is cleaned with my own method the result is a natural holographic sound.... By the way I does not doubt that technically speaking this kind of dac rolls off the higher frequencies,but that is not at all my impression when I listen to it,but this is purely subjective impressions yes, but that speak volume to the musical quality of this dac... I think that the quality of cleanliness of the electrical grid play a great part in this final sound quality results of mine...Interestingly when I plug this dac in my system the first time,his sound was not completely satisfying and the high frequencies were rolled off indeed, but my electrical grid were not completely cleaned at this time( many months ago).... After burn in and cleaning of the grid the results were very different...

I will like to compared with another dac like the one you designed, but I pay 50 dollars for the SPS all in all and the results are so astonishing for me that I am a bit afraid to try something else because I cannot fault it in my audio system.... Perhaps you can say to me and to all what cheap dac will be better design and are an interesting buy.... It will be very interesting for me to compare with...Thanks very much for your useful and more than interesting analysis.... My best to you....

My dac is the dac2 of SPS I use it with hifimediy isolator and the hifimediy dac as optical convertor and the results are great...

One last remark: my audio system is I think very good vintage one but not absolute HIGH END, and I dont doubt that in a more refined absolutely high end system this dac will reveal some of his technical limitations... I say that because the majority of people may have good audio gear but not many had the high end system that will reveal the limitations of this musical and cheap dac....I say that because in my experience the cleaning of the electrical grid has at least a comparable impact and for me at the end greater on the sound quality at the end than almost any change of gear...

 

[stellarelephant]

That long quote from Malefoda was illuminating, perhaps brutally so. I was surprised by some of it...do the SPS DACs indeed have electrolytics for output coupling caps? I have tried just about every type excluding Black Gates in my cap rolling tests, and I have finally concluded that NONE of them sound transparent.

This was a surprise to me:

the TDA1387 is the last, up to 24bit/384kHz, even if useless that tells about internals manufacturing level, when the older TDA1543 is 16/192

The 1387 spec sheet reads:

"The TDA1387T (CC-DAC) accepts serial input data format of 16-bit word length"

How can we reconcile these two statements??

 

[gvl2016]

Also it is up to 384kHz word frequency, so 1387 is really a 16/192kHz chip.

 

[automojo]

You can always remove the caps, or put ones of smaller value.
Many amps are direct coupled.
As long as you don’t have a failure of a transistor, etc you’re in good shape
I do have to wonder, in a blind test with electrolytic caps and film caps of the same value, if you could honestly tell the difference.
The main advantage of film caps is stability over time, but size is a drawback.
If the design of the Muse takes into account the cap type and value, I see little point in changing those, or even removing them entirely.
The DAC Box S Fl is direct coupled.
I have a Muse coming, and will be interested in comparing to the Project.
I think Richard is correct in his assessment of dealing with noise in feeding it clean power. Especially if it’s design doesn’t filter out the noise from the probably poorly filtered stock switching power supply
Plus a lot of this depends on your system resolution your speakers headphones etc.
From what I’ve read so far, the muse seems to be a better choice for playing through speakers then headphones.

 

[Malefoda]

Hi there,

About my miswritten sample frequencies story, it's 16-bit / fws 192kHz for the TDA1543 and 24-bit / fws is 384kHz for the TDA1387. As for "bits" its plays like a champ with 24-bit samples, been sold as a real 24/192 SPDIF DAC which it is, (and tested further in frequency via I2S, more or less stable BTW). The TDA1387 is in no way "limited" to 16/192.

Must add that I've never wanted to spit on Christophe's DAC3 or other people efforts, these were pure guesses in private messages, apologies if inaccuracies or hurted anyone feelings. And for english mistakes...
Have fun and enjoy the music!

 

[richard51]

I will add only a simple remark about cleanliness of the electrical grid.... You cannot listen the optimal potential of any dac, even if the dac seems out of the grid connected to a lithium battery like mine, because you listen your dac connected upward from a source, computer or turntable, and you listen to it through your amplifier, and your amplifier is immersed directly in your house electrical grid like your computer or turntable, then what I had already said is simple, for me the most important upgrading improvement, more than any new amp buying or new dac, or new speakers etc is the way we clean the grid or not....The floor noise do not forgive...If you want to listen to the very good gear already in your posession, dont upgrade, clean the main line.... And vibrations, micro seismic effects, does not forgive also, then it is mandatory to separate of one another all pieces of gear,mechanically decouple them, damp them from internal resonance also and electro magnetically shield them also.... After that and only after that you will be ready for a costly upgrade... Without all that at the end you will change your gear without having listen to their true optimal sound quality because an audio system embedded in a non clean grid cannot reveal their optimal quality and your insatisfaction will misguide you .... And this is only my story before I discover all that.... Thanks to all for your patience with my rant....

 

[Haden2866]

Hi there,

About my miswritten sample frequencies story, it's 16-bit / fws 192kHz for the TDA1543 and 24-bit / fws is 384kHz for the TDA1387. As for "bits" its plays like a champ with 24-bit samples, been sold as a real 24/192 SPDIF DAC which it is, (and tested further in frequency via I2S, more or less stable BTW). The TDA1387 is in no way "limited" to 16/192.

Must add that I've never wanted to spit on Christophe's DAC3 or other people efforts, these were pure guesses in private messages, apologies if inaccuracies or hurted anyone feelings. And for english mistakes...
Have fun and enjoy the music!

Hi Malefoda, thanks for joining us.
I felt that your comments to me were interesting and relevant to the discussion here, and I tried to share them here in the correct context. If I have overstepped the mark or you feel I have unfairly represented you then please let me apologise. And again, I'd like to emphasise that the discussion on this thread is largely around how much people are enjoying this remarkable DAC, despite its apparent lack of technical sophistication. So to anyone just joining this thread, please don't be put off purely by the comments I quoted.

 

[Malefoda]

Hi Haden,
No problem, I wanted to be clear and avoid any misinterpretation about other's work. You know how forums can can suddenly take fire

 

[gvl2016]

Hi there,

About my miswritten sample frequencies story, it's 16-bit / fws 192kHz for the TDA1543 and 24-bit / fws is 384kHz for the TDA1387. As for "bits" its plays like a champ with 24-bit samples, been sold as a real 24/192 SPDIF DAC which it is, (and tested further in frequency via I2S, more or less stable BTW). The TDA1387 is in no way "limited" to 16/192...

First I stand corrected on the Fs for 1387, not sure what what I was thinking but 384kHz Fws means it can do up to 384kHz or 8x oversampling.

I stand on the supported bit-depth however, it is 16 bits not 24. The beauty of I2S is that the sender and receiver don't have to work on the same word length, if the sender sends a word shorter than the receiver can support it is padded with 0-s in LSBs. Inversely, if the sender sends more bits than the receiver can support they are simply truncated. Therefore a DAC that has a 24-bit SPDIF receiver and a 16-bit DAC chip and I2S in between will "play" 24-bit content at 16-bit resolution. Same goes if the DAC has an I2S input, you can send 24 bits or maybe even 32 depending on the Fs, it will play but only 16-bits it can do. Truncation is actually a problem and you should dither your 24-bit files to 16 bits in software prior to sending the signal to the DAC.

In other words 1387 is a 16/384kHz DAC, given the 18.4MHz max data clock frequency it will probably stand up to 24bits/384kHz serial data stream (2*24*384,000=18,432,000) but per above will only use the 16 bits per sample for the actual D/A conversion.

 

[stellarelephant]

I do have to wonder, in a blind test with electrolytic caps and film caps of the same value, if you could honestly tell the difference.

Differences between film and electrolytic caps in the signal path are quite easy to hear. In fact, 'lytics introduce enough of a sonic imprint that I can tell very clear differences between the different brands! In contrast, I find films to be pretty transparent, but even so, it isn't too hard to hear the difference between a polyester and a nicer dielectric material such as polypropylene or PPS. I have never tried a boutique cap such as a Mundorf before. But i assume they are better yet, or else they are grossly overpriced.

 

[buldogge]

What size are the caps in the SPS DAC3, and has anyone tried swapping films in? ...or perhaps bypassing the electrolytics with small films??

-Mark in St. Louis

 

[automojo]

s

Differences between film and electrolytic caps in the signal path are quite easy to hear. In fact, 'lytics introduce enough of a sonic imprint that I can tell very clear differences between the different brands! In contrast, I find films to be pretty transparent, but even so, it isn't too hard to hear the difference between a polyester and a nicer dielectric material such as polypropylene or PPS. I have never tried a boutique cap such as a Mundorf before. But i assume they are better yet, or else they are grossly overpriced.

I too use film caps in crossovers, for the most part, because they are more stable over time, but the new electrolytic are far better then 20 years ago.
But I have recapped speaker with electrolytic as well and they sound perfect, as you are getting the crossover back in to spec. That's what usually accounts for the difference in sound.
However you say you can hear the difference, but in a properly conducted blind test I have to wonder.. The mind convinces us of a lot of things. You read the (Muse) mod pages here, and some frankly don't hear any differences. And I have heard the same thing over the years..in speaker forums, amp preamp forums etc. This assumes the caps are functioning properly in the intended design.
Frankly, even a small 1/2 db level difference between to products can make one "sound" better then the other.
My main point is simply swapping out for films doesn't guarantee "better" sound
As mentioned I use film caps in my speaker crossovers, but I have to say, if a electrolytic caps was functioning properly, and has the same voltage rating, I can't really hear much if any difference.
Frankly, unless your super worried about DC, you might just remove them completely. As far as I can see, that appears to be their main function, DC protection. If you can get the job done with less components, I think you reduce the risk of incompatibility more then anything.
If it makes you feel better to use (film) them, I say have at it!!
I think what would be interesting is to see what the creator of the DAC feels about Film vs electrolytic, and why he chose to use coupling caps. To me that would make the most sense. Instead of randomly redesigning the DAC vs someone who actually designed, created and understands how his particular circuits work, and the goals of those designs. Circuits can be optimized for certain components, types, power supply, types/etc and take into account their drawbacks.

 

[automojo]

Well after spending 3 hours with the Muse TDA 1543x 4, comparing it to the direct coupled Project DAC Box S FL (4x TDA 1543), the are very close to call.
Running the Myse on the same 9 volt IFi power supply.
The Project has a very slight refinement edge, as it did over the tubed Maverick D1.
System used. H/K PT2300/PA2400, Marantz CD4001, modded ESS AMT3’s, with the Great Heil tweeter. (Lots of mods for electrical control, as mechanical). Transparent, but more musical then anything. (That was my goal with the AMT3 woofer and cab/crossover mods). But it’s very easy to hear slight differences with this set up particularly with the speakers.
The Heil is smooth, but not masking.
I could be happy with either DAC not lose any sleep, worrying about caps, etc.
I did find a huge difference running it off the 9 volt IFi power supply vs stock 12 volt, and a linear 12 as well.
It really became more balanced with the 9 volt supply. Being more at “ease” is a easier way of describing it.
Maybe 12 volts is over driving it?
Not sure.
Again I think I would be perfectly content with the stock Muse...it sounds (almost/slight) everybody is good is the $300 Project in my system anyway.
You could pick at the slight difference in smoothness, but for myself I wouldn’t lose any sleep!
Maybe spend the money on noise control, quiet power supply, and good mechanical isolation.
Pretty amazing little unit for $50 right out of the box.
Gives you that smooth tubed type analog sound with out the price penalty!!!

 

[gvl2016]

This gent thinks that the Muse is using fake TDA1543s: http://forums.hardwarezone.com.sg/h...3/muse-4-x-tda1543-dac-nos-sound-5430099.html , that are better than the originals