Well... I think sorfbrw said:
I am not B_online but:
"PD- I know there are some other designs around, but I need to build something I can easily get the PCBs, mounting instructions and some kind of feedback or help should anything go wrong, since this wll be my first ever project besides the Crossover from my speakers....."
Well, it's a kit with pcbs. So it should be doable. BTW it is a non-oversampling DAC and I think it's a good one! Though that's a wild guess. If I would have the money I would try and build one, but...
Fedde
If the Curcio dac is non-oversampling then I'm a flying pig in a quite convincing wig. Perhaps you ought to read the details on the website http://www.curcioaudio.com/cddes_3.htm especially the bit about the PLL interfacing with the digital filter and more references to the dac here. http://db.audioasylum.com/cgi/m.pl?forum=digital&n=54779&highlight=curcio&r=&session=
ray.
ray.
I am sorry, you are a pig
Check:
http://db.audioasylum.com/cgi/m.pl?forum=digital&n=45531&highlight=curcio&session=
http://db.audioasylum.com/cgi/m.pl?forum=digital&n=59056&highlight=curcio&r=&session=
And a nice review:
http://db.audioasylum.com/cgi/m.pl?forum=digital&n=41221&highlight=curcio&r=&session=
"A call to Joe Curcio confirmed that the chip operates in non-oversampling mode (his webpage is inaccurate on this point), with no "brickwall" digital filter, although a digital filter chip is present to furnish glue logic".
Fedde
Check:
http://db.audioasylum.com/cgi/m.pl?forum=digital&n=45531&highlight=curcio&session=
http://db.audioasylum.com/cgi/m.pl?forum=digital&n=59056&highlight=curcio&r=&session=
And a nice review:
http://db.audioasylum.com/cgi/m.pl?forum=digital&n=41221&highlight=curcio&r=&session=
"A call to Joe Curcio confirmed that the chip operates in non-oversampling mode (his webpage is inaccurate on this point), with no "brickwall" digital filter, although a digital filter chip is present to furnish glue logic".
Fedde
Hi guys - I have the Curcio DAC. In stock form, it works in oversampling mode. Typical digital sound, hash, little decay etc....
I modified it with new caps and also took it to non-oversampling and it is far better (IMHO). i.e. I took 3 wires from the Pins 1,2,3 of the SAA7220 and wired them directly to Pins 1,2,3 of the TDA1541.
In stock, the TDA1541 pins 1,2,3, are connected to the last pins of the SAA7220.
VinceL
fedde said:I am sorry, you are a pig
Check:
http://db.audioasylum.com/cgi/m.pl?forum=digital&n=45531&highlight=curcio&session=
http://db.audioasylum.com/cgi/m.pl?forum=digital&n=59056&highlight=curcio&r=&session=
And a nice review:
http://db.audioasylum.com/cgi/m.pl?forum=digital&n=41221&highlight=curcio&r=&session=
"A call to Joe Curcio confirmed that the chip operates in non-oversampling mode (his webpage is inaccurate on this point), with no "brickwall" digital filter, although a digital filter chip is present to furnish glue logic".
Fedde
Complete and utter pig swill. Anyone capable of reading the SAA7220 datasheet can see through such marketing nonsense. BTW both the SAA7220 and the TDA1541 can accept an I2S input so glue logic is not needed.
ray
Guys,
Calm down! - firstly, I know its easy given the faceless nature of the Internet to get excited and second, often the language/cultural barriers give rise to misunderstandings/misinterpretations.
Hence, could we please refrain from name calling/flaming etc...?
Okay. Down to business. I have not heard every DAC out there but guided by Thorsten and others, I do believe that the TD1541-based dac gives the most "life like" sound. If you open up older players with this chip, you'll know what I mean.
I found the Curcio DAC to be the most suited TD-1541 based DAC in my budget of US$1,000 (assembled PCB). My dream DAC is the Zanden DAC but that baby is in the stratosphere - and as much as I would like to go out to get it,
MrsL unfortunately would not let me back in the house!
Anyway, I plan to write a review on it when I get the chance and post in AA.
Overall, I like the DAC as in my quite modified form, it plays very easy on the ears, soothing and enjoyable music and should last a long time.
Joe Curcio, the designer decided to put the effort into the DAC stage while less attention was paid to the power supply stage. My gripe here is that I would have preferred a comensurate amount of effort to have also been put into the power supply section - but hey, us greedy modders can't have it all!
Pls do not flame me in my 2 cent attempt to clarify the misunderstanding behind the aforementioned heated debate on nonoversampling and oversampling. AFAIK the chap who reviewed the Curcio DAC in AA misunderstood the oversampling/non-oversampling aspect and posted the DAC as non-oversampling in the review. The Curcio website does not say anywhere that it is non-oversampling.
You HAVE TO take it to non-oversampling yourself as I did and I enjoy the DAC much more for that.
VinceL
Well, I modified my Marantz CD-94 by replacing all electrolytics to Elna and HFQ, also used BG whenever possible with N Hi Q (14 of them) bypass at the DAC chip. I used OPA 627 in the output with Vishay for feedback and stripped down original configuration. It is TDA 1541, no oversampling.
I still prefer TDA1543 with better PS (passive filtering and Elso recommended regulatior). I use reclocking and MIT RTX coupling caps. To my ears this DAC sounds smoother and I don't think the highs are so bad as others claim. It is definitely better than my Broadhurst design based DAC with CS8420, PDF 1704 and parallel PCM 1704, and all 8 transformers. That DAC is just not as natural sounding as non oversampling designs and you feel the touch of "digital".
I still prefer TDA1543 with better PS (passive filtering and Elso recommended regulatior). I use reclocking and MIT RTX coupling caps. To my ears this DAC sounds smoother and I don't think the highs are so bad as others claim. It is definitely better than my Broadhurst design based DAC with CS8420, PDF 1704 and parallel PCM 1704, and all 8 transformers. That DAC is just not as natural sounding as non oversampling designs and you feel the touch of "digital".
VinceL: thanks for your reaction. We were just joking about the pigs, nothing serious really!
Peter: there are probably still a lot of things you can do to improve your TDA1541 CD-P. What capacitor do you use on the chip (I forgot the pins), 470 pF ? I forgot the details, but I recall that was important for the performance. Also I connected the clock (Guido clock) directly to pin4. It shouldn't work, but seems a big improvement! Try biasing the opamp into class-A. And try if a buffer after that opamp helps...
(maybe use a separate supply for the opamps !? Or tubes
)
Fedde
Peter: there are probably still a lot of things you can do to improve your TDA1541 CD-P. What capacitor do you use on the chip (I forgot the pins), 470 pF ? I forgot the details, but I recall that was important for the performance. Also I connected the clock (Guido clock) directly to pin4. It shouldn't work, but seems a big improvement! Try biasing the opamp into class-A. And try if a buffer after that opamp helps...
(maybe use a separate supply for the opamps !? Or tubes
)Fedde
Earlier in this thread there is a reference to the Retro DAC. (http://www.geocities.com/yury_g/dac.htm) What is the advantage of using two DACs in the complimentary fashion shown in the schematic? You certainly don't need complimentary DACs to produce a balanced output. The only function I see for the complimentary DACs is to screw-up the linearity.
Consider the simple data sequence of 1, 2, 4. Each data point is twice the value of the previous one. Converting that sequence with the Retro DAC, the plus DAC gets 1, 2, 4, the minus DAC gets -2, -3, -5, and the resulting differential output is 3, 5, 9. Each value is not twice the previous one. Similarly, with the input sequence of -1, -2, -4, the plus DAC gets -1, -2, -4, the minus DAC gets 0, 1, 3, and the resulting differential output is -1, -3, -7.
The data coming from the CD is a signed binary number. To properly compliment it you need to invert every bit and then ADD ONE.
Consider the simple data sequence of 1, 2, 4. Each data point is twice the value of the previous one. Converting that sequence with the Retro DAC, the plus DAC gets 1, 2, 4, the minus DAC gets -2, -3, -5, and the resulting differential output is 3, 5, 9. Each value is not twice the previous one. Similarly, with the input sequence of -1, -2, -4, the plus DAC gets -1, -2, -4, the minus DAC gets 0, 1, 3, and the resulting differential output is -1, -3, -7.
The data coming from the CD is a signed binary number. To properly compliment it you need to invert every bit and then ADD ONE.
Broadhurst design based DAC
Sorry for jumping in but I felt I should offer Peter Daniel my experience with this dac.
I built it myself with the pcm1704, balanced output and passive iv and initially found it way to "digital"
I also use a dvd player as source and when fed 20 or 24 bit pcm code it was a completely different dac, all the digital harshness was gone.
Due to the dvd experience (and indirect advice on this forum) I eventually started thinking it might be bit depth problems.
To the point...before giving up I rebuilt the digital out on my cdp by putting in a 8420 in hw mode 2, 16 bit input and 24bit out.
Suddenly it was sooo... smooth and the "digital" feel was gone.
I can not conclude anything else then that the digital feel was caused by quantization noise.
The 8420 in hw mode 1 has no way of handling different bit depths, it treats everything as 24 bits.
So my conclusion regarding this dac (or any other using the 8420 as DIR and 24 bit DF and dac) is that one has to make sure the 16 bit output from a cdp is properly dithered or scaled before feeding the 8420.
Sorry for jumping in but I felt I should offer Peter Daniel my experience with this dac.
I built it myself with the pcm1704, balanced output and passive iv and initially found it way to "digital"
I also use a dvd player as source and when fed 20 or 24 bit pcm code it was a completely different dac, all the digital harshness was gone.
Due to the dvd experience (and indirect advice on this forum) I eventually started thinking it might be bit depth problems.
To the point...before giving up I rebuilt the digital out on my cdp by putting in a 8420 in hw mode 2, 16 bit input and 24bit out.
Suddenly it was sooo... smooth and the "digital" feel was gone.
I can not conclude anything else then that the digital feel was caused by quantization noise.
The 8420 in hw mode 1 has no way of handling different bit depths, it treats everything as 24 bits.
So my conclusion regarding this dac (or any other using the 8420 as DIR and 24 bit DF and dac) is that one has to make sure the 16 bit output from a cdp is properly dithered or scaled before feeding the 8420.
Re: Broadhurst design based DAC
To do that, do you need 2 CS8420 receivers (one in cdp, one in DAC) or you can simply configure the one in the DAC only?
I was reading review of Shanling 100 and it appears that it sounds much better when oversampling mode in 8420 is disabled.
I'm not that familiar with Crystal chip and I don't remember the exact implementation in Broadhurst design, but I guess 8420 is in oversampling mode. How one can disable that?
A 8 said:
To do that, do you need 2 CS8420 receivers (one in cdp, one in DAC) or you can simply configure the one in the DAC only?
I was reading review of Shanling 100 and it appears that it sounds much better when oversampling mode in 8420 is disabled.
I'm not that familiar with Crystal chip and I don't remember the exact implementation in Broadhurst design, but I guess 8420 is in oversampling mode. How one can disable that?
It is not ideal but -yes, I put in another 8420 in the cdp.
To get a similar solution in the dac you would need to add a DIR (like the 8412) that output 16 bit serial data, then use the 8420 in mode 2 serial in, serial out and set it to receive 16 and output 24 bits.
The problem with this is that you can not make good out of a native 20-24 bit spdif without it cutting off the 4-8 LSB's.
Well, you can disable it but it will still not manage the bit depths properly and the whole benefit of the added master clock disappears, it gets similar to the 8414 in its function.
A 16 bit dac must be great as long as you feed it from a cdp.
Personally I would not want to exclude access to the sonic benefits of some of the native 20 or 24 bit LPCM dvd's out there..
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.