Any DIY hints for a DAC based on PCM1704 out there?
Cheers, mates.
I'm still massively impressed by having just read Charles' website, for example this page about R2R DACs.
Well, most of what's written on his site is not totally new to me, but the compelling nature of his approach does make me to want to build a battery powered R2R DAC right now.
The problem is: I'm inexperienced in electronics DIY, I'm still lacking a lot of knowledge, and without a circuit diagram, a parts list and a detailed description about what to do and what to avoid, I'm helpless here.
Can anybody point me to a resource where I can find some help or maybe even detailed advice?
Thanks in advance!
Ah, and don't be afraid to stop me if this project is downright nonsense. Be prepared to receive a lot of request for technical and physical evidence though.
Greetings from the snowy Alpes,
Axel
Cheers, mates.
I'm still massively impressed by having just read Charles' website, for example this page about R2R DACs.
Well, most of what's written on his site is not totally new to me, but the compelling nature of his approach does make me to want to build a battery powered R2R DAC right now.
The problem is: I'm inexperienced in electronics DIY, I'm still lacking a lot of knowledge, and without a circuit diagram, a parts list and a detailed description about what to do and what to avoid, I'm helpless here.
Can anybody point me to a resource where I can find some help or maybe even detailed advice?
Thanks in advance!
Ah, and don't be afraid to stop me if this project is downright nonsense. Be prepared to receive a lot of request for technical and physical evidence though.
Greetings from the snowy Alpes,
Axel
i don like tda1543 designs nor nonoversampling cause, even if 1 of 1000 tda1543 is close to theoretical perfomance (ok be it 15bit) , channel mismatch is still there. That motheroftone site ignores nonoversampling issues I'd rather be aware of.
Inertial: I don know how much difference there is between linear and minimum phase oversampling or is it percievable at all or just marketing talk (wadias neural net patent , EMM labs implementation, well , interesting at least), I think there are more important things downstream.
Inertial: I don know how much difference there is between linear and minimum phase oversampling or is it percievable at all or just marketing talk (wadias neural net patent , EMM labs implementation, well , interesting at least), I think there are more important things downstream.
Hi tritosine,
thanks for comment.
Based on my listening experience with Wadia DACS and big expansive 2 chassis top of the line DENON of 12-14 years ago ( do not remember model, mabe DA-s1 DP-s1?) I think those "ringing" are not only marketing claims.........
IMO they sounded like a great turntable, not digital sound at all
This is the reason why I would be curious to see a indicial response graphs/ impulsive response graphs of one of those Meitner DSD type
But I do not know where to find one
Cheers,
Paolo
thanks for comment.
Based on my listening experience with Wadia DACS and big expansive 2 chassis top of the line DENON of 12-14 years ago ( do not remember model, mabe DA-s1 DP-s1?) I think those "ringing" are not only marketing claims.........
IMO they sounded like a great turntable, not digital sound at all
This is the reason why I would be curious to see a indicial response graphs/ impulsive response graphs of one of those Meitner DSD type
But I do not know where to find one
Cheers,
Paolo
schiller said:
Dear Schiller,
glad to see your posting...the question is the same to me...
irritating is, that none of those engineers who rated the PCM1704U-K top of the "world" give statements on a comparison between 1704 to 1794/1792...but they all build newer equipment using 1794/1792 or Wolfson 8740...hence they keep a version left using the PCM1704U-K.... maybe this is the statement ;-)
(Heyink : www.dse-heyink.de/ , www.dse-heyink.de/DSEProFRQ.htm ; -not too expensive, well built, individualities possible; http://www.medianet-home.de/home.php?id=611 there click on dsp technologies ;
http://www.hoer-wege.de/dacup1704cmf.htm rather expensive, but fine too, different concept
only lesslossaudio keeps their statement onto the pcm1704
http://www.lessloss.com/page.html?id=30
on the other hand: now there are available
the V-dac of Musical Fidelity
and the new
dacmagic of Cambridge
and the prices are so low...
....I don't know either what to do...but I like the PCM1704U-K
;-))
one argument posted by engineers is often: the new 1792/1794 are combining both of those two worlds: R2R and 1-Bitstream ....but if this approach is better than pure R2R nobody really comments up to know...
so long
suchtgutenklang
Re: Re: PCM1704 or newer chips?
Hi suchtgutenklang,
excuse my little OT,
I am curious about current CD/SACD players Denon. They employ
1796(DCD CX3) , 1791(DCD700 and DCD1500), 1790(DCD2000),
1792(DCD SA1).
Are they HYbrid DACs or pure R-2R type?
Advanced Segment....is it a sort of DeltaSigma?
thanks,
Paolo
suchtgutenklang said:
one argument posted by engineers is often: the new 1792/1794 are combining both of those two worlds: R2R and 1-Bitstream ....but if this approach is better than pure R2R nobody really comments up to know...
so long
suchtgutenklang
Hi suchtgutenklang,
excuse my little OT,
I am curious about current CD/SACD players Denon. They employ
1796(DCD CX3) , 1791(DCD700 and DCD1500), 1790(DCD2000),
1792(DCD SA1).
Are they HYbrid DACs or pure R-2R type?
Advanced Segment....is it a sort of DeltaSigma?
thanks,
Paolo
...sorry but I missed the point...the only aim for me is......
Hi tritosine,
...sorry but I missed the point in your commment: what do you like to point out?
...the only aim for me is to get the "best" sounding d/a-conversion and up to now I only see (and heard) the pcm1704U-K as an unmatched IC.
Interesting thereby is, that no one seems really to contradict/disagree on this statement. ..and this is the situation which make Schiller(?) and me wondering I guess....
...maybe the ~new~ designed ICs seem able to work/ are better fitting on ~new~ samplefrequencies and dataformats but that means only that something is new...nothing else..or?
If for now the pcm1704U-K seems to give me/us the "best" sound I will still prefer it now...looking on until a "better sounding " alternative appears.
so long
suchtgutenklang
tritosine said:dont forget 1704 was "not recommended for new design" for years, Analog Devices discontinued all their audio R2r, and TI made 1704 ROHS compatible only last year, thereby lifting "not recommended for new design".
Hi tritosine,
...sorry but I missed the point in your commment: what do you like to point out?
...the only aim for me is to get the "best" sounding d/a-conversion and up to now I only see (and heard) the pcm1704U-K as an unmatched IC.
Interesting thereby is, that no one seems really to contradict/disagree on this statement. ..and this is the situation which make Schiller(?) and me wondering I guess....
...maybe the ~new~ designed ICs seem able to work/ are better fitting on ~new~ samplefrequencies and dataformats but that means only that something is new...nothing else..or?
If for now the pcm1704U-K seems to give me/us the "best" sound I will still prefer it now...looking on until a "better sounding " alternative appears.
so long
suchtgutenklang
A comment about multibit versus "single bit"
It is possible to build really good DAC's using either single-bit or multi-bit DAC chips. Today, most DAC chips are a combination of both. They use 4-5 bits sigma-delta modulated.
The problem with low number of bits is that your clock needs to be more precise (jitter - i.e short term variations). The advantage is linearity. You are also limited to a relatively low oversampling rate. These units typically measure way better than the ones below. The units typically have integrated digital filters which may or may not be designed for good sound but again measure really well. The process technology is typically on one domain (i.e digital). There is a lot of investment in this technology by manufacturers. I am going to be very imprecise and compare how the technology works with "LCD displays or CMOS sensors for digital cameras" - i.e take something in the wrong domain and tweak it into the domain you desire..
The problem with "fully populated bits" is that the cost is high and linearity is hard to achieve. The clock sensitivity is much lower. Many audiophiles like the sound, and i is easy to do non-oversampled designs and for many chips high oversmapling designs as well. The process technology is typically in two domains (both analog and digital making it harder to produce). There is not a lot of investment in this technology by manufacturers. I am going to be very imprecise again and compare this with "Plasma Displays or CCD sensors for digital cameras" - i.e take something near the right domain and try to make a good job of it even though some elements are hard to achieve.
I am sure a lot of others have opinions on these matters and will excuse my ranting and silly "comparisons". Again, excellent results can be had from both camps. I am about to buy a TV now BTW - it will be LCD. I am also buying another camera with a CMOS sensor. My next DAC though will be multi-bit.
Petter
It is possible to build really good DAC's using either single-bit or multi-bit DAC chips. Today, most DAC chips are a combination of both. They use 4-5 bits sigma-delta modulated.
The problem with low number of bits is that your clock needs to be more precise (jitter - i.e short term variations). The advantage is linearity. You are also limited to a relatively low oversampling rate. These units typically measure way better than the ones below. The units typically have integrated digital filters which may or may not be designed for good sound but again measure really well. The process technology is typically on one domain (i.e digital). There is a lot of investment in this technology by manufacturers. I am going to be very imprecise and compare how the technology works with "LCD displays or CMOS sensors for digital cameras" - i.e take something in the wrong domain and tweak it into the domain you desire..
The problem with "fully populated bits" is that the cost is high and linearity is hard to achieve. The clock sensitivity is much lower. Many audiophiles like the sound, and i is easy to do non-oversampled designs and for many chips high oversmapling designs as well. The process technology is typically in two domains (both analog and digital making it harder to produce). There is not a lot of investment in this technology by manufacturers. I am going to be very imprecise again and compare this with "Plasma Displays or CCD sensors for digital cameras" - i.e take something near the right domain and try to make a good job of it even though some elements are hard to achieve.
I am sure a lot of others have opinions on these matters and will excuse my ranting and silly "comparisons". Again, excellent results can be had from both camps. I am about to buy a TV now BTW - it will be LCD. I am also buying another camera with a CMOS sensor. My next DAC though will be multi-bit.
Petter
Re: Re: Re: PCM1704 or newer chips?
Hi Paolo,
regarding the several mentioned ICs polease look up the datasheets like:
http://focus.ti.com/docs/prod/folders/print/pcm1794.html;
but you understood right I guess:
'Figure 30. Advanced Segment DAC
The PCM1794 uses TI’s advanced segment DAC architecture to achieve excellent dynamic performance and
improved tolerance to clock jitter. The PCM1794 provides balanced current outputs.
Digital input data via the digital filter is separated into 6 upper bits and 18 lower bits. The 6 upper bits are converted
to inverted complementary offset binary (ICOB) code. The lower 18 bits, associated with the MSB, are processed
by a five-level third-order delta-sigma modulator operated at 64 fS by default. The 1 level of the modulator is equivalent
to the 1 LSB of the ICOB code converter. The data groups processed in the ICOB converter and third-order
delta-sigma modulator are summed together to create an up-to-66-level digital code, and then processed by
data-weighted averaging (DWA) to reduce the noise produced by element mismatch. The data of up to 66 levels from
the DWA is converted to an analog output in the differential-current segment section.
This architecture has overcome the various drawbacks of conventional multibit processing and also achieves
excellent dynamic performance.'
means a sort of the upper 6 bits to be processed multibit-alike
so far
suchtgutenklang
1796(DCD CX3) , 1791(DCD700 and DCD1500), 1790(DCD2000),
1792(DCD SA1)
Are they HYbrid DACs or pure R-2R type?
Advanced Segment....is it a sort of DeltaSigma
Paolo [/B]
Hi Paolo,
regarding the several mentioned ICs polease look up the datasheets like:
http://focus.ti.com/docs/prod/folders/print/pcm1794.html;
but you understood right I guess:
'Figure 30. Advanced Segment DAC
The PCM1794 uses TI’s advanced segment DAC architecture to achieve excellent dynamic performance and
improved tolerance to clock jitter. The PCM1794 provides balanced current outputs.
Digital input data via the digital filter is separated into 6 upper bits and 18 lower bits. The 6 upper bits are converted
to inverted complementary offset binary (ICOB) code. The lower 18 bits, associated with the MSB, are processed
by a five-level third-order delta-sigma modulator operated at 64 fS by default. The 1 level of the modulator is equivalent
to the 1 LSB of the ICOB code converter. The data groups processed in the ICOB converter and third-order
delta-sigma modulator are summed together to create an up-to-66-level digital code, and then processed by
data-weighted averaging (DWA) to reduce the noise produced by element mismatch. The data of up to 66 levels from
the DWA is converted to an analog output in the differential-current segment section.
This architecture has overcome the various drawbacks of conventional multibit processing and also achieves
excellent dynamic performance.'
means a sort of the upper 6 bits to be processed multibit-alike
so far
suchtgutenklang
Ak4395 beats pcm1704
I have an Assemblage Dac 2.7 which has DF1704 filter and PCM1704 dac chip. It's been a long time since I have done anything with it because a DEQ2496 with the direct out mod and dac swap to AK4395 absolutely kills it's sonics. Just as the Dac2.7 creamed my old Enlightened Audio Designs dsp1000 with the pcm63. There are no legacy dacs. Technology marches on. I would hate to have paid $15,000.00 for a Mark Levinson 10 years ago.
I have an Assemblage Dac 2.7 which has DF1704 filter and PCM1704 dac chip. It's been a long time since I have done anything with it because a DEQ2496 with the direct out mod and dac swap to AK4395 absolutely kills it's sonics. Just as the Dac2.7 creamed my old Enlightened Audio Designs dsp1000 with the pcm63. There are no legacy dacs. Technology marches on. I would hate to have paid $15,000.00 for a Mark Levinson 10 years ago.
Hi Scott,
The complete units may sound better or worse depending on how the audio signal is treated after the D/A process. I have seen several examples of this over the years.
I would love to build a PCM 1704 DAC. I do have a Denon DCD S-10. That's the nicest machine I've ever heard in that price range. It uses PCM 1702 DACs and sounds utterly natural.
The PCM 1702 (and PCM 1704) use the same resistor ladder for two DACs. These DACs are feed complimentary data and are used push-pull inside the chip. That means that each chip has canceled out non-linearity to a great extent. The idea works.
Now, at this level of performance, it's terribly easy to destroy the sonics. PCB layout and all the rest of the chain becomes critical. That's why the same DAC can be used in different machines that have dissimilar audio quality.
Denon did a surprisingly competent job with this.
Now, can we do any better? Can we equal the performance of the Denon units. I'm only using this as a reference as I'm sure there are better units out there using the same DACs and filters.
I did lust after the one model up from this. Single chassis DCD "S" series single chassis design. My feeling is that you lose too much by combining the three (minimum) signals and leaving the chassis. If your signals remain inside and separate, they can be treated properly and your D/A section has much less jitter to deal with.
-Chris
The complete units may sound better or worse depending on how the audio signal is treated after the D/A process. I have seen several examples of this over the years.
I would love to build a PCM 1704 DAC. I do have a Denon DCD S-10. That's the nicest machine I've ever heard in that price range. It uses PCM 1702 DACs and sounds utterly natural.
The PCM 1702 (and PCM 1704) use the same resistor ladder for two DACs. These DACs are feed complimentary data and are used push-pull inside the chip. That means that each chip has canceled out non-linearity to a great extent. The idea works.
Now, at this level of performance, it's terribly easy to destroy the sonics. PCB layout and all the rest of the chain becomes critical. That's why the same DAC can be used in different machines that have dissimilar audio quality.
Denon did a surprisingly competent job with this.
Now, can we do any better? Can we equal the performance of the Denon units. I'm only using this as a reference as I'm sure there are better units out there using the same DACs and filters.
I did lust after the one model up from this. Single chassis DCD "S" series single chassis design. My feeling is that you lose too much by combining the three (minimum) signals and leaving the chassis. If your signals remain inside and separate, they can be treated properly and your D/A section has much less jitter to deal with.
-Chris
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.