Interesting. The build still looks intimidating because of the mounting distances etc. I sort of enjoy plug-n-play solder project kits though
In its basic form, it's one of the simplest audio projects ever, not much more complicated than a stereo Gainclone
You essentially have just:
- CS8412 IC and its power supply
- TDA1543 IC and its power supply
- 1 input resistor for CS, 3 output resistors for TDA, 2 output caps and 1 resistor+1cap for PLL filter. I think that's all.
The schematic only looks daunting because the power supplies are very sophisticated, but you could use a much simpler arrangement and still get very good sound.
Nichicon KG are certainly not the worst choice, but I would suggest trying 2x 2200uF Elna Starget in the TDA PSU. I think they are a bit clearer than Nichicons (of the Nichicon family, I liked FG series best overall). A similar arrangement would work for the CS PSU, but you could try a 1000uF Sanyo Os-Con instead of the second Starget, I think it works a bit better here.
If you are using Bobken regs, I would go for a 15-33uF Os-Con at the output, although you might prefer Starget in the TDA regulator.
Keep in mind, these are still not quite as good as the Black Gates - but very good nonetheless; the best I could find, at any rate. And you can of course wood-mod them, which will greatly enhance the sound.
For bypassing, I would simply go for the best film cap that meets your size / budget constraints, the same for output caps. Your current choices are ok for their price.
Also, it is not on the new schematic, but you could try installing a 10nF cap at the DAC input. With most caps, this will degrade dynamics and colour the sound; however the best ones might actually be an improvement. I'm talking Jupiter Beeswax / V-Cap / Duelund sort of quality (or my hand-made caps!
).
Cap choices
I enjoyed Peter's usb dac for quite some time with the supplied BGs. In most situations I prefer BG sound to most everything. I got bitten by the tda1541a and for some time used Peter's dac as the USB to I2S converter. Effectively stealing the I2S from the PCM2706. It worked well. In my mania to perfect the new 1541a dac I borrowed the BGs from Peter's Dac. I replaced the 1000u BG std in C1,3 with pretty generic 16v 5800u caps (could be Panasonic) and replaced the BGN 1000u at C2,4 with OSCON SEP 16v 150u. Also changed supplied caps @ C12 and C8 with OSCON SEP 16v 150u.
I expected to kind of impair Peter's Dac with these 'inferior' caps. Big surprise. To my ear sound opened up. More bass better controlled and a bit of a veil removed across the board. Peter's Dac eventually migrated out of the picture and was replaced fully by the WaveIO as a USB to I2S converter. This is just in a different league altogether.
Point of this is there are cap choices that can work very well. I am guessing that the pcm2706 really mates well with the Oscon SEP and most anything in the first stage of the crc will suffice.
I enjoyed Peter's usb dac for quite some time with the supplied BGs. In most situations I prefer BG sound to most everything. I got bitten by the tda1541a and for some time used Peter's dac as the USB to I2S converter. Effectively stealing the I2S from the PCM2706. It worked well. In my mania to perfect the new 1541a dac I borrowed the BGs from Peter's Dac. I replaced the 1000u BG std in C1,3 with pretty generic 16v 5800u caps (could be Panasonic) and replaced the BGN 1000u at C2,4 with OSCON SEP 16v 150u. Also changed supplied caps @ C12 and C8 with OSCON SEP 16v 150u.
I expected to kind of impair Peter's Dac with these 'inferior' caps. Big surprise. To my ear sound opened up. More bass better controlled and a bit of a veil removed across the board. Peter's Dac eventually migrated out of the picture and was replaced fully by the WaveIO as a USB to I2S converter. This is just in a different league altogether.
Point of this is there are cap choices that can work very well. I am guessing that the pcm2706 really mates well with the Oscon SEP and most anything in the first stage of the crc will suffice.
If you use an USB/I2S interface like the one wlowes mentioned, you only need the TDA1543 chip + its PSU and output components.
@wlowes - which version of NOS DAC you were using? I have not heard any TDA1541-based DAC, that would match Peter's latest version for naturalness and raw energy, so I'm kinda curious what pushed you towards 1541?...
Right, to build a TDA1543 DAC mated to WaveIO. It would likely be a good combination. In my case I used for a while peter's kit minus the 1543 to supply I2S to the 1541a. Good but clearly inferior to WaveIO.
I use a home built 1541a DAC. Basically nothing more than the chip with a Grundig style DEM circuit, smt organic polymer decoupling caps on the pins, followed by 2np2 tube buffer..russian PIO coupling caps. Power supplies well filtered with BGN and discrete low noise regs.
It is to my ear pretty good. I evolved in this direction prior to the revised approach to 1543 described by this thread. It may well be better than my 1541a. No question however, my 1541a driven by Peter's I2S segment was much better in naturalness than Peter's original....the WaveIO took it into another realm. I expect you and Peter have created something awesome, but I have no opportunity to compare as I have not built it. At this stage for me the 1541a mystique has grabbed me, and I am happy with the audio result.
I continue to watch this thread as I can imagine building this one day for an alternate experience.
I have one of those MUSE 1543x 4 DACS that uses the DIR9001 receiver chip. The specifications claim 50pSec jitter for the DIR9001 and 200pSec for the CS8412.
I am planning on building my own DAC and have read that NOS DAC's are less critical of jitter. Has anyone tried both receivers or know of any sonic differences in using either one?
I am planning on building my own DAC and have read that NOS DAC's are less critical of jitter. Has anyone tried both receivers or know of any sonic differences in using either one?
If I understand this new build it is just under the premise that better power is the only thing the 1543 needs... Then some DC coupling output, correct?
Is the improved power about regulation, or noise rejection? I haven't been under the impression that power regulation improves sound as much as noise rejection, yet.
Is the improved power about regulation, or noise rejection? I haven't been under the impression that power regulation improves sound as much as noise rejection, yet.
Hi,
I am also one of the DIYers who recently asked Peter for a DAC kit unlikely sold out.
I was able to built Peter's chipamp kit, but it was fundamental Peter's thread on gainclone instructions for beginners. Can you address me to a similar thread?
Thank you
Renato
I cannot comment on the regulation vs noise issue, but I do know that Peter and Bobken went to great lengths optimising passive components in their latest DACs. By optimising I mean choosing the best sounding value, as well as choosing the best sounding brand/type, in each position.
I'm afraid the NOS DAC is a bit more complex than a Gainclone... I mean it is simple for a DAC, especially if you use IC regulators, but there are still things to consider, like managing digital/analogue grounds, shielding, input isolation etc.
I mean, you could just solder it together on a protoboard as per Peter's schematic, and it would likely work fine - all I'm saying is that the range of trouble you could potentially run into is much broader than with a Gainclone... So I do not think there is a thread with instructions for beginners because without the kit, it is not really a project for beginners...
Thank you Uncle Leon. Let's wait, may be Peter changes idea and prepares a new batch of kits.
Renato
Here is the most basic NOS DAC circuit that you can attempt to build, that will still work (and likely sound reasonably good too). I took Peter's "classic" schematic and edited it to simplify it. I also incorporated some of the recent updates into it. Have a look and see if this is something you would feel confident attempting.
This simplified DAC should actually sound better than original - with the following reservations:
- the original NOS DAC used very high quality passive components (Black Gate, Riken, Caddock), some of which are no longer available.
- the IC regulators that Peter liked are also no longer produced, and somewhat hard to obtain, which means you may have to resort to LM78xx-type regulators (which will likely sound worse). Where simplicity is not that important, one could go with any of the new low-noise regulators, or buy/make a "super regulator" (such as the Bobken reg discussed here earlier).
- using different regulators could mean that the optimal values for their output caps are now different.
- your shielding arrangements, and wire routing choices will have some impact on the sound, and may sound a bit worse than Peter's PCB (although it may sound better too).
- the classic DAC uses a ferrite bead at the input if the regulator U3. I prefer the sound without it, but some people like it. It does (theoretically) make sense to install it if you use a single trafo, because it will limit "spreading the contaminations" in the power supply.
One could probably go even further with the simplification, and feed the regulators from the same PSU, but I think that would be going a bit too far in terms of sound quality compromise, and it would not simplify things that much (if you know how to build one diode bridge, then you certainly know how to build another).
Attachments
Thank you! Now I know how to spend Christmas holidays... at least studying the diagrams. I'll try.
Renato
This post shows the most cost effective way to convert my original DAC into recently updated version. For even simpler approach, discreet regulators on receiver chip could be also omitted: http://www.diyaudio.com/forums/audio-sector/187748-pushing-limits-tda1543-nos-dac-2.html#post2551745