uhmmmm...
I´m working on a PCB for this chipset: AD1896, low jitter osc, DF1704, PCM1704, AD8610, AD8620 for output buffer AND balancing chip. Thinking on several input formats, I2S, EIAJ, bypass on the AD1896 too, etc...
receiver still not defined (maybe CS8414 ??).
I am considering to include 2 pcm per channel and already have a board with pcm1702 working with 4 per channel.
I do intend to share the circuit and board design if you want, no problem. But I think I will share design early as you can do some suggestions on it, helping to define some "y" ways in the board design.
If there is somebody who want took part of it, please e-mail me and lets share both design and effort. As Joko said... time is the most expensive part of the project.
Cheers...
Euclides.
I´m working on a PCB for this chipset: AD1896, low jitter osc, DF1704, PCM1704, AD8610, AD8620 for output buffer AND balancing chip. Thinking on several input formats, I2S, EIAJ, bypass on the AD1896 too, etc...
receiver still not defined (maybe CS8414 ??).
I am considering to include 2 pcm per channel and already have a board with pcm1702 working with 4 per channel.
I do intend to share the circuit and board design if you want, no problem. But I think I will share design early as you can do some suggestions on it, helping to define some "y" ways in the board design.
If there is somebody who want took part of it, please e-mail me and lets share both design and effort. As Joko said... time is the most expensive part of the project.
Cheers...
Euclides.
hagtech said:
No, I meant easiest. The hard parts would be 0603 sizes, a DF1704 (TSSOP), SOT-323, etc.
Just wondering if I should offer the HagDac as a blank circuit board. Figured the surface mount would be a problem for most DIYers. The responses here indicate it is not.
jh
You are of course correct I did not consider the fact that the DF170* are finer pitch than the PCM as it makes no difference if you have the correct technique (not pin at a time method) regarding the 0603, SOT excreta they are less dependant on method so in my opinion easier to learn (opinion gained from training many operators over the years).
Selling a pcb that is primarily SMT will limit your sales to those that are confident or daring enough, but the camp of people that have not learnt the required skills will not have the chance to build or modify modern electronics unless they take a chance and learn.
There are several very good posts on this forum explaining the best way to build with these devices.
Dual differential configuration with DF1704/6 and PCM1704 + balanced analog stage could be of interest to people who use balanced preamps / amps and believe that balanced has benefits when it comes to cancellation of certain distortions, immunity to noise (cancellation of the same), better dynamics and ability to follow balanced analog signal further down to their preamps and amps and benefit from balanced analog even more.
I also understand DIYers who would like simplicity in single 1704 per channel and would like to preserve connectivity to their existing single ended setups, i.e. nonbalanced RCA’s
I’ll be nice and contribute the 'balanced 1704' idea by sending the diagram of DENON 1650SR player to those thinking of (yes?!) putting this KIT together. It is a proprietary upsemler and dual differential 1704 + balanced I/V, followed by summing analog filter/amplifier, and follower design.
Extreme_Boky
I also understand DIYers who would like simplicity in single 1704 per channel and would like to preserve connectivity to their existing single ended setups, i.e. nonbalanced RCA’s
I’ll be nice and contribute the 'balanced 1704' idea by sending the diagram of DENON 1650SR player to those thinking of (yes?!) putting this KIT together. It is a proprietary upsemler and dual differential 1704 + balanced I/V, followed by summing analog filter/amplifier, and follower design.
Extreme_Boky
hagtech said:
A late reply to this. Somehow I missed the thread earlier. I for one would almost certainly buy a board of you made it available. The current built up price is sadly outside my means at the moment (well, outside my discressionary fun budget
) but I would probably be able to justify a kit price. Whilst not exactly what I am looking for it is easily close enough, and my current ideas of designing my own mostly a pipe dream with my current commitments.Surface mount is really just becoming a fact of life. Perhaps a little less forgiving, but in the end probably no worse in terms of difficulty than doing a really good and careful through hole job.
If anyone figures they'll do this board and sell it to me I would probably only be interested in a pcb design that is truly made for a dual diff design using the 1704, completely symmetrical in all aspects of the layout. There should be no "special fixes" or wierd add ons to get the dual diff function.
If there are to be any options on the PCB I would rather have them on the spdif reciever side and I/V conversion side.
On the reciever side there could be a 24/96+ capable DIR, an option for a ASRC that could be bypassed, perhaps a "tent type"~can clock layout on the board with close and direct connections, through individual single gate buffers, individually ps supplied to bitclock on the DAC's, MCKL on the filter and the ASRC . Include buffered clock out to slave a transport for those who don't want ASRC (or want to run it synchronus....
)
Each and every power supply should be possible to regulate individually and by choice, basically just a ground and supply entry point on the board.
For IV I would be happy with solid ground and signal exits from the pcb.
When it comes to SMD there are really no alternatives and I think whoever makes the pcb should make the most of everything including using smd as much as possible.
If there are to be any options on the PCB I would rather have them on the spdif reciever side and I/V conversion side.
On the reciever side there could be a 24/96+ capable DIR, an option for a ASRC that could be bypassed, perhaps a "tent type"~can clock layout on the board with close and direct connections, through individual single gate buffers, individually ps supplied to bitclock on the DAC's, MCKL on the filter and the ASRC . Include buffered clock out to slave a transport for those who don't want ASRC (or want to run it synchronus....
) Each and every power supply should be possible to regulate individually and by choice, basically just a ground and supply entry point on the board.
For IV I would be happy with solid ground and signal exits from the pcb.
When it comes to SMD there are really no alternatives and I think whoever makes the pcb should make the most of everything including using smd as much as possible.
more encouragement
I dont think I've shown this screendump before but its from a dual diff 1704K prototype dac, actually hardwired using the DF1704, passive balanced IV and measured using some serious averaging. The signal is taken at a (I think+28dB) possible +32dB +- 50v poweramp output.
As some of you might know the human ear is pretty good at averaging so if you can play loud enough and have the material you might just hear a -129dB signal using the PCM 1704.
I dont think I've shown this screendump before but its from a dual diff 1704K prototype dac, actually hardwired using the DF1704, passive balanced IV and measured using some serious averaging. The signal is taken at a (I think+28dB) possible +32dB +- 50v poweramp output.
As some of you might know the human ear is pretty good at averaging so if you can play loud enough and have the material you might just hear a -129dB signal using the PCM 1704.
Attachments
Balanced or not it is going to be a truly hi-end DAC and I am up for it.
I don't know if it's possible to make over-sampling and reclocking on-off function with this DAC but this would satisfy everybody's taste.
Reclocking makes much sense to me because we can use any
CD transport and make it perfect. At least this is in theory.
Let's keep this project alive.
I don't know if it's possible to make over-sampling and reclocking on-off function with this DAC but this would satisfy everybody's taste.
Reclocking makes much sense to me because we can use any
CD transport and make it perfect. At least this is in theory.
Let's keep this project alive.
Banned
Joined 2002
Hold on a moment. I too have long been interested in this idea, but it seems to me that the "meat" in this thread is devoted to discrediting it.
I've believed, like many others, that the one way to squeeze more resolution out of CD technology would be to have two sets of DACs per channel, decoding the phase and antiphase signals, and sending them through a balanced output.... thus having each half of your push-pull amplifier driven by a complete SPDIF/DAC system.
That, I had thought, is what it would take to get me away from my PCM63 DAC -- with its parallel internal DACs and used with 8x oversampling and transformer-coupled output, I think it's audibly superior to more recent chips like the 1704. Short of a more advanced version of the PCM63 being released, I had thought that some kind of balanced system, as described in this thread, would be the only route to improvement.
But this thread seems to have shot down this theory. Jocko et al have argued that paralelling DACs in phase-inverted structure is no better than paralelling them in an in-phase structure -- and in fact can be worse, because of the difficulties in matching the phase/antiphase DAC sets.
Has anyone offered a plausible counterargument to this? I can't find it on this thread. It seems to have gone like this: 1) We'd really like to see a balanced 1704 DAC; 2) You know, it really isn't such a good design; 3) Yeah, but we'd really like to see a balanced 1704 DAC.
That said, I'd probably snag a PCB if you made one. At the very least it would be a conversation piece.
I've believed, like many others, that the one way to squeeze more resolution out of CD technology would be to have two sets of DACs per channel, decoding the phase and antiphase signals, and sending them through a balanced output.... thus having each half of your push-pull amplifier driven by a complete SPDIF/DAC system.
That, I had thought, is what it would take to get me away from my PCM63 DAC -- with its parallel internal DACs and used with 8x oversampling and transformer-coupled output, I think it's audibly superior to more recent chips like the 1704. Short of a more advanced version of the PCM63 being released, I had thought that some kind of balanced system, as described in this thread, would be the only route to improvement.
But this thread seems to have shot down this theory. Jocko et al have argued that paralelling DACs in phase-inverted structure is no better than paralelling them in an in-phase structure -- and in fact can be worse, because of the difficulties in matching the phase/antiphase DAC sets.
Has anyone offered a plausible counterargument to this? I can't find it on this thread. It seems to have gone like this: 1) We'd really like to see a balanced 1704 DAC; 2) You know, it really isn't such a good design; 3) Yeah, but we'd really like to see a balanced 1704 DAC.
That said, I'd probably snag a PCB if you made one. At the very least it would be a conversation piece.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.