Hi everyone,
I have tried to find some info on this subject, but either to the poor keyword choice or by my own sloppiness, I haven't been able to find specific information.
So, sorry if this is obvious, redundant or trivial, but I have to ask:
I am contemplating my next build, and one image keeps returning: no "digital" stuff after DAC.
I like the idea of transformers, and am working towards winding my own, so custom transformers should be considered available.
Two main "problems" with this approach, as it seems, are:
- low signal level on the DAC output
- challenge to meet wide bandwith requirements along with others (SNR etc)
The bandwidth requirement can be loosened in a multi-way system by making dedicated DACs for single channels (x-over in digital domain before DAC).
This is what I intend. This should make transformer design a little easier.
The other problem is that the DAC (my first candidate for the moment is PCM1794) needs to see a reasonably low impedance to work properly, which leads us to very low output voltage.
The workaround for this would be to use more DACs (I am computer scientist, so this attempt of horizontal scale-out may not be appropriate in analog-field).
My naive idea is to make a transformer with one (in case of PCM1794 two, since balanced) primary winding per DAC (secondary would be as usual, depending on the design of the next stage).
Total ratio could be 1:1 or 1:2, maybe slightly more depending on the whole circuit requirements.
The winding ratio seen from a single DAC would be ( total winding ratio / X ) , X := # of DAC chips.
Effects I imagine:
- lower load seen by the DAC, corresponding to the # windings in one DAC Primary winding to secondary ratio
- seemles coupling of multiple DAC signals (summed by the transformer)
- other good stuff, DC isolation etc
The general idea is to get 2V line output capable of driving the next stage.
Since the DAC chips are reasonably priced (assume 6 channels X 4 DACs = 24, even for the expensive PCM1794 a managable amount, given the prices in this hobby).
Has anyone done this or something similar?
I still have to model it in LTSpice, but as a noob, I am grateful for any misconceptions you could destroy for me, or any affirmation to continue playing with this idea
Best regards,
Nikola
I have tried to find some info on this subject, but either to the poor keyword choice or by my own sloppiness, I haven't been able to find specific information.
So, sorry if this is obvious, redundant or trivial, but I have to ask:
I am contemplating my next build, and one image keeps returning: no "digital" stuff after DAC.
I like the idea of transformers, and am working towards winding my own, so custom transformers should be considered available.
Two main "problems" with this approach, as it seems, are:
- low signal level on the DAC output
- challenge to meet wide bandwith requirements along with others (SNR etc)
The bandwidth requirement can be loosened in a multi-way system by making dedicated DACs for single channels (x-over in digital domain before DAC).
This is what I intend. This should make transformer design a little easier.
The other problem is that the DAC (my first candidate for the moment is PCM1794) needs to see a reasonably low impedance to work properly, which leads us to very low output voltage.
The workaround for this would be to use more DACs (I am computer scientist, so this attempt of horizontal scale-out may not be appropriate in analog-field).
My naive idea is to make a transformer with one (in case of PCM1794 two, since balanced) primary winding per DAC (secondary would be as usual, depending on the design of the next stage).
Total ratio could be 1:1 or 1:2, maybe slightly more depending on the whole circuit requirements.
The winding ratio seen from a single DAC would be ( total winding ratio / X ) , X := # of DAC chips.
Effects I imagine:
- lower load seen by the DAC, corresponding to the # windings in one DAC Primary winding to secondary ratio
- seemles coupling of multiple DAC signals (summed by the transformer)
- other good stuff, DC isolation etc
The general idea is to get 2V line output capable of driving the next stage.
Since the DAC chips are reasonably priced (assume 6 channels X 4 DACs = 24, even for the expensive PCM1794 a managable amount, given the prices in this hobby).
Has anyone done this or something similar?
I still have to model it in LTSpice, but as a noob, I am grateful for any misconceptions you could destroy for me, or any affirmation to continue playing with this idea
Best regards,
Nikola
I'm, working towards doing something similar, using TDA1387 as the DAC. I reckon I'll need something in the region of 48 chips in parallel to get a decent output (each is good for around 1mA) level to drive transformers direct. PCM1794 has 8mA (but is considerably more than 8X the price of TDA1387) so by that reckoning your 4 DACs should do fine (I'm using passive filtering which results in 6dB loss).
Your problem (challenge) is ascertaining what output voltage is tolerated by the PCM1794 - there may well be devotees of this chip lurking to give the answer to that.
In practice a typical line level is a bit less than the CD standard 2V so I'll be aiming in my design for something in the region 0.5 - 1V.
Your problem (challenge) is ascertaining what output voltage is tolerated by the PCM1794 - there may well be devotees of this chip lurking to give the answer to that.
In practice a typical line level is a bit less than the CD standard 2V so I'll be aiming in my design for something in the region 0.5 - 1V.
abraxalito, thanks!
Assuming that the winding ratio from one DAC's primary is 1/4 of the secondary (4 DAC chips), can it be assumed that the impedance of the primary is 1/16 (1/4²) of the secondary impedance?
(I am not sure I understood this detail right)
For reference, this thread is all about performance of PCM1794 under various loads.
http://www.diyaudio.com/forums/digital-source/221743-testing-pcm1794.html
Assuming that the winding ratio from one DAC's primary is 1/4 of the secondary (4 DAC chips), can it be assumed that the impedance of the primary is 1/16 (1/4²) of the secondary impedance?
(I am not sure I understood this detail right)
For reference, this thread is all about performance of PCM1794 under various loads.
http://www.diyaudio.com/forums/digital-source/221743-testing-pcm1794.html
Last edited:
Have a look at the below linked project website and related diyaudio thread, which essentially covers what you hope to do:
DDDAC 1794 NOS DAC - Non Oversampling DAC with PCM1794 - no digital filter - modular design DIY DAC for high resolution audio 192/24 192kHz 24bit
http://www.diyaudio.com/forums/digi...-nos-192-24-dac-pcm1794-waveio-usb-input.html
DDDAC 1794 NOS DAC - Non Oversampling DAC with PCM1794 - no digital filter - modular design DIY DAC for high resolution audio 192/24 192kHz 24bit
http://www.diyaudio.com/forums/digi...-nos-192-24-dac-pcm1794-waveio-usb-input.html
I've been reading this document:
http://www.analog.com/static/imported-files/application_notes/AN_912.pdf
The right side of page 4 reads:
"Note that vA and vB are twice as large as vNORM and vCOMP. What
is the reason for the discrepancy? The answer lies in the fact that the two primary windings interact with each other. "
They are considering only one differential pair.
I am trying to figure out how primaries from more than 1 DAC would interact..
http://www.analog.com/static/imported-files/application_notes/AN_912.pdf
The right side of page 4 reads:
"Note that vA and vB are twice as large as vNORM and vCOMP. What
is the reason for the discrepancy? The answer lies in the fact that the two primary windings interact with each other. "
They are considering only one differential pair.
I am trying to figure out how primaries from more than 1 DAC would interact..
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.