Varieties of DACs

I think all of the past/present DAC "topologies" are as follows:

(1) Discrete (R2R, etc. Examples: Soekris, Denafrips, etc)

(2) ASIC (single chip, may also have digital filter inside. Delta-Sigma, Bit-stream/MASH, R2R, DEM, CC, etc. Examples: Philips TDA1541, AD1862, PCM63, AKM 4499, ESS9038)

(3) Firmware (FPGA. Example: Chord)

Other:

Parallel ASIC. Examples: Accuphase "multi-multi" ( 4 parallel PCM 63), diy: tda1543 http://www.diyparadise.com/8x1543dac.html)

Anything missing from this list?
 
(3) doesn't make sense to me, firmware can do some digital preprocessing but not the conversion from digital to analogue.

Besides fully discrete or fully integrated, you have designs that are a bit of both. For example the DSC2, with integrated circuits (not ASICs, but simple shift registers) driving discrete resistors.


The word "topology" makes me think more of the working principle than whether it is made out of discrete circuitry, ICs or a bit of both.
 
Last edited:
So do two of my DIY DACs, but always with some circuit after the FPGA to do the conversion.

(As an aside, digital designers usually regard configuring an FPGA as a kind of hardware design, but you could indeed just as well regard it as a kind of firmware design. It's in the grey area.)
 
  • Like
Reactions: Paddy Garcia
The TDA1541 is one of the first chips with dynamic element matching, although the TDA1540 was even earlier. It only used plain R-2R networks for generating the currents for the lower bits.
Right, in my OP, I mentioned DEM. And also CC (continuous calibration), which Philips seem to have followed up with.

Not sure about what the early Japanese CDPs used, like Sony. Anyone?
 
Last edited:
How would you categorize the DAC that's described in this?

https://linearaudio.net/sites/linearaudio.net/files/03 Didden LA V13 mvdg.pdf

It's a DAC using an FPGA to convert the signal to a single-bit format by sigma-delta modulation, either straightforward single-bit sigma-delta modulation or quasi-multibit with an embedded pulse width modulator, followed by discrete components and standard logic ICs for level shifting, and then a discrete valve circuit for the actual conversion from digital to analogue.
 
Mentioning the 16/44 Sony "CX" DAC series makes one wonder why stand-alone DAC projects -- diy or otherwise -- did not source CX dacs?
Did Sony tightly control their distribution (unlike, say, Philips whose TDA dacs were found in several non-Philips devices)?
I did see a few Sony CX dacs for sale on eBay. DIP size. Not sure whether anyone here at diya has experimented with them.
 
FPGA dacs:
I looked at the videos you linked to. According to the second one, the Chord DAC uses an FPGA for digital signal processing, but standard logic flip-flops for digital-to-analogue conversion. That makes sense, you can then give the flip-flops a clean supply and clock.

The animation at the end gives me the impression they use pulse width modulation (I guess embedded in a noise shaping loop/multibit sigma-delta modulator) and some sort of FIRDAC.
 
Last edited by a moderator:
In my opinion, the most basic classification needs to come from the fact that a DAC is a system (mostly on a chip) to most technical people, but is a "box / appliance" like a refrigerator / radio / CD player, to most non-technical people.

Then, of course there're the variations within each kind, as already listed my many others above.
 
Mentioning the 16/44 Sony "CX" DAC series makes one wonder why stand-alone DAC projects -- diy or otherwise -- did not source CX dacs?
Did Sony tightly control their distribution (unlike, say, Philips whose TDA dacs were found in several non-Philips devices)?
I did see a few Sony CX dacs for sale on eBay. DIP size. Not sure whether anyone here at diya has experimented with them.
There is a preference for low hanging fruit.