What are the different types of DACs?
DACs can be divided into two groups: single-bit (aka Delta-Sigma) and multi-bit (aka R-2R or ladder). There are integrated circuit (chip), discrete, and FPGA versions of both.
Can all DACs decode all file formats?
No. Single-bit DACs are native to DSD and multi-bit DACs are native to PCM.
Single-bit DACs have to convert PCM to DSD before they can decode it.
Multi-bit DACs have to convert DSD to PCM before they can decode it.
This conversion from DSD to PCM or PCM to DSD is often done by an FPGA.
What is an FPGA?
FPGA stands for “Field Programmable Gate Array” which is an integrated circuit (chip) that can be programmed to perform a specific set of functions.
The versatility of FPGAs makes them very popular for things like converting digital music formats, upsampling, oversampling, and error correction.
What are the advantages of FPGAs?
FPGAs are programmable which makes them infinitely versatile. They can be customized and optimized to perform any number of functions. They allow smaller manufacturers to create unique products without the expense of developing specialized chips.
FPGAs can also be reprogrammed, so a high-end audio manufacturer can easily update their programming (aka firmware) and improve the performance of their products.
What are the disadvantages of FPGAs?
The problem with FPGAs is longevity: their programming will eventually get corrupted. If the original manufacturer no longer supports that specific product, it is unrepairable.
What is a single-bit Delta-Sigma DAC?
Single-bit Delta-Sigma and single-bit FPGA DACs interpolate rather than decode.
They use noise shaping modulation that operates at several times the sampling frequency of the original recording. They take in the bit stream faster than the music is being played, noise shape it, upsample it, analyze it, error correct it, and output what the algorithms believe to be the correct waveform.
What is a multi-bit R-2R ladder DAC?
R-2R ladder DACs are the only DAC typology that actually decodes a PCM bitstream.
R-2R DACs consist of a ladder of resistors where each bit in the digital word is passed through a corresponding rung on the resistor ladder. Each rung weights the voltage of the corresponding bit to the relative volume quantized during the recording process.
Simple and effective.
What are the advantages of single-bit Delta-Sigma DACs?
Single-bit DACs have a higher signal-to-noise ratio and higher linearity than multi-bit DACs.
Single-bit DACs also cost significantly less, are significantly smaller, have significantly higher efficiency, and have the ability to correct errors in corrupted bit streams.
That is why single-bit DACs have become the standard in modern consumer electronics.
What are the disadvantages of single-bit Delta-Sigma DACs?
Single-bit DACs are very sensitive to clock jitter because they modulate the time domain as opposed to the amplitude domain. Multi-bit DACs are less sensitive to clocking.
Since error correcting algorithms can’t tell the difference between the musician playing with a bit of style and a corrupted bit stream the music tends to sound more analytical and less emotionally engaging from single-bit than multi-bit conversion. What are the advantages of multi-bit R-2R DACs?
Multi-bit DACs have a simpler digital noise-shaping loop, less dithering, a much simpler analog smoothing filter, and are much less sensitivity to clock jitter.
Because of their simplicity multi-bit DACs tend to have better timing, articulation, musical flow, and emotional expression than single-bit DACs. This is also partly due to multi-bit DACs having no noise shaping or error correction algorithms.
What are the disadvantages of multi-bit R-2R DACs?
Multi-bit DACs have the potential to have linearity issues due to mismatching of the resistors in their ladder(s). Linearity issues are significantly worse in discrete R-2R DACs vs those where the resistors are laser matched on a single Silicon wafer.
Multi-bit DACs are also subject to zero-crossing distortion where the sign wave changes from positive to negative. Single-bit DACs don’t have this issue because they have a fixed amplitude and modulate through the time domain.
Are there more than one type of R-2R DACs?
Yes. There are single-ladder and multiple ladder “segmented” R-2R DACs. There are also hybrid DACs where part of the signal is decoded using R-2R and part using Delta-Sigma.
A true R-2R can only be achieved by laser matching a single resistor ladder on a Silicon wafer. To achieve greater than 20-bits multiple resistor ladder segments are used and then weighed and summed. Quite a bit more complex. Quite a bit less musical.
Because of their simplicity a single ladder R-2R DAC has the best musical flow, most harmonic coherency, and best emotional expression of any DAC typology.
What about other vintage R-2R DAC chips like the AD1865?
The AD1865’s data sheet states it “employs a partially segmented architecture.” Since it is not a true single ladder R-2R like the PCM58 it can never match its musicality.
The AD1865 DAC chip was created as a budget product with a lot of features in one package. It’s 18-bit, stereo, runs on +/-5V, and has both current or voltage outputs.
In its day the AD1865 was the budget alternative to the significantly better and more expensive 20-bit, mono, current output only, +/-12V power supply AD1862.
The AD1865 and AD1862 do not employ the same no-compromise Nichrome resistors as the PCM58 which gives the PCM58 a more neutral, natural, and transparent presentation.
What about other vintage R-2R DAC chips like the TDA1543?
The TDA1543 is a true single ladder 16-bit R-2R DAC chip that could run on a single +5V power supply with no external component parts. It was revolutionary in its day.
The TDA1543 was designed as a budget alternative to the much larger, much more expensive, much more complex 16-bit TDA1541, that required multiple positive and negative voltage power supplies as well as several external component parts.
Because of the compact size, low cost, and abundant availability of the TDA1543, and the simplicity of requiring no external parts and only a single positive voltage power supply, the TDA1543 gained popularity among DIYers and manufactures.
A common ultra-purist configuration was to use a bank of several TDA1543 chips in parallel without any output stage. Since the TDA1543’s had a voltage output and didn’t require any additional amplification stages or external parts it was revolutionary in its day.
The TDA1543 is an amazing sounding budget R-2R DAC chip, but it is no comparison to the more advanced top-shelf R-2R DAC chips like the TDA1541, AD1862, PCM58, and PCM63.
What about discrete R-2R ladder DACs?
Those so-called discrete R-2R ladder DACs are not true R-2R but rather a segmented FPGA compensated R-2R hybridization. Quite an amazing bit of engineering.
Both the technology and the sound of those discrete ladder DACs is halfway between that of a true R-2R and a Delta-Sigma DAC.
Quite nice sounding but far from the coherency and expressiveness of a single ladder R-2R.
Aside from digital conversion what else effects musicality?
To achieve world-class musical reproduction from a digital signal you need a power supply that is low enough in noise to reveal low-level details and responsive enough to neither lag nor overshoot during dynamic passages (see our power supplies FAQs ).
Component parts, shielding, and anti-resonance also play an important role in minimizing coloration and preventing outside influences from degrading a DAC’s performance.