luka

Great videos. I love analog computers - so much more elegant than all that thrashing about with 1s and 0s. If you ever get the chance to visit the Computer History Museum, they have a fantastic collection: Nordsieck My favourite, though, is the humble lens. A 2D Fourier transform at the speed of light: An Intuitive Explanation of Fourier Theory

Op-amp Varieties

Thanks for the clarification. I missed the statistical reference in your previous post - must have been added after my reply.

You mean dynamic element matching? That just minimises the impact of mismatches in levels. Whatever, it doesn't gain you exactly 3dB in THD every time you double the number of converters.

SNR improves when using multiple converters because the noise is each converter is generally uncorrolated whilst the signal is not. The same is not true however for THD as distortion products are correlated with the signal.

When I was referring to multi-bit, or multi-level, I was talking about the number of bits/levels in the SDM quantizer, the two being related by: bits = log2(levels). To me the two terms are interchangable since one can be derived from the other. DACs that use dynamic element matching may encode the modulator output to some other number of levels, but that is not what I was referring to. Why does 'proper' have to equate to large amounts of heat? Modern process geometries can deliver huge amounts of processing power without the need for heatsinks. In hardware you can have large numbers of highly-specialised processes running simultaneously, so you don't always need a super-high clock rate. Really poor in what way? Are you refering to the modulator or the upsampling algorithm (or both)? 14dB is quite a lot though, isn't it? I don't really understand what you are saying here. I guess you could sign-extend the DSD bitstream to fit any word length, but I don't think that's what most multi-bit DACs do. You could certainly use some of the 1-bit output elements to make some sort of analog FIR output filter. This would at least get rid of some of the quantization noise, but your 'perfect' DSD impulse response will now be replaced by the impulse response of the FIR filter. Well, if the TI data sheets are to be believed, the multi-bit output of the PCM4222 has lower noise AND lower distortion compared to the 1-bit output of the PCM4202. True enough, but many DACs have an 8fs input that allow most of the upsampling to be bypassed. What we need is a DAC equivalent to the PCM4222 that has a direct interface to the modulator =) Yes, but the DSD64 analog filter also has to filter out the shaped quantiztion noise, so it probably ends up being even steeper than the PCM image filter, especially if SAH stages are replaced by proper filters. Nice plots. Not much between them in terms of out-of-band noise. What DAC was this from? I guess the point I'm trying to make here is that 1-bit sigma-delta no longer represents the state-of-the-art in either DAC or ADC design, so why introduce a 'new' format that is based on out-of-date technology. I can see why HiFi companies and specialist record labels would like DSD it as it gives them a chance to differentiate their products and sell more kit, but it is not really pushing the envelope over what can currently be achieved by 192kHz/24bit PCM. As music consumers, we should be much more concerned with the quality of the recording/production/mastering process and less concerned with what is essentially 'packaging'. Enjoy the music!

Not sure I agree with your reasoning here Miska. A 1-bit SDM is not difficult to do in silicon. As I understand it DAC manufacturers were initially attracted to 1-bit SDM because it meant they could cut costs without sacrificing performance - standard CMOS process vs laser trimmed BiCMOS. However, they soon discovered that 1-bit SDM comes with it's own limitations. The main problem with 1-bit SDM, or any 1-bit format such as PWM, is jitter sensitivity. Any variation in edge timing results in an error that spans the entire output range. Timing jitter is akin to frequency modulation so the jitter signature modulates the output of the SDM which leads to shaped quantisation noise being mixed down into the audio band. It therefore becomes very difficult to achieve a noise floor that approaches that of the best multi-bit converters. The attraction of 1-bit converters, however, is that they are at least theoretically capable of perfect linearity. Unfortunately, this is impossible to achieve in practice due to the imperfect performance of real world transistors. Thankfully, DAC design has progressed significantly since the early days of 1-bit sigma delta. Multi-bit SDM DACs have low jitter sensitivity and lower levels of quantisation noise but they no longer have the inherent linearity of 1-bit systems. However, thanks to some clever processing it is possible to transform the distortion generated by multi-level linearity errors into noise which can then be shaped in much the same way as quantisation noise is shaped in a SDM. ADC designers have begun moving to multi bit too (e.g. PCM4222), again because it reduces the noise floor. Most 1-bit modulators can usually only be used up -6dB or so, so maybe the reduced distortion is due to reduced analog signal levels. The distortion of most modern DACs only gets above the noise floor in the upper few dBs. Regarding ultra-sonic output, for a PCM input DAC this really depends on how they get to the modulator rate. Some DACs have filtered upsampling beyond 8fs which gives improved performance over a basic sample and hold. Also, a DSD spectrum will have multiple images too, but both DSD and PCM noise/images can be dealt with easily with an appropriate analog filter. Me too!!!

Thanks Miska. I found a Cirrus patent that describes the process - basically using the multi-bit DAC elements as an analog FIR filter. Ingeneous.

I have no axe to grind. My first post was a genuine question to which I fully expected the answer to be yes. My second post expressed my disbelief in psme's response and was meant to be semi-humorous, as was my third (seriously, no Monty Python fans?). I build audio gear in my spare time and I have shamelessly copied many designs, so I cannot really claim any moral high ground, but it does give me a perspective which enables me to spot similar plagiaristic attempts. If no one else agrees, then fair enough, I'll leave it at that.

I don't think I have made any such accusations. I am just pointing out similarities as I see them. Also the reference design in the Wolfson DAC data sheet does not use transformers.

I have only the photographs I linked to earlier. I have tried to decipher the Klimax DAC circuit from photographs for my own homebrew DAC project. That's why the Lumin DAC circuit caught my eye. If you look closely, and with a bit of electronics knowledge, you will see many similarities.

Is the Wolfson Direct DSD mode really direct? The filter graphs in the data sheet make me think some processing is being done (not natural to have such sharp nulls)

OK, so apart from the DAC chips, the output transformers, the transformer drive circuit, the output relay, the connectors, and the case construction, what have the Romans ever done for us? they are completely different.

More than a bit similar I think . Have their lawyers been in touch yet?