Soekris DAM1021 R2R DAC Measurements

[Superdad]

Burr Browns last R-2R Sign Magnitude chips are by many regarded the best sounding ever made, but are not manufactured any more...

It is my understanding that TI still sets up and runs the PCM1704 about once every 20 months. It is a pain for them as the fab equipment required is now quite old. And the resulting parts are no longer cheap (about $71 each).
Running a quad of them with no hardware-constrained OS filters, and feeding them 768KHz (SRC done on i7 computer) sounds wonderful.

Of course the above comments are not meant to take anything at all away from Soekris' excellent modern offering of sign-magnitude R-2R ladder DACs at very affordable prices with volume control built in. Now that he is offering them in full conceived products (I never had time for his first DIY boards), there is a very good chance I may spring for one this year. Speaking of spring, the Holo Spring (which I do happen to own) is another interesting R-2R DAC. It is not sign-magnitude style, but it does have a separate ladder network compensating for resistor imprecision. And it has a completely separate resistor ladder for DSD. Can't say I full understand it.

 

[Wombat]

Thanks for the nice words. This place is interesting, seems to be at a high technical level, it seems to be too much focused on what you can measure....



Not between the positive and negative itself, that would just result in increased even harmonics.... The distortion come from general mismatch between the resistors, and will improve with more precise resistors, although the ones I use are typical much better than the max specs....

Methodical subjective listening assessments are a form of measurement albeit lacking precise accuracy and consistency for wider application.
Conducting these tests is beyond the capacity of most forums. Manufacturers are better placed in this regard but mostly don't do it or let on that they do.

This forum covers subjective topics, e.g.

https://audiosciencereview.com/forum/index.php?forums/audio-reference-library.30/

https://audiosciencereview.com/forum/index.php?forums/psychoacoustics-science-of-how-we-hear.20/

And in other discussions.

 

[March Audio]

Thanks for the nice words. This place is interesting, seems to be at a high technical level, it seems to be too much focused on what you can measure....



Not between the positive and negative itself, that would just result in increased even harmonics.... The distortion come from general mismatch between the resistors, and will improve with more precise resistors, although the ones I use are typical much better than the max specs....

Well the technical focus is a rationale of the forum. Myself and others have too much experience of the frailty of subjective opinion. Personal subjective opinion can be found on any forum and ultimately it is fairly useless information. Rarely performed with any control, it is riddled with bias and often plain simple vivid imagination.

Whilst correlation between measurement and subjective opinion is extremely difficult, I don't subscribe to the view of some that we can't measure things that are heard. I often find the converse is true where we measure and get concerned about minutia that are below any audible impact. Having said that I see no reason to buy a DAC with demonstrably serious issues like Schitt ones recently measured. Measurements have allowed us to cut straight through the marketing BS.

 

[DonH56]

Thanks Don, what Im really getting at here is there any reason to think the sign magnitude design is inherently immune from low level linearity errors. Should you always expect if you input a -138dBfs signal it will output precisely that. This seems to be the implication from Soekris.

Edit, just seen your reply above. So essentially there still are multiple factors which prevent ultimate linearity. This was my understanding. My pleading ignorance was really more to encourage discussion of this subject. Sorry

No worries, I don't understand why the apology?

I really want to avoid stepping on @soekris but in general no, sign magnitude design does not obviate low-level linearity issues that can arise from a multitude of sources. What it does, and I believe he has stated, is that it can solve the zero-crossing discontinuity problem.

I have a lot of DAC model and measured data, somewhere, might have to dig it up sometime. I thought I had a slide set I gave at a conference years ago but the slides used a program I no longer have and is apparently no longer around (AmiPro). And I am not sure I even have the actual files anymore... I may still have the Matlab program but a lot of that was lost a few years ago (HD crash, and turns out had been going bad so the backup was also corrupted, grrr...)

 

[amirm]

Hate to put you to work Don , but we badly need one of your tutorials on ladder DACs. As a DAC (silicon) designer, you are uniquely qualified to write such a thing.

 

[DonH56]

On delta-sigma DACs: There are a lot of subtle things about the beasts. Bear in mind I have not piddled with them for a while now, and I was more focused on RF applications (last one I piddled with targeted 5 - 10 GS/s modulator rates). But the basics are all there...

LF (i.e. audio and instrumentation) delta-sigma DACs may use multi-bit DACs to reduce the complexity of the digital modulator and reduce the required oversampling ratio (i.e. they don't need to be clocked so fast). At RF this is a bigger deal as high-speed multipliers are a royal PITA and take a lot of power. I am not all that familiar with the audio variety (just have not looked recently; it has been years since I designed one) and so I do not claim to know what is currently in vogue. The catch with multi-bit DACs in a delta-sigma design is that, even if it is only say a 5-bit DAC, the steps and linearity must meet the desired output requirements. That is, if you are building a 16-bit delta-sigma DAC using a 4-bit DAC at the output of the modulator, that 4-bit DAC must have 16-bit linearity. So even though it only has 16 steps (2^4), those steps match match to 1/2^16 (1 part in 65,536) to create a 16-bit DAC. That usually means trimming, calibration and compensation, or some sort of dynamic element matching scheme (van de Plassche presented the latter in papers and a book on data converters; I actually piddled with it for a while).

Idle tones happen when the finite filter length in a delta-sigma modulator or demodulator repeats the same values endlessly. This was a big problem when folk started using DS ADCs for precision DC and LF measurements -- the cycling of filter values causes fairly large spurs in the output spectrum. Bad at audio, natch. Fortunately higher-order loops, multibit designs, and adding dither pretty much gets rid of them. But there are other concerns, including keeping those very high-speed clock signals and switching spikes away from the output (fried tweeters, anyone? Not to mention driving an amplifier into instability with ultrasonics...) That is in addition to the basic issue of noise shaping; all the quantization noise pushed out of baseband (audio band) by the delta-sigma modulator ends up at HF where it must be filtered away from the system. The scheme is to push it high enough that noise filtering isn't too hard, and the high oversampling ratio (clock frequency) means images (at multiples if half the sampling frequency) are more easily filtered. In practice, when talking about 16-24 bits and more, it takes some pretty fancy filters. A sneaky problem is noise modulation and this is a common concern for audio DACs. Essentially the signal interacts with those digital filters to modulate the noise floor. These creates "humps" in the noise floor for steady-state signals, and noise modulation looks like modulation or dynamic "pumping" of the noise floor with time-varying signals (e.g. music). It may audible; I do not know. And I mean that sincerely -- I do not know how audible the effect. It can be large enough that I suspect so with the right signal and DAC.

I'll stop there, hard to write decades or research in a few paragraphs, and while I have designed and worked with them in the past, I'd have to bone up a bit to be a competent source. Most of my designs were more conventional DACs (segmented ladder types), and while I have designed and built 16-bit DACs from 5 to 100+ MS/s 10+ years ago, I have not really done much with audio DACs.

 

[DonH56]

Hate to put you to work Don , but we badly need one of your tutorials on ladder DACs. As a DAC (silicon) designer, you are uniquely qualified to write such a thing.

You know, that would be interesting, but my work week is usually 60+ hours lately and we have some family stuff going on. Let me think on it... In the meantime, the basics are outlined in the sampling and DAC threads, covering conventional and delta-sigma designs, that (I think) you pulled from WBF. It has been quite a few years now, but I think one of the missing pieces is a more detailed discussion of switching glitches and ladder bow and other nonlinearities. Have to look through what you pulled over, or dig up my copies (assuming they weren't lost in The HD Crash a few years back).

Must note they weren't all Si, however... Had to throw in InP, GaAs, and SiGe just for fun.

 

[amirm]

Must note they weren't all Si, however... Had to throw in InP, GaAs, and SiGe just for fun.

You should build one out of GaAs for audio and could make millions!!! We clock it at 1 Ghz externally and tell people everything else is pedestrian.

 

[DonH56]

Piece of cake. Put in a 1 GHz GaAs or YiG oscillator with a divider down to a few MHz to clock a regular old audio DAC and market the 1 GHz GaAs technology. Should sell like hotcakes.

 

[amirm]

The trick would be to find a good excuse for me not to be able to measure it!!! I guess I could say it is too fast for my gear....

 

[March Audio]

OK, as mentioned previously, the last thing we want to do here is put up inaccurate measurements so I have taken a look at the process and taken a second set. I have made 5 changes.
1. Measured the single ended output instead of balanced
2. Moved the test frequency to 2kHz where there are no mains or other spuria
3. Limited the measurement bandwidth to 1990 to 2020Hz - see the RMS L dBr value at top of plot.
3. Increased the FFT resolution to 262144 points - this was not about resolution but taking a longer sample to minimise the effects of random noise which is becoming significant at the lower levels
4. Make sure it averaged 50 readings readings (although it states 50 on the plot I was stopping earlier after the FFT had visually stabilised.

Note its flat top window so ignore the spectral spread.
Currently I cannot track down the cause of the excessive 50Hz pick up, however there are no harmonics at the 2kHz test tone. Being a DIY build, although I cant see anything fundamentally wrong, there is room for problems such as this hence why I am very keen to see one of Sorens finished DACs measured.

Soren, how significant do you consider thermal effects on the ladder?

 

[March Audio]

For comparison Motu 8A single ended output
note that the 8A may be considered to have have a slight advantage as its 0dBfs output voltage is 5v whereas the 1021 is 4 volts

I will post a more comprehensive set of measurements of the Motu 8A in a new thread.

Myself and Amir will perform the AP v QA comparison. My current view after looking at the spec sheet for the ADC is that I dont trust the QA beyond 21-22bit for this specific linearity test.

 

[soekris]

Soren, how significant do you consider thermal effects on the ladder?

I use resistors with very low temperature coefficients, but thermal effects is not really an issue as voltage out from a R-2R network is a ratio between resistors, and the resistors track each other pretty good, and the power dissipated is so low so self heating is also not a issue.

One big difference between the QA401 and the AP when measuring low levels, is the the QA only have a simple two step attenuator (2V RMS and 20V RMS ranges) while the AP have many steps. That mean when measuring low levels, the QA have to use the whole range of its ADC, while the AP use can have its ADC stay in its most linear range as it can add gain as needed for low level signals....

 

[March Audio]

This is a good point, although it has to be said the es9038 board still managed to measure well with a 2 volt FS output. Amir does the AP always gain change through the linearity measurement? If so do you know the ranges?

I do have a THAT diff amp board Im evaluating which could be an ideal inst amp.

 

[Jimster480]

Now we're moving into subjective statements....

I, as I sell only R-2R DACs, believe that Delta Sigma DACs have all kind of side effect, some large and clear, like idle tones, other smaller, all a result of the modulation and noise shaping process the they need to do to get something useful out of those one to five bits DACs. Those tings are hard to measure while simple signals, like a Sine, is much easier for a Delta Sigma DAC to handle and result in those nice numbers, while there isn't really any good way to measure with real music, except your ears....

While the R-2R DACs are pure, directly converting from the input PCM data to Audio output, without needing to do any tricks at all, resulting in a more natural sound. It's not without reason that some of the considered best DACs in the world are R-2R types....

Its true that they do have noise shaping, but with modern chips running at such high clockspeeds; it could be possible for any DS chip to run at 100-1000mhz allowing it to do hundreds if not thousands of passes on every sample in real time.

The R2R design is definitely simpler because there isn't always a software aspect involved in shaping the sound as the passes are made but that doesn't necessarily mean that the results come out differently.

 

[DonH56]

The limitation in DS converter clock rates is not really the logic, it is closing the loop and (mostly) getting the output stage to slew and settle. There are plenty of schemes to speed up the loop(s) but hard to get the analog input or output to settle to umpteen bits no matter the clock rate. A DS converter (ADC or DAC) still has a few circuit blocks that require full N-bit precision for an N-bit converter. The noise shaping loop does not obviate the need for a few good cells. As an analog guy that makes me happy since I remain employed, natch.

There are fundamental differences in architecture and the way the circuits operate that mean an R2R and DS design are going to have different performance and spec trades. Noise modulation is rarely an issue for an R2R DAC (it happens but in a different way) and DS DACs don't generally have to deal with large nonlinear switching glitches every time you step levels. They are pretty different beasts... Whether the results come out audibly different is, as always, a hot topic.

 

[Rene]

Now we're moving into subjective statements....

I, as I sell only R-2R DACs, believe that Delta Sigma DACs have all kind of side effect, some large and clear, like idle tones, other smaller, all a result of the modulation and noise shaping process the they need to do to get something useful out of those one to five bits DACs. Those tings are hard to measure while simple signals, like a Sine, is much easier for a Delta Sigma DAC to handle and result in those nice numbers, while there isn't really any good way to measure with real music, except your ears....

While the R-2R DACs are pure, directly converting from the input PCM data to Audio output, without needing to do any tricks at all, resulting in a more natural sound. It's not without reason that some of the considered best DACs in the world are R-2R types....

I, too, agree with Soekris that R2R dacs can sound more natural than sigma-delta dacs. Without getting into conjecture about why this so, or what the sigma-deltas get wrong, as no one has been able to give me a good technical explanation, I will say that one of the goals we strove for at Berkeley Audio Design was making the Alpha dac with its sigma-delta converter sound as close to the Pacific Microsonics Model 2 dac as possible. It never quite made it. There is an openness and solidity to the PMI dac, especially, to my ears, in the bass region, that we could not duplicate.

 

[Rene]

The limitation in DS converter clock rates is not really the logic, it is closing the loop and (mostly) getting the output stage to slew and settle. There are plenty of schemes to speed up the loop(s) but hard to get the analog input or output to settle to umpteen bits no matter the clock rate. A DS converter (ADC or DAC) still has a few circuit blocks that require full N-bit precision for an N-bit converter. The noise shaping loop does not obviate the need for a few good cells. As an analog guy that makes me happy since I remain employed, natch.

Yes, Don, your explanation sheds light. Even if the dac chip itself is well behaved, there is still the challenge to the following I-V conversion or buffer stage to be well behaved in the presence of all the out-of-band digital noise.

And if you can get the signal through that stage properly, a well behaved analog filter is a must.

 

[svart-hvitt]

I, too, agree with Soekris that R2R dacs can sound more natural than sigma-delta dacs. Without getting into conjecture about why this so, or what the sigma-deltas get wrong, as no one has been able to give me a good technical explanation, I will say that one of the goals we strove for at Berkeley Audio Design was making the Alpha dac with its sigma-delta converter sound as close to the Pacific Microsonics Model 2 dac as possible. It never quite made it. There is an openness and solidity to the PMI dac, especially, to my ears, in the bass region, that we could not duplicate.

Aren’t you suggesting here that R2Rs have a sound that measurement tools can’t...measure?

If so, there is a sound aspect in R2Rs that need to be heared, not measured?

Which again raises the question: Why measure DACs if measurements are below hearing limit, and there are audible aspects that cannot be measured?

This would leave blind testing the only valid process for testing high-end audio gear.

Agree?

 

[Rene]

Aren’t you suggesting here that R2Rs have a sound that measurement tools can’t...measure?

If so, there is a sound aspect in R2Rs that need to be heared, not measured?

Which again raises the question: Why measure DACs if measurements are below hearing limit, and there are audible aspects that cannot be measured?

This would leave blind testing the only valid process for testing high-end audio gear.

Agree?

Agreed.