Measurement of R2R DAC

[March Audio]

Thanks. I would really like to see nearly full scale 12 Khz tone/J-Test with FFT length of 32K so it matches my measurements. The ones shown use much longer FFTs which artificially lowers the noise floor a lot.

As it is, the distortion spikes are quite high compared to even cheap DACs I have test.

Sure. Im away at the moment but Im sure my friend who now has the board will be quite willing to let me take more measurements. Probably next week

 

[March Audio]

Don’t R2R DACs typically require slightly different measurements to find the pimples? What would be the best measurement to show glitch errors? Also, aren’t the NOS R2R DACs rolled off in HF?

As Don says glitches would show up in the time domain. I never noticed any in the measurements I took. You can see a -100dB sine above. EDIT also found this comment from the designer
The advantage of the sign magnitude DAC is that the MSB bits are not switching if there is no need to, so you don't have high levels switching noise going into low level signals more than absolute necessary, in contrary to a regular DAC where the MSB are always switching as the signal crosses zero.

The board oversamples to 3 MHz and has a range of selectable filters. Roll off isnt therefore a problem. You can even design and load your own, which was my primary interest in the board - to use as an active crossover for my speaker design, I am agnostic to the architecture beyond that. Unfortunately despite sounding good, the fpga simply did not have the power to do fir filtering.

Massive diyaudio thread here
http://www.diyaudio.com/forums/vend...iscrete-2r-sign-magnitude-24-bit-384-khz.html
And
http://www.diyaudio.com/forums/digital-line-level/269776-filter-brewing-soekris-r2r.html


The dam1021 is a DAC module based on a discrete R-2R sign magnitude DAC design, with FPGA based FIFO buffering/reclocking and custom digital filters, < 1 ps jitter clock generator, with 28 bit resolution so there is headroom, oversampling up to 3.072 Mhz. Up to 24 bit / 384 Khz input from SPDIF, I2S and USB (via USB to I2S interface board), with isolation on the I2S interface. The board is very flexible, with digital volume control and filter parameters that can be downloaded. The board is fully firmware upgradable over a simple serial connection, which enable new features later on. Now full support for DSD when using firmware vers 1.05 or newer.

Product Name dam1021-12 dam1021-05
THD @ -1db 0.006% 0.015%
THD @ -60dB 0.03% 0.09%
Resistor Precision 28 bit, 0.01% - 0.02% Resistors 28 bit, 0.05% Resistors
Clock Jitter RMS 0.8 pS typical
S/N 20 Khz Bandwith 127 dB unweighted
Frequency Range +0.1dB -1.0dB 20hz - 20Khz
Input SPDIF Coax and Toslink, Up to 24 bit / 192 Khz
Input I2S PCM Fully Isolated, Up to 24 bit / 384 Khz
Input I2S DSD Fully Isolated, Up to DoP128 and DSD256
Output unbuffered Single ended 1.4V RMS, Zout 640R
Output buffered Balanced 4.0V RMS, Zout 20R
Power input 7-8V AC or +-7-15V DC max 5W
Board size 3.2" x 5.8" (81 x 147 mm)

 

[Wayne]

Does anyone know what year/decade were R2R DACs at their peak or most popular?

Thanks

 

[DonH56]

I'd guess the 1980's into the 90's. By then single-bit and then multibit delta-sigma designs had started to take over as digital processing technology advanced. There were some prized conventional chips that survive today though are not used as much AFAIK. However, most high-speed ADCs and DACs used in RF and wideband systems are conventional'ish.

 

[March Audio]

I know it doesn't change what is being measured Don. I am asking for the shorter FFT with the same source signal so that his graphs can then be compared to my DAC measurements without hand compensation for the difference in FFT size. They larger FFT also visually makes things look better even more than mine than they are.

Hi,

This is one thing I have been meaning to mention for some time. Shall we establish set of parameters fo measurements posted, easier for everyone if we are comparing apples with apples.

 

[amirm]

Yes for sure. I plan to write an article on my measurements and I will outline all of this. We had also talked about putting together the same test tones for all of us to use. More on my TODO list when I get back from CES 2018.

 

[Jakob1863]

Yes I agree, but imo fwiw, I did not find it to be a noticeable problem. In the real world 0.006% THD is still low.
For info I was performing subjective comparisons to a Chord Mojo and Meridian Explorer 2. I liked ( note I say liked not stating better) the Mojo then soekris then Meridian.

View attachment 9954

I just realize that in your graph ARTA already provided a calculated THD, and that 0.024% number in the same ballpark with what i had in mind (it depends on the phase alignment, but ARTA did know that part from the internal calculation), so due to the higher harmonics it is supposedly not as harmless as the same number only consisting of first and second harmonic would be.

Don´t worry about the subjective part; i assume that readers know that a subjective comparison is exactly that even if you write at the end "better" instead of "like it more"

 

[March Audio]

I just realize that in your graph ARTA already provided a calculated THD, and that 0.024% number in the same ballpark with what i had in mind (it depends on the phase alignment, but ARTA did know that part from the internal calculation), so due to the higher harmonics it is supposedly not as harmless as the same number only consisting of first and second harmonic would be.

Don´t worry about the subjective part; i assume that readers know that a subjective comparison is exactly that even if you write at the end "better" instead of "like it more"

Should be making some further measurements next week on the device and we can see if that tallies with the designers figures which are 0.015% for that board. The 0.1% resistor board is 0.006% for -1dB.

I do worry about the subjective part as so many implied "difinitive" statements are made in forums regarding uncontrolled nebulous and unverifiable subjective opinions.

As an aside, and its for another thread really, do you think you can can hear the effect of a spray of harmonics 100dB down in a music track?

 

[Jakob1863]

Should be making some further measurements next week on the device and we can see if that tallies with the designers figures which are 0.015% for that board. The 0.1% resistor board is 0.006% for -1dB.

Will be interesting to see....

I do worry about the subjective part as so many implied "difinitive" statements are made in forums regarding uncontrolled nebulous and unverifiable subjective opinions.

Understood, but me thinks, if somebody writes without further details that he did a subjective comparison, that i should take it as an informal approach but not as a controlled listening test.

As an aside, and its for another thread really, do you think you can can hear the effect of a spray of harmonics 100dB down in a music track?

Always tricky to answer such a question without more detailed information about the case, but under usual circumstances my hypothesis would be that i can´t hear it.
In a specific case it will depend, as masking theory would suggest, because spray of harmonics is a bit to vague. Masking region of the fundamental is more spread out frequencywise above than below, but if the spray of fundamentals is extended beyond that region and is not high in frequency to leave the sensitive region than it can be audible although one would think "100 dB down" can´t be.
This masking effect was imo the reason why (wasn´t it D.E.L. Shorter? again) in the 1970s people already stated that higher harmonics should always be fall in level compared to the lower harmonics, preferably in an exponentially way.

Why did you specially ask for the "100 db down" case?

 

[March Audio]

Because if you look at the plot from the designer for -1dB the harmonics are about -100dB.

 

[dat yeti]

For testing a recently made r2r dac, I would suggest the combo dac/headphone amp r2r-11 from audio-gd as it would be about $450 usd tops, shipping included. Its also their cheapest r2r dac from what I can tell. Also, their nfb-11 has pretty much the same hype train.

http://www.audio-gd.com/Products-EN.htm

Audio-gd has pretty high hype from everywhere except sbaf so there might be something to their products, and maybe their r2r-11 unit as well.

 

[amirm]

I noticed that the other day. I love their honesty:

Sounds like they have no mitigation for glitches and hence will suffer the same fate as Schiit multibit products.

Since it is a bit pricey, let me wait on that a bit as I am purchasing the RME AD-2 Pro ADC/DAC.

 

[DonH56]

Hmmm... I'd dispute the advantages in their list but it's not my design...

BTW, it is hard to design a de-glitcher that doesn't add glitches of its own. Usually much easier to make sure the glitches the DAC generates do not add distortion then filter them out.

 

[amirm]

Hmmm... I'd dispute the advantages in their list but it's not my design...

Yeh I just focused on the weaknesses .

 

[Superdad]

Sounds like they have no mitigation for glitches and hence will suffer the same fate as Schiit multibit products.

Please don't tar the venerable PCM1704 with the feathers of Schiit's weird AD5791!

 

[amirm]

Please don't tar the venerable PCM1704 with the feathers of Schiit's weird AD5791!

Oh, I thought they had a discrete implementation. Is it not?

 

[Superdad]

Oh, I thought they had a discrete implementation. Is it not?

Oh, if you are referring to the Audio-Gd line, they offer some DACs with PCM1704 and recently, since PCM1704s are getting scarce and expensive, some with discrete resistor ladders (which I suspect won't test near as well as the sign-magnitude architecture of PCM1704 and Soekris).

What I think all your DAC tests fail to take into account is the great variation in SQ between the I-V and analog output stages of various DACs. If you think a quick and cheap op-amp will sound as good as say a discrete output stage in an Ayre or Berkeley DAC, well then...
I recognize and accept that many here at ASR believe that your distortion and jitter measurements tell the entire story audio quality. Even setting aside high-end audio manufacturers, I don't think that pro-sound/studio equipment firms would agree.

But anyway, have a great weekend all. We are getting our first low-elevation snow of the year here in the Sierra Foothills. Quite lovely!

 

[Cosmik]

If you think a quick and cheap op-amp will sound as good as say a discrete output stage in an Ayre or Berkeley DAC, well then...

Probably...

 

[Purité Audio]

Almost certainly,
Keith

 

[Rthomas]

I noticed that the other day. I love their honesty:

View attachment 10869

Sounds like they have no mitigation for glitches and hence will suffer the same fate as Schiit multibit products.

Since it is a bit pricey, let me wait on that a bit as I am purchasing the RME AD-2 Pro ADC/DAC.

Hi Amir,

Good to hear that you're getting the RME. I'm waiting for your measurements before deciding on an RME ADI-2 DAC or Soekris 1541. I have no idea what to choose!