Reference DAC Module - Discrete R-2R Sign Magnitude 24 bit 384 KHz

[Painkiller]

Believe also it worths the "money risk" as it is versatil with input/output and it can match your active crossover goals : with 3 boards you can have your treble/medium/bass with embeded croosover and personal EQ !

I'm hurry to compare it with your AD1862 which is certainly the best R2R chip ever maid (TDA 1541 is not exactly R2R ???), even better with its 20 bits than the PCM 1704 with its 23-24 bits (dryer sound for many !).

Ihmo with OEM chips the best DACchips are TDA1541/AD1862 : first, PCM63 (second, near the ad1862, but not exactly the same, its softer in upermid-treble). Third : pcm1704, AD1865,tda1545A,tda1543....PCM1702...

Yes, I'm actually in for 4 boards, just for versatility sake, in case I'd ever want to build a 4-way speaker system.

The higher order harmonics at 0 dBFS are actually lower with the AD1862 compared to the measurements in the first post in this thread. But at -60 dB the measurements of this discrete dac is the best I've ever seen. Can't miss this opportunity.

 

[spzzzzkt]

my clock board is a very normal clock, nothing of special, but I think this DAC cannot use any external clock.

And that is the problem.... your clock board is nothing special.

On the other hand Soekris has a R2R DAC design that is unique in the DIY market. That is why there is interest regardless of your opinion of the clocking scheme.

Your comments about Soekris's design sound like the whining of a jealous child who is angry that someone else has stolen the limelight.

 

[Superdad]

... single ladder

Why would you want a single ladder? The sign-magnitude architecture is THE key to the design! Without it we would probably get only about 10-bits of monotonicity. Resistors precise enough for 18 bits on a single R2R ladder don't exist on the shelf. My math is bad, but I think you would need something like 0.000000001% resistors for that to work. The sign-magnitude method makes it possible to get a reasonable 18-bits with 0.01% tolerance.

 

[andrea_mori]

And that is the problem.... your clock board is nothing special.

On the other hand Soekris has a R2R DAC design that is unique in the DIY market. That is why there is interest regardless of your opinion of the clocking scheme.

Your comments about Soekris's design sound like the whining of a jealous child who is angry that someone else has stolen the limelight.

I think you have a few problem to understand.

First: true, my clock "is nothing special", only 30 dB better than the SiLabs, but maybe you don't understand the difference.
Second: this is not a diy project, you can buy a finished board only, no schematics and so on. Do you think AMR dac are diy project? If yes, ok this is a diy project, otherwise this is not.
Third: my comments are related to this commercial product, exactly like comments about AMR or Wadia products.
And when someone is clearly wrong claiming SiLabs are low jitter oscillator, I have to correct him. No angry or jealous, simply "the datasheet". You can read yourself the Si514 datasheet to discover it's not designed for audio, and it's not a low jitter device.
There are tons of document on the internet that explains phase noise and jitter, I posted some of these several time, please take a look to understand.

Last but not least: my projects are really diy, every schematics are provided, I have no commercial interest.
I do this little thing for the community, not to gain money.

 

[gentlevoice]

@TNT: If you are about to use the NZ2520SDs you might be interested in contacting with this person:

analog.research on eBay

He has done extensive measurements on the NZ2520SDs and apparently there's quite some (up to 30 dBs) difference between the oscillators. To my knowledge he may also sell oscillators with a low (measured) phase noise value.

I know this may be a bit late mentioning this since you have ordered the oscillators ... but maybe of use/interest ...?

Cheers,

Jesper

 

[gentlevoice]

Now - not to "intrude" in a thread that is dedicated to another topic but just briefly:

@andrea_mori:

First: true, my clock "is nothing special", only 30 dB better than the SiLabs, but maybe you don't understand the difference.

Hmmm... it somehow surprises me that you say this, Andrea. When comparing the measurements you posted here:

http://www.diyaudio.com/forums/digi...se-jitter-crystal-oscillator.html#post4051205

... with the ones of the Pulsar OCXO it looks to me as if they are the same:

http://www.diyaudio.com/forums/vend...clock-ultra-low-noise-ocxo-3.html#post4089885

Same frequency (11.28 MHz) and close to the same performance. Makes me wonder why you think it's nothing special ...

.... ...

Well, a digression.

Jesper

 

[andrea_mori]

Why would you want a single ladder? The sign-magnitude architecture is THE key to the design! Without it we would probably get only about 10-bits of monotonicity. Resistors precise enough for 18 bits on a single R2R ladder don't exist on the shelf. My math is bad, but I think you would need something like 0.000000001% resistors for that to work. The sign-magnitude method makes it possible to get a reasonable 18-bits with 0.01% tolerance.

Sorry, but you are wrong. There is no difference in monotonicity between sign magnitude and other notation. Using 0.01 ladder resistors you can reach 13 bit precision (2^13 = 8192 - 1/8192 * 100 = 0,012%).
There is no "window" that change math analysis due to sign magnitude. When the 14th bit and upper bits switch, the precision decreases.
Using dual ladder you are doubling the error, that's the reason BB used a single ladder in its PCM1704.

 

[TNT]

I think you have a few problem to understand.

First: true, my clock "is nothing special", only 30 dB better than the SiLabs, but maybe you don't understand the difference.
Second: this is not a diy project, you can buy a finished board only, no schematics and so on. Do you think AMR dac are diy project? If yes, ok this is a diy project, otherwise this is not.
Third: my comments are related to this commercial product, exactly like comments about AMR or Wadia products.
And when someone is clearly wrong claiming SiLabs are low jitter oscillator, I have to correct him. No angry or jealous, simply "the datasheet". You can read yourself the Si514 datasheet to discover it's not designed for audio, and it's not a low jitter device.
There are tons of document on the internet that explains phase noise and jitter, I posted some of these several time, please take a look to understand.

Last but not least: my projects are really diy, every schematics are provided, I have no commercial interest.
I do this little thing for the community, not to gain money.

Andrea, i would back you up on all these aspects. You clock seems quite special!

//

 

[andrea_mori]

Now - not to "intrude" in a thread that is dedicated to another topic but just briefly:

@andrea_mori:



Hmmm... it somehow surprises me that you say this, Andrea. When comparing the measurements you posted here:

http://www.diyaudio.com/forums/digi...se-jitter-crystal-oscillator.html#post4051205

... with the ones of the Pulsar OCXO it looks to me as if they are the same:

http://www.diyaudio.com/forums/vend...clock-ultra-low-noise-ocxo-3.html#post4089885

Same frequency (11.28 MHz) and close to the same performance. Makes me wonder why you think it's nothing special ...

.... ...

Well, a digression.

Jesper

Jesper,

I meant nothing of particular at implementation level, maybe "special" was not the appropriate definition.
I meant that my clock can be implemented like any other oscillator such Crystek and so on, just fitting it close to the dac or using u.fl. cable.
Then, there are "commercial projects" like Ian's Fifo buffer that can accommodate external clock (you can also slave it from master clock), while other "commercial projects" like Soekris DAC can't accommodate any external clock.
As I said above, every choice has done by the vendor, so not a diy project.

 

[rogerchou]

Hi Guys here, please don't dance with him and treat him as air. In Ian's thread, no one gave him any response and then everything went better. Let's focus on Soren's design, maybe it is time for us to discuss digital filter or which output stage can match this R2R Dac better...

 

[Geekey]

Soekris could you please address the questions I recently asked .

1. In case I am not going to use DAC’s internal analog output stage , is it possible to power the rest of DAC’s circuit with just one single secondary winding of the transformer, or I still need two secondarys.

2. There was unanswered question[post #50] about how to disconnect the unused onboard analog output section.

3. And what about availability of the remote control for this dac.
Will you provide or offer one?

 

[Eldam]

Please Budha, if the sound can be heard where you live without air and forgett what I'm going to write as the wisdom is far from me : so please give me your pardon !

Answers

1: You will know all the détails as a single secondary or a double one is a detail !

2: two outputs : a single ended or a symetric :connect where you need it !

3: TOTALDAC prices

3 bis : TOTALDAC prices

Sorry can't resist ! Hey maybe you will never receive the dac as this is a Madof plan at this price... but in case of doubt: go to answer number 3 for a better serviceplease

 

[TNT]

Eldam, you don't make any sense. Google translate?

//

 

[Eldam]

Yes , sorry, here the translation : TOTALDAC prices

don't have link to translate in USD !

 

[TNT]

I was think more your language...

//

 

[Eldam]

Ah, you mean it doesnt make any sense....

I'm sure you understood ! : numbers are universal....

I was thinking on a more accurate translation without injures :

Please Budha, if you can hear me where you live in a place whitout air to transport the vibration of sound about what I'm talking about : "I beg your pardon for what I'm going to say as the wisdom seems to be a very long way totry to keep the silence" :

Answers :

1) pfffff, it's very not important, the price of a traffo with one secondary or two is the same but if I would have the answer I give you with pleasure but have not !
2) oups, I have myself asked something very similar, it seems you can hack easily the SE output as there is a connector. No traces to cut !
3) you are very focused on the delay of Soekris's answering but he "offers" to us a good development at a friendly price and if you want have a better service pay the price in a commercial company like Total DAC

Better translation

 

[TNT]

"without injures" :-D

...

Jepp, its ok


//

 

[Superdad]

There is no difference in monotonicity between sign magnitude and other notation. Using 0.01 ladder resistors you can reach 13 bit precision (2^13 = 8192 - 1/8192 * 100 = 0,012%).
There is no "window" that change math analysis due to sign magnitude. When the 14th bit and upper bits switch, the precision decreases.
Using dual ladder you are doubling the error, that's the reason BB used a single ladder in its PCM1704.

Andrea, can you explain a bit more to me? I am not very technically math oriented.
Perhaps I should not have referred to the monotonicity, but still I believe there is significant benefit to sign-magnitude style in that the MSBs are not constantly switching. Otherwise 0.01% resistors would be nowhere near good enough for the first 4-5 MSBs.

And to quote what Soren said: "The DAC will not be be monotonic down to more than maybe 14 bits, but thanks to the sign magnitude principle those 14 bits will still be there at the -60 db level, and it will be level linear down to the last bit, way below noise…."

But what I don't understand about what you said Andrea is about use of a "dual ladder" on this board versus how it is done in PCM1704. On the photo of Soren's prototype board I see 4 ladders, but these cover stereo 2 channels and each pair is for +/-, sign-magnitude style. Is this not what you see? What do you see that makes you refer to it as a dual-ladder design?

Thanks in advance.
AJC

 

[soekris]

Soekris could you please address the questions I recently asked .

1. In case I am not going to use DAC’s internal analog output stage , is it possible to power the rest of DAC’s circuit with just one single secondary winding of the transformer, or I still need two secondarys.

2. There was unanswered question[post #50] about how to disconnect the unused onboard analog output section.

3. And what about availability of the remote control for this dac.
Will you provide or offer one?

First, to increase the change of answers significant, turn html off in you postings....

1) The power supply have been dimensioned for a trafo with two secondary windings, but if you reduce loading by removing the output buffers it will probably work just fine with one winding.... You still need + and - supply, but one windings can provide that.

2) R-2R output before buffers is available on a connector. If you believe that the buffer somehow affect that signal, you can a) cut traces and remove power to buffers, or b) remove buffer opamps completely. I will not do custom versions (unless you order enough...), but once in full production I'll be happy to provide detailed howto's.

3) I currently have no plans to provide remote control, need to limit manpower needed, already doing too much.... But you can control the DAC over a std serial port, so maybe remote control and a display is a little diy project....

 

[soekris]

Please Budha, if the sound can be heard where you live without air and forgett what I'm going to write as the wisdom is far from me : so please give me your pardon !

3 bis : TOTALDAC prices

Sorry can't resist ! Hey maybe you will never receive the dac as this is a Madof plan at this price... but in case of doubt: go to answer number 3 for a better serviceplease

I'm sorry for the low prices. Maybe I should start using buzz words like "Platinum", "Signature", "Femto Clock", "Carefully tested", "Select component", "Exclusive Technologies".... And then add $100 per word used