• These commercial threads are for private transactions. diyAudio.com provides these forums for the convenience of our members, but makes no warranty nor assumes any responsibility. We do not vet any members, use of this facility is at your own risk. Customers can post any issues in those threads as long as it is done in a civil manner. All diyAudio rules about conduct apply and will be enforced.

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

While we continue to wait for the manual, can anyone offer an opinion as to how to best control the volume level using a pot for 4 boards in an active 4 way system? Would it work to use a one pot and wire it to each of the 4 boards, or to use a 4 deck pot and wire an output to each board? I am thinking to use a motorized alps pot to have remote control capability and they come in various configurations....Any insights appreciated...
 
Member
Joined 2003
Paid Member
Yes, i totally agree.. they give low distortion, but always found op-amp sound boring. A well designed discreete amplifier stage sounds alot more alive, and shouldn't give much more distortion either.

RollE2K (or anyone else) have any good alternative stages?

All I have for stages following a voltage output DAC are either OpAmp based (AD815, OPA1632/LME49724, LME49710/LME49990), no-gain buffers, or tube.

All of the interesting alternatives are I/V stages with very low input impendences.

I'll get to tube eventually on this, but for now I want to run on the +-12v-15v rails I have available plus I need gain.

What else should I look at?

Greg in Mississippi
 
This statement is...

There's nothing superior about balanced. In fact in most cases it will be inferior as the signal will be passed through some opamps to generate the balanced out. With this DAC the unbuffered single ended output is the purest as long as you feed it to a reasonably high Z load.

Somewhat misleading and too general in nature. Balanced circuits, when properly implemented, will outperform single ended ones.
Soren's DAC is not by nature of its design a balanced circuit, and I agree that adding a "balanced" output with an IC OPA could affect the sound for the worse: but some buffer circuit after the ladder is going to be necessary in most systems to get appropriate gain and current drive.

It is not true that: "in most cases it will be inferior as the signal will be passed through some opamps to generate the balanced out.", Especially when considering DACs. Most DAC chips already have balanced outputs, and the balanced signal does not need to be "generated" by anything, suggesting that balanced output from a DAC generally requires the balanced signal to be made by an IC OPA circuit is incorrect.

Best performance of Soren's DAC will likely be realized by a fully balanced approach, using two DAC boards, and taking the outputs from the ladders and buffering via a fully balanced discrete circuit.
 
RollE2K (or anyone else) have any good alternative stages?

All I have for stages following a voltage output DAC are either OpAmp based (AD815, OPA1632/LME49724, LME49710/LME49990), no-gain buffers, or tube.

All of the interesting alternatives are I/V stages with very low input impendences.

I'll get to tube eventually on this, but for now I want to run on the +-12v-15v rails I have available plus I need gain.

What else should I look at?

Greg in Mississippi

I'll be running se direct to lampizator 6111 board, just waiting for a few parts from mouser. In the meantime the direct se output to alps 10K is sounding great. Dynamics, freq response, volume, soundstage, etc. all present. What tube will you be using?
 
Member
Joined 2003
Paid Member
I'll be running se direct to lampizator 6111 board, just waiting for a few parts from mouser. In the meantime the direct se output to alps 10K is sounding great. Dynamics, freq response, volume, soundstage, etc. all present. What tube will you be using?

Not decided yet. I have a good stock of 6DJ8/6922/6N1P variants and will probably start my re-tube journey there. I've also had a recommendation of a 6N5 or 6N6-based Aikido stage.

And my first implementation of the board will go in a chassis where I already have the needed supplies including a +-12v-15v supply for a SS output stage, so that's why a SS stage will come first... and the base output level is likely not high enough, but I will try it. If it does turn out to be sufficient, a simple FET or monolithic buffer would work well.

Greg in Mississippi
 
If I understand the TP system, I think the Chronos board is designed to work with their cape and provide the clocking. But then, what (if any) synch happens with the DAM is not clear to me. I must admit to not understanding much of the basics involved in this--my DIY projects (at least in digital audio) have been the result of others' knowledge and sharing.

Alan

Alan, in this context the TP Chronos etc. is irrelevant; the BBB/Botic combo has no dependency on TP hardware. I already use a BBB/Botic with an Acko SO3 very successfully. The SO3 provides isolation and high quality clocking of the I2S data and in theory the Soekris DAC should be able to do the same.

In simple terms...

The starting point is to be aware that the BBB only has a single clock at 24MHz, which is fine for 48KHz family data rates but it means 44KHz data is resampled to 48KHz - not desirable. To avoid the resampling you disable the BBB clock and feed it the correct masterclock rate from an external source, enabled by the Botic software.

To facilitate the correct external clock rate the BBB/Botic sets a frequency select flag, which tells the masterclock source which rate to transmit. The BBB/Botic then uses the correct masterclock rate to 'play' the incoming data correctly (i.e. it sends the correct I2S out via its header pins, through the isolators, gets reclocked in the FIFO, etc.).

The Soekris DAC has a frequency select input and a masterclock output so the theory seems to say it will work but the devil could be in the detail and the Soekris uses a programmable clock rather than discrete fixed frequency clocks; it is that which needs to be tested.

Ray
 
RollE2K (or anyone else) have any good alternative stages?

All I have for stages following a voltage output DAC are either OpAmp based (AD815, OPA1632/LME49724, LME49710/LME49990), no-gain buffers, or tube.

All of the interesting alternatives are I/V stages with very low input impendences.

I'll get to tube eventually on this, but for now I want to run on the +-12v-15v rails I have available plus I need gain.

What else should I look at?

Greg in Mississippi

I'm not actally a designer in anyway, myself and usually like simple discrete circuits and there are plenty good ones around, then from those i often try to experiment to get what i want. The thing is that it all depends on what you want to get out of the stage. Are you thinking about running this straight into power amplifier after the amplifier stage you are planning on using? And would you want it to be balanced output, or SE? - what loads will you be driving with the output? it all matters to what output stage that will be optimal.

Myself i'm going to build a 6N6P Parafeed LTP with either plate chokes OR gyrators, haven't decided yet, the output transformer i have decided to be LL1689AM, and then balanced output to my headphones or line out.
 
Somewhat misleading and too general in nature. Balanced circuits, when properly implemented, will outperform single ended ones.
Soren's DAC is not by nature of its design a balanced circuit, and I agree that adding a "balanced" output with an IC OPA could affect the sound for the worse: but some buffer circuit after the ladder is going to be necessary in most systems to get appropriate gain and current drive.

It is not true that: "in most cases it will be inferior as the signal will be passed through some opamps to generate the balanced out.", Especially when considering DACs. Most DAC chips already have balanced outputs, and the balanced signal does not need to be "generated" by anything, suggesting that balanced output from a DAC generally requires the balanced signal to be made by an IC OPA circuit is incorrect.

Best performance of Soren's DAC will likely be realized by a fully balanced approach, using two DAC boards, and taking the outputs from the ladders and buffering via a fully balanced discrete circuit.
Agreed if balanced is inherently generated by the design. Then a transformer conversion to single ended works well.

In general though, balanced is not inherently superior to single ended. Often it complicates unnecessarily, and mostly is just a result of adherence to marketing dogma that has been foisted on the audio market by those with a vested interest in selling their products.
 
Disabled Account
Joined 2005
Haha, that one is even worse :rolleyes: :D


What does he call a null exactly?
I was under the impression that the attenuation needed to be high enough at fs/2, but that something like -40dB would be enough, or even less (-20dB is already quite low), as it was only there to mask an already subtle artifact.
There is no evidence that this ringing will occur exactly at fs/2 either, depending on the designs of the filters used in the production process.
So in the end all that remains is the idea of attenuating the signal before any other LP filter that might have occurred during production, and do it with shallow (less total ringing) and minimum-phase (only post ringing) filters.

In the 2003 AES conference paper Craven says:

It may be considered prudent to make the apodising filter have a response that is essentially zero by 90% of Nyquist, so that uncertainty about the transition- band response of other filters in the chain does not affect the final response.

I've seen comments in other places that Meridian implement a null at fs/2 for 44.1kHz. Craven was consulting for Meridian at the time he wrote the apodization papers.

Unfortunately papers Craven published only discusses sampling rates greater than 48kHz. It seems virtually impossible to have "essentially zero" response at 90% of nyquist (fs/2) at 44.1kHz sampling rate as this would put the null point at 19.845kHz and would result in very audible attenuation of HF.
 
Last edited:
Disabled Account
Joined 2005
Hi Soekris,
1. You have already mentioned “then to +-4V by precision low noise medium current opamps”; “-4V reference is sent though an inverter with 0.01% resistors generating the +4 reference”.
What the low noise medium current opamps are you used?
2. Why you used opamps instead a simple resistive divider, for example?

Søren can probably answer this better but I think the op-amps are used to buffer the output of the reference. Walt Jung used this technique in an article on ultra low noise voltage references (and possibly in his super-regs). Using an op-amp buffer lowers the output impedance, and can potential lower the noise output of the reference to that of the op-amp with additional filtering. It's a method that is often used in precision ADC's that require a highly accurate, low noise voltage reference.
 
Last edited:
Søren can probably answer this better but I think the op-amps are used to buffer the output of the reference. Walt Jung used this technique in an article on ultra low noise voltage references (and possibly in his super-regs). Using an op-amp buffer lowers the output impedance, and can potential lower the noise output of the reference to that of the op-amp with additional filtering. It's a method that is often used in precision ADC's that require a highly accurate, low noise voltage reference.
Yep, prevents loading of the reference... Isolates it from whatever it's feeding.
 
Some questions :

1. Is A+/- on J2 raw filtered (by 820uF caps) DC after the bridge rectifier?
2. Does A+/- run the output OPAMPS? This drawing seems to imply it does : https://hifiduino.files.wordpress.com/2015/01/r2routput.png

3. I know it is not recommended (and I hate to ask this) but if the SMD 8L05/9L05 regulators are removed from the PCB can +/-5V be injected on J2?

I think a + and - are the supply for all regs except the OPAMPS for the 4v
these OPAMPS are powered by the 8L05 9L05 regs
so ill remove these and take in power either at the 100uf caps or on the 8L05 pads
j2 might not be the best place to take in external power

just to add I expect most of the Sq to come from better filters

there's also plenty of risk
so need to be prepared to buy new boards if a mistake is made
 
Last edited:
I expected a diagram of the DAC to better understand the operation but Soekris is very busy and has been explaining the DAC in numerous interventions. I have made a selection of this to understand. I made the diagram that I could understand about the initial design. The current limits without changing architecture are:

Up to 1016 tabs at 44.1K/48K input sample rate.
Up to 508 tabs max at 88.2K/96K input sample rate.
Up to 252 tabs max at 176.4K/192K input sample rate.
124 tabs at 352.8K/384K input sample rate, but normally bypassed.
FIR2 is operating at 2.822M/3.072M and can have up to 120 tabs.

To better understand the level of DAC DAM1021 I have to meet other products on the market that have been mentioned in this thread..

MSB has one pass filter at 16x (705.6 Khz.) with 3200 taps. Or 32x (1.4112 Mhz.) With 6000 taps. The filter have noise shaping. The accumulators have 80-bit. Not round the result before adding next sample, intermediate data is stored at 80 bit resolution. Finaly convert to 24 or 26 bits including ultrasonic dither and noise shaping to have in sonic frecuences more than 24 o 26 bit resolution.

Ayre has one pass filter at 26x (1,1466 Mhz.) with 64-bit accumulators.

Dam 1021 have 2 or 3 cascade filters 2 FIR and 1 IIR. (64x 2,822400 Mhz.) with 67 bit MAC accumulator. Unlike the rest of DAC allows to manufacture their own filters, perform low pass, bandpass, highpass for speakers, Room Correction. Very interesting for diy builders and experiments in the forum community.
 

Attachments

  • filtro.JPG
    filtro.JPG
    106 KB · Views: 957