Transcendent Sound Forum
Latest Topics
 
 
 


Reply
  Author   Comment   Page 1 of 6      1   2   3   4   Next   »
Wolfgang
Reply with quote  #1 
I am a big fan of R2R DACs as they don't have the "digital" footprint of Delta-Sigma.
Now they are available as stand-alone gear for those who don't want to build from scratch. I think the R2R work extremely well with TS amps. Different filters can help to fine tune any system even without DSP/EQ.
The only thing missing for complete happiness for owners of TS amps would be a TS line stage/buffer without volume control based on a single triodes, like the MP but without the input stage/volume control. And the PP amps could sound even better with some added even order harmonics from a pure triode stage. Who knows?
Here are some links with the details:
http://www.soekris.dk/dac1541.html
http://www.soekris.dk/dac1421.html
Ray P
Reply with quote  #2 
Wolfgang, I'm currently looking to make a purchase of an R2R DAC but not one of those you link to. Why? Because, as you know, I like DSD reproduction and the Soekris R2R converts DSD to PCM before passing it to the R2R DAC.

The DACs I'm looking at are different in that both manufacturers cater for both PCM and DSD by having dedicated R2R sections for each; they're essentially a PCM DAC and a DSD DAC in one box (they automagically detect the incoming data type). The DACs I'm looking at are;

Holo Audio have a DAC called 'Spring'. Here's a review;

https://www.computeraudiophile.com/ca/reviews/holo-audio-spring-level-3-kitsune-tuned-edition-r2r-dac-review/

Denafrips have three DACs called Pontus, Venus and Terminator. They're basically the same approach but with increasing levels of 'build quality' (i.e. better power supplies, clocking etc. Denafrips also do a budget R2R DAC, the Ares, though it doesn't have DSD capability. Here's a link to the web page of their international distributer;

https://www.vinshineaudio.com/products

I expect to purchase one of them within the next week or so

Ray
Wolfgang
Reply with quote  #3 
Ray, I really like the "Terminator"- specs and of course the Amanero USB interface. The Soekris DACs cannot compete in this respect. Don’t know how much it’s audible and what kind of USB interface the Soekris DAC uses. The I2S input option is also a wonderful feature which you don’t find often at this price.
I would be very interested to hear about your impressions with these DSD/PCM DACs!

I am still hesitant regarding DSD because of two things. DSD downloads are still expensive (look at HD tracks for example) compared to HiRes PCM and I don’t hear so much difference in SQ between different higher sample rates/resolution. But I am anyway limited if I use  the SHARC module for DSP (max 96/24). Maybe I should try one of these DSD/PCM DACs at some point as I hardly need DSP in my room. With different JPlay engines and DAC filters I can control the sound almost as effectively as with DSP but with more realistic sounding results. But JPlay alone is already very effective.

Sound studios are still struggling with the costs and efforts it would take to make the upgrade to DSD recording on the equipment side. The new approach from Meridian (MQA) can be easier implemented and will probably be more profitable in the long run (streaming etc). We will see. I simply have no idea which format -MQA or DSD - will have the bigger impact and will win this race. Just in case somebody is interested in more details about DSD/PCM - not the last word on this topic but interesting: http://www.mojo-audio.com/blog/dsd-vs-pcm-myth-vs-truth/

My biggest concern is that these DACs don't have any option for uploading different filters. Or did I miss this part? I think this is the most important feature of the Soekris DAC as it deals directly with the biggest "digital-sound" problem: the influence of the filter design on the analog sound that comes from the DAC and how it interacts with the rest of the system/room response (bloated bass, too much bass, harsh sound, digital sound etc). I think it’s more convenient if the user can chose from different filters than be stuck with the filter that sounds “best” to the designer ears in his test environment.

Interesting also that after about 2 years of listening with my R2R DAC I can say with confidence that new digital recordings (not so much old analog recordings on CD) sound still different but clearly better than even very good analog. “Different” simply because these are two different ways of reproducing the recorded signal which will always be audible as “different”, “better” because the virtual space, the definition, the dynamics are simply outstanding compared to even very good analog. At least in my opinion this article gives some possible explanations:
http://www.endino.com/archive/arch2.html
Ray P
Reply with quote  #4 

Wolfgang, I don’t generally buy DSD downloads but use HQPlayer to upsample FLAC files to high rate DSD. HQPlayer provides plenty of filter options for both PCM and DSD and you can apply room correction with it, notwithstanding its excellent sound quality as a player.

I’m away for the weekend but will be happy to answer any questions when I get home.

Ray

Wolfgang
Reply with quote  #5 
Ray, I know that you use HQ player/upsampling FLAC to DSD but I thought this was only in connection to your “no-DAC” DSD output device.
So I assumed that if you spend the extra money for a good R2R DSD DAC , sacrificing the option for uploading different filters, you must have native DSD in mind.
On second thought I cannot imagine how and why upsampled PCM to DSD would suffer much if it’s converted back to PCM (like in the Soekris DAC) as it’s not native DSD in the first place.
I downloaded the demo version of the HQ player in order to see how it performs in my system.
Ray P
Reply with quote  #6 
Quote:
Originally Posted by Wolfgang
I know that you use HQ player/upsampling FLAC to DSD but I thought this was only in connection to your “no-DAC” DSD output device. So I assumed that if you spend the extra money for a good R2R DSD DAC , sacrificing the option for uploading different filters, you must have native DSD in mind.


HQPlayer is my normal player software for all my playback, not just 'No DAC' and I generally use it to upsample to DSD even when playback is via my ES9018 DAC. The reason I started to explore DSD was because I preferred it (upsampled via HQP) on the ES9018 DAC to PCM. As you'll find out with your trial, HQP has filter options.

Quote:
Originally Posted by Wolfgang
On second thought I cannot imagine how and why upsampled PCM to DSD would suffer much if it’s converted back to PCM (like in the Soekris DAC) as it’s not native DSD in the first place.


It's not whether is suffers with being converted back or not that is the issue, it's simply that I prefer DSD playback, which the Soekris cannot do. With the Soekris it isn't worth upsampling to DSD in the first place. The Holo and Denafrip DACs can perform D to A conversion of DSD data as well as PCM (as mentioned before, they're essentially two DACs, on for DSD and one for PCM, in one box)


Quote:
Originally Posted by Wolfgang
I downloaded the demo version of the HQ player in order to see how it performs in my system.


To upsample to DSD (especially >DSD128) you'll need a powerful computer but if you have no means to playback DSD (as with the Soekris) there would be no point. You will be able to try out different PCM filters and oversampling options though. I have found the optimum configuration to be with an HQP NAA (Network Audio Adaptor: your main computer does the heavy lifting and streams the output to a lightweight silent computer that feeds the DAC) as the computer doing the processing, which can be noisy - audibly and electrically) can be remote from the audio system. Will be interested to see how you get on.

Ray
Achim
Reply with quote  #7 
Hi Wolfgang and Ray,

I have a Soekris DAM 1541 on order, delivery date was pushed back once already to late June 2017, let's see when I will have it.

I talked to Soeren at the Highend show in Munich - he showed me the testpoints on the board where on can tap the raw signal after the R-2R stage before the opamp output. I intend to take it from there to my Masterpiece or the 300B power amps directly.

The 1541 uses switchmode power supplies which I hope won't rain on the parade. Soeren was of some conviction that these are a better compromise than a linear PSU with its magnetic field in the chassis. I think it's a pity he did not go for a seperate PSU.

I also use HQplayer into my Oppo 105D. At first I used it to upsample to DSD - but have gone "back" to upsampling to 384 PCM. I'll see what I like more with the R-2R.

Cheers,
Achim


PetrL
Reply with quote  #8 
Technically, why they don't use through hole resistors? It seems perfectly feasible considering the DAC size, it would fit into standard width chassis (who of the people shelling out more than 1000 Eu for a dac needs that small chassis anyway?) They may have higher inductance, but this can be decreased by paralleling them in critical places. The chip resistors are not of the best quality and stability.
Wolfgang
Reply with quote  #9 
Ray, I don't think I will do the 2 computer version but I want to get to know HQ player (finally!) and upsampled DSD playback. I hope of course there will be interesting native DSD downloads over time with music I like. I build all my audio computers from scratch and can make them running silently enough so that I cannot hear them anymore from my listening position. In the worst case I just go for a fanless CPU version and SSD for processing the data.

Achim, interesting development. Also interesting that you went back to PCM (compared to Ray). There would be no problem to get rid of the SMPS and use linear psu. You wouldn't have to solder SMD parts which is a nightmare. I did this already 2 times because I got one of the first board versions that needed some upgrading (as per Soekris' instructions). The DAC (direct out!) works absolutely fine with direct connection to the OTls but it sounds almost too clean and clinical. I recommend a pure triode line stage which can drive the OTLs to unbelievable SQ levels. This cannot be achieved with regular preamps! I don't know anymore how it came up but Ray sent me one day a link to a RCA26 linestage build. I did my own and little different version based on my previous 2A3/300B buffers/experiences but the RCA26 beats them in combination with the Soekris DAC and OTLs.This is the closest to live music I have ever come in home reproduction.

Petr, you wouldn't be able to lasertrim regular resistors and you need at least 0.005 for these DACs.And when it comes to digital, things change quite dramatically regarding circuit design, ground plane, several layer PCBs etc if you want a really top high end product. Not a good DIY project like analog audio. Wrong capacities and inductivities as a result of a bad design, too much space inbetween parts etc would ruin the whole thing.SMD is perfect in this respect.
Ray P
Reply with quote  #10 
Quote:
Originally Posted by Wolfgang
Ray, I don't think I will do the 2 computer version but I want to get to know HQ player (finally!) and upsampled DSD playback. I hope of course there will be interesting native DSD downloads over time with music I like. I build all my audio computers from scratch and can make them running silently enough so that I cannot hear them anymore from my listening position. In the worst case I just go for a fanless CPU version and SSD for processing the data.


I generally build my own PCs too. My current audio 'heavy lift' PC that I use for upsampling to DSD has a Xeon processor (don't recall which one off hand) that benchmarks at around 9000. Upsampling to DSD256 is a load of about 40% and DSD512 around 80%. Obviously the Xeon, being a server processor, has a noisy fan, hence the NAA use to allow it to be remotely located.

Quote:
Originally Posted by Wolfgang
...interesting that you went back to PCM (compared to Ray).


One of the things a Holo or Denafrip will enable me to do is more easily compare PCM vs DSD.

Quote:
Originally Posted by Wolfgang
Petr, you wouldn't be able to lasertrim regular resistors and you need at least 0.005 for these DACs.And when it comes to digital, things change quite dramatically regarding circuit design, ground plane, several layer PCBs etc if you want a really top high end product. Not a good DIY project like analog audio. Wrong capacities and inductivities as a result of a bad design, too much space inbetween parts etc would ruin the whole thing.SMD is perfect in this respect.


There are some DIY DSD projects that use through-hole resistors in FIR filters with reported excellent results, for example;

http://www.xtremeplace.com/yabbse/index.php?topic=190645.0

One of the 'features' of the Holo Spring DAC is a compensation circuit that reduces the need for very tight resistor tolerances and compensates for drift.
Wolfgang
Reply with quote  #11 
I realized that the DSD capable DACs mentioned above don't have digital volume control.
I don't like analog volume control as it loses some subtle but important part of the signal in the case of pots /attenuators and doesn't sound as crisp as digital volume control. TVCs have different problems.

So I was thinking to control the output volume in a triode line stage with variable negative feedback. This idea is not new, there are even patents, but all with transistors or opamps.

Rather than sinking part of the signal into amp ground I could imagine that it would sound better if we reduce the signal level by adding some part of it in opposite phase reducing at the same time distortions at the output. Cannot be too bad as OTLs use it a lot.

Does anyone have some practical experience with this?
Kelvin Tyler
Reply with quote  #12 
Wolfgang. Many years ago I implemented an electronic 'volume control'  using a cathode follower with a variable  cathode resistor and a fixed bias. Can not remember the exact details, but I think it solved the problem I was having with high frequency loss in some setup I was working on.
Regards,
Kelvin
Wolfgang
Reply with quote  #13 
Kelvin,
so where would you start with a triode like this (see attachment)? The gain is about 12dB.
It’s not necessary to cover the whole volume range like with a regular pot as practical experience has shown that I need only about 10dB of volume control.

I would take the signal from the output and feed it back with variable Rf to the grid.

But I have some questions:
First, I don’t use a grid stopper (Rg) as it limits transients (practical evaluation) and adds resistor noise in this circuit . So I would only calculate with Rf . Ideally Rf shouldn’t be too high (resistor noise).

As Rf needs to be higher than Ra (1k2 choke) so that it doesn’t affect the anode load for the tube I guess the lowest possible resistance for one corner point (only theoretically) could be around 10k. I was also wondering if it makes any difference for calculating NFB that I use a choke instead of a simple resistor ?

The input impedance of the line stage should be able to handle at least 200-600ohm with 10dB NFB.

Is NFB volume control possible under these conditions: Rf not too high & no Rg, 10dB neg gain, choke loaded anode, input impedance not too low?

What would be the best way to approach this project? Simply calculate/define the corner points for 10dB NFB and then chose a pot/stepped attenuator that reaches up to the highest value of Rf for around 0dB while its lowest value would be a 10K (or higher) fixed resistor in series with the pot? The frequency dependent value for the output cap/Rf could be found later if it works at all.

I hope you can give some input.


Wolfgang

Attached Images
Click image for larger version - Name: triode.jpg, Views: 42, Size: 245.57 KB 

Kelvin Tyler
Reply with quote  #14 
Wolfgang. You have had me blowing the dust off some 60 year old notes![smile]
Thinking about adding Miller type feedback to your choke loaded 26 preamp, I guess what you would end up with would be an anode (plate)-follower type arrangement. I have sketched out the usual set-up in the attached fig(a). I note you have only 12dB open-loop gain and, assuming the plate choke is of the order of 100-200H. the AC load impedance will be quite high except at very low frequencies. Both these considerations may be problematical. In the conventional  set-up in fig(a) the gain could certainly be varied by adjusting the feedback resistor R2. If R1 was taken as 47K say, then R2 might go from 47K ~470K perhaps. Given your dislike of series resistors in the grid circuit you may not fancy this scheme very much. Personally, I would not be too worried by resistor noise at these signal levels. The audio bandwidth is not that wide really and we are not talking phono preamps here.

The cathode follower idea shown in fig(b) certainly works with modest signal levels. I see from my old notes that I stole the idea from Langford-Smith's 'Radio Designer's Handbook', Fourth Edition, 1953, p797. As stated there, way back I used a low gm sharp cut-off pentode (6SJ7) so the switched resistors did not need to be ridiculously low values, and adjusting the screen voltage gave me further control. I was not using the circuit in the context of audio preamps at that time.
Kelvin

Attached Images
Click image for larger version - Name: followers.jpg, Views: 39, Size: 213.01 KB 

Wolfgang
Reply with quote  #15 
Thank you, Kelvin.
I will try version (a)as only the signal from the anode has the extreme transients that I like so much and gives me this "live music" feeling. The signal from the cathode sounds less dramatic and more laid back. And with only one tube I need all the gain I can get (anode).
The choke is 200H,btw. The impedance of the choke at 1kHz is about 1.26Mohm. After a rough calculation it all looks ok except that the fixed output cap which is part of the NFB will give of course different roll-off frequencies at different values of Rf. So I chose 10Hz for Rf max and on the opposite side I will end up at 0.1Hz. I hope this will not give me problems. Will see.

My main goal is to find out how this kind of volume control sounds in comparison to conventional style and if it works at all. On the other hand I have found out that the feedback resistor can have an extreme influence on the sound of an amp and in this case this kind of volume control could turn out to be something interesting.
Previous Topic | Next Topic
Print
Reply

Quick Navigation:

Easily create a Forum Website with Website Toolbox.