Transcendent Sound Forum
Latest Topics
 
 
 


Reply
  Author   Comment   Page 2 of 3      Prev   1   2   3   Next
Kelvin Tyler
Reply with quote  #16 
W. Nothing like a bit of experimentation.Your options are a bit limited if you have a one tube preamp.
I have just revisited Norman Crowhurst's excellent monograph 'Understanding Hi-Fi Circuits', (1947!!). He has a section on feedback type volume controls in chapter 9. Most of his examples are for two stage preamps, but he does give one example of a single tube implementation. This simply consists of inserting a variable resistor between the lower end of the cathode biasing components and ground, with the grid resistor returned to the junction. This obviously gives increasing degeneration as the resistor value is increased from zero. He does mention that this should reduce distortion at low volume settings. Very easy to try, apart from disturbing your 'star' grounding!
Hope you find a result that pleases you. Keep us posted.
Kelvin
Kelvin Tyler
Reply with quote  #17 
Correction! Book date should be 1957 not 1947!
Kelvin
Kelvin Tyler
Reply with quote  #18 
W. On reflection, I feel Crowhurst's scheme mentioned above would not be suitable for your preamp as the mu of the 26 tube is only ~8 and the gain is dominated by the high choke impedance. The required degenerative resistor would have to be so high it would drop far too much voltage.
Apologies for my multiple posts!
Kelvin
Wolfgang
Reply with quote  #19 

No need to apologize, Kelvin, your input is highly appreciated!
Here is what I did so far. I started with using a 10k resistors for Rg, a 100K TKD stereo pot for Rf and 0.1uF Mundorf caps for the output coupling cap. Higher values are not practical as I would lose too much gain. With these I lost about  60%. 

I wired one channel for variable NFB as in figure (a) of your post #14, the other channel had the pot wired as voltage divider.

I didn't expect this much difference between these two ways of volume control (see pics). The NFB keeps the square wave in perfect shape, the typical pot rounds the upper edges.


Then I used a 50k Alps pot for Rf with 1k for Rg and got of course almost all gain back. That’s a very good compromise and I couldn’t find any problems with the NFB . Perfect square wave, sound is a tiny little bit more controlled compared to no grid stopper/no NFB like in the original version. No need for a resistor in series with the pot so that Rf never can get 0 as the high volume will keep us from going there.


If I ever buy one of these DSD/PCM DAC without digital volume control I will use the NFB volume control as there is obviously no difference in sound, transients, “live character” between digital volume control and volume control with NFB. I cannot think of anything better right now. Unfortunately I didn't have a TVC handy and  from my vague memory of how it sounds I cannot say anything about it in this context.The limiting influence of a traditionally wired pot, however, is audible.
The only thing left is to calculate Zin and Zout with these values for Rf and Rg and the final test in my main system. I did all tests with headphones and lower grade equipment.

Attached Images
Click image for larger version - Name: square 1khz.jpg, Views: 6, Size: 68.85 KB  Click image for larger version - Name: square 5kHz.jpg, Views: 6, Size: 67.61 KB 

Kelvin Tyler
Reply with quote  #20 
Good stuff,W. Looks like you have achieved your goal. Hope all works out with your main system.
Kelvin
Wolfgang
Reply with quote  #21 
After listening for some hours with the new NFB feature, experimenting with different combinations etc. I came to the conclusion that it is more or less unpredictable how minor changes in electronic circuits can affect the sound. I like the results very much but had expected something different.

If I compare the practical outcome of the NFB volume control to what I have tried before, like changing NFB in OTL amps, different volume controls, experimenting with digital filters, I would summarize it like this:
Volume control with a regular pot clearly softens the sound and slows down transients compared to the NFB version but the audible difference is less dramatic as the 2 pics in post # 19 might suggest (1kHz, 5kHz square wave) although it's clearly audible.
The NFB in the RCA 26line stage can be used as volume control but -as expected- it sounds “tighter” than a regular pot because of the NFB. This can be experienced as positive at lower volume levels as it sounds extremely “clean” and details are more audible compared to a regular volume control at these levels. At higher volume levels the sound gets more “live” again. There is no abrupt change or coloration in this process. It’s very natural and homogenous.
This version of NFB creates a wonderful separation between instruments and a wider and deeper virtual sound stage, and it has the most incredible crisp trebles (but not hard or unnatural sounding compared to the rest of the frequency spectrum). This kind of differentiation and amount of details- but always kept 100% stable in a homogenous frame - I have only heard before from top high-end preamps. It reminds me very much of what different output filters can do in a R2R DAC but without changing certain elements at the same time, like moving some parts more to the forefront or to the side etc. With NFB in this line stage all these elements don’t change no matter what style of music or at what volume level, while different digital filters can have completely different results with different musical genres: one can sound absolutely stunning with Jazz but really bad with classical music and vice versa.
I prefer the combination of digital volume control with 70% NFB volume control and I favor it before any digital filter.
Just in case: the effect of NFB in the RCA26 line stage on the sound, virtual sound stage has nothing in common with changing the overall NFB in OTL amps. First of all I wouldn’t recommend it as it would affect so many unforeseeable factors (most important impedance), or at least only at a very minor degree. In the case of the 300B OTL a little less NFB (2k2 to 5k6) shifts the sound a little to “darker” but doesn’t change anything else.
Wolfgang
Reply with quote  #22 
Sorry, what I meant is 30% NFB.I wired the pot so that it work like a regular volume control: the higher the volume (clockwise), the less NFB.
Ray P
Reply with quote  #23 
Returning to the original theme, another R2R DAC new to the market;

http://www.audio-gd.com/R2R/R2R7/R2R7EN.htm

Ray
Wolfgang
Reply with quote  #24 
Here is the old question: How is it possible to get 24bit resolution with resistors with only 0.01% or even more tolerance as used in the Soekris DAC?
Denafrip claims to use 0.005% resistors which don’t exist according to Audio-GD.
Audio-GD uses a different way altogether (correcting the ladder) in order to resolve this problem but all 3 DACs claim a S/N ratio of -120dB which would be an indicator for true 24 bit resolution.
So, what’s the true resolution of these DACs? 18bit, 20bit, 24bit? Theoretically at least the Audio-GD DAC could be a true 24bit DAC.

But that’s just theory. How they sound is what matters and so far the Soekris DAC is still the best sounding DAC I ever had in my system at home. Considering the price, warranty, and the really nice output stage the Audio-GD would be my first choice at this point.
Gary Kemp
Reply with quote  #25 
Slightly o/t: How much of a difference (roghly! but choose your units!) is there between my CD/SACD players (Marantz, can't remember the model but it's about £450, and a Sony SCD-ES777), and these hi-end DACs? 
Wolfgang
Reply with quote  #26 
Gary,
I still use my old Marantz SA-11 S2 CD/SACD player as music source for testing equipment. So I am pretty familiar with its sound. It also has 3 different filters and a noise shaper but they are rather subtle like most filters of the HQ player.

https://www.stereophile.com/hirezplayers/marantz_sa-11s2_reference_sacdcd_player/index.html

My scratch build Soekris R2R DAC (and one of its neutral sounding original filters) with Wave I/O USB interface, JPlay 6 player, SOtM USB card, optimized Win 8 for JPlay, RCA 26 line stage, 300B OTLs reproduces the instruments, placement relative to each other, placement in space where the recording took place, transients, harmonics etc very realistically . Everything seems in the right position, absolute and relative to each other. It feels “real” like “live”.
If I play the same song with the Marantz (with preamp volume control, GG or other pre) it sounds very good but the reality, the magic to be “there”, has somehow disappeared. It sounds more artificial..like good HiFi. This magical thing doesn’t happen anymore that makes me listen with full attention for hrs. The space in the music, the vibration of each note as single note , the context from which all the single elements draw their meaning is almost gone. The music plays on but it’s more difficult to connect to its context, its “soul”, or however one wants to describe this important element that makes music the enjoyable thing it can be.
If I use one of the less neutral sounding filters (can be downloaded), tweaked for certain elements (minimal phase, less ringing, roll-off points at different frequencies, more weight to the time domain, more weight to filter accuracy etc) I can get even more authentic sounding music reproduction but it’s limited to certain music styles. I really can reproduce the experience of sitting in a Jazz club for example or I should rather say: the experience comes to me automatically. I don't have to focus or concentrate or imagine something. Analog also gets very close in this respect, btw, but with a lot more efforts and fine tuning.

I hope this answers your question or points at least in the right direction.
Ray P
Reply with quote  #27 
Quote:
Originally Posted by Wolfgang
Here is the old question: How is it possible to get 24bit resolution with resistors with only 0.01% or even more tolerance as used in the Soekris DAC? Denafrip claims to use 0.005% resistors which don’t exist according to Audio-GD. Audio-GD uses a different way altogether (correcting the ladder)...


Both the Holo Spring and Audio-GD R2R-7 use ladder correction technology. I previously enquired of Denafrips whether they use something similar but have not had a response yet (also pointed out that the front panels of their Pontus and Venus DACs have a spelling mistake!).

The Audio-GD looks to be very well considered with attention to detail where it matters but, compared with the Holo and Denafrips, is a bit more clunky as it uses internal jumpers to change options instead of front panel buttons.

But, as you say, at the end of the day it is the sound that matters and there is the rub as getting to audition any of these DACs, let alone compare them, will be pretty difficult.

Ray
Gary Kemp
Reply with quote  #28 
Wolfgang, 
Thank you very much for this thoughtful reply. I listen mostly to vinyl, where I have a pretty good set-up, but have long wondered whether my digital set-up could be improved. It does sound pretty expensive! 

Quote:
Originally Posted by Wolfgang
Gary,
I still use my old Marantz SA-11 S2 CD/SACD player as music source for testing equipment. So I am pretty familiar with its sound. It also has 3 different filters and a noise shaper but they are rather subtle like most filters of the HQ player.

https://www.stereophile.com/hirezplayers/marantz_sa-11s2_reference_sacdcd_player/index.html

My scratch build Soekris R2R DAC (and one of its neutral sounding original filters) with Wave I/O USB interface, JPlay 6 player, SOtM USB card, optimized Win 8 for JPlay, RCA 26 line stage, 300B OTLs reproduces the instruments, placement relative to each other, placement in space where the recording took place, transients, harmonics etc very realistically . Everything seems in the right position, absolute and relative to each other. It feels “real” like “live”.
If I play the same song with the Marantz (with preamp volume control, GG or other pre) it sounds very good but the reality, the magic to be “there”, has somehow disappeared. It sounds more artificial..like good HiFi. This magical thing doesn’t happen anymore that makes me listen with full attention for hrs. The space in the music, the vibration of each note as single note , the context from which all the single elements draw their meaning is almost gone. The music plays on but it’s more difficult to connect to its context, its “soul”, or however one wants to describe this important element that makes music the enjoyable thing it can be.
If I use one of the less neutral sounding filters (can be downloaded), tweaked for certain elements (minimal phase, less ringing, roll-off points at different frequencies, more weight to the time domain, more weight to filter accuracy etc) I can get even more authentic sounding music reproduction but it’s limited to certain music styles. I really can reproduce the experience of sitting in a Jazz club for example or I should rather say: the experience comes to me automatically. I don't have to focus or concentrate or imagine something. Analog also gets very close in this respect, btw, but with a lot more efforts and fine tuning.

I hope this answers your question or points at least in the right direction.
Ray P
Reply with quote  #29 
Quote:
Originally Posted by Ray P
Both the Holo Spring and Audio-GD R2R-7 use ladder correction technology. I previously enquired of Denafrips whether they use something similar but have not had a response yet (also pointed out that the front panels of their Pontus and Venus DACs have a spelling mistake!).


The Denafrips DACs do not use any ladder correction techniques so I'm guessing are much closer to the Soekris R2R philosophy. I'm trying to find out more.

Ray
Kelvin Tyler
Reply with quote  #30 
I have always had a liking for the R-2R DAC topology. So elegant and so fast. Ray P makes a very good point when he mentions the difficulties involved in auditioning and comparing these latest devices. Not withstanding this caveat, Wolfgang, and no doubt Gary, will be amused to learn that I have ordered a Soekris 1541 to replace my trusty Arcam irDAC which has served me well in recent years. Now my hearing is not all that it was in my younger years, but this can hardly be a retrograde step, if not exactly inexpensive. I will keep you posted in due course.
Returning briefly to Wolfgang's 26 tube amp with NFB volume control. Where has the extra 6dB of open loop gain gone(mu~8)? Is the preamp driving a load of the order of Rp, the 26 tube plate impedance (~7.5K)?
In the anode-follower, the feedback resistor is a direct load on the output so the output impedance will vary as the gain control is altered. Taking a 50K linear pot with the wiper connected to the grid and set at about one third rotation, ie R2~ 2xR1, I calculate the gain to be about 1.3 and Rout to be about 1.2K. The input impedance would then be about R1, ~17K.
Previous Topic | Next Topic
Print
Reply

Quick Navigation: