CD6000OSE - tube output stage - help appreciated :-) - diyAudio
Go Back   Home > Forums > Source & Line > Digital Source

Digital Source Digital Players and Recorders: CD , SACD , Tape, Memory Card, etc.

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 30th September 2009, 04:59 PM   #1
diyAudio Member
 
Join Date: Oct 2007
Location: Cologne
Default CD6000OSE - tube output stage - help appreciated :-)

Dear community

I'm up to modify my CD6000OSE. I already changed most Electrolytics, some rectifiers and added some damping to the chassis.
All this made it a better player, but it is still far away from what I'd like to get - and I think that would be possible to get out of the basic concept.

I read a lot of threads here about modifying this player. Unfortunately I did not find a thread about a tubed outputstage. But this is exactly what I'd like to implement.

I know that there was a player sold under the label HEART. They added a transformer (??? for filtering AND converting balanced to SE???) and two 6922 tubes.
This player was reviewed very well. I don't know the circuit they used and I don't know how they implemented their circuit. What I found was a link to "Thermionic Valvestages" by Thorsten Loesch. He describes different DAC principles and how to implement which circuit. Well, thats it, I thought.

My main issue now is, that my player - unlike the easier tunable CD63/67 - has two DACs(SM5872BS) instead of one. Each of them is fed by the same integrated circuit, that modifies the I2S signal for each DAC (L/R flag and DATA content). Each DAC has four analogue outputs - normally: L+L-R+R- (two balanced channels) - here in this case one DAC has got 1L+, 2L+, 1L-, 2L-, the other has 1R+, 2R+, 1R-, 2R-. Now Marantz uses two differential buffers to get again two single balaced signals (L+L-R+R-). One more buffer (HDAM) module per side convert this signal into SE signals. So far so complicated. I hope this is understandable and right. Here the sheme:

So in my eyes I have a few options:

++ use just one of the positive outputs, i.e. 1L+ and 1R+ right off the DACs and add Thorstens tube stage.
++ use one balanced pair of each dac, add a transformer to make it SE and a tube stage
++ use the first part of the players HDAMs to summarize all the doubled channels and add a tube stage
++ build four differential tube stages followed by 2 balanced2 SE tube stage (6 double triodes)

++ deeply modify this player, deactivate one DAC, deactivate the separating IC (TC160G11AU) and feed the other DAC with an external clock and the I2S stream right out of the TDA 1302T (drive logic). As an analogue stage I'd use a tube like Thorstens one.

My questions now are:

?? Is my analysis of the circuit right?
?? Which of the options might be the best?
?? Which would be the best simple one?

Thanks for reading. Please post just positive answers! I'm not skilled, I just read a lot and have a huge amount of dangerous quarter knowledge. I might be wrong and hope to get some help here.

Cheers Ernst
  Reply With Quote
Old 11th October 2009, 10:48 AM   #2
diyAudio Member
 
Join Date: Oct 2007
Location: Cologne
Hmm, no reply so far... Anything wrong with my post? Questions too specific?
To be honest, I would be happy about any reply regarding tube outputs for my player. Maybe somebody knows a thread in another forum?

Thx
Ernst
  Reply With Quote
Old 11th October 2009, 11:16 AM   #3
diyAudio Member
 
Miniwatt's Avatar
 
Join Date: Nov 2004
Location: Den Haag
Quote:
Originally Posted by ernesternest View Post
Dear community
I know that there was a player sold under the label HEART. They added a transformer (??? for filtering AND converting balanced to SE???) and two 6922 tubes.
This player was reviewed very well. I don't know the circuit they used and I don't know how they implemented their circuit. What I found was a link to "Thermionic Valvestages" by Thorsten Loesch. He describes different DAC principles and how to implement which circuit. Well, thats it, I thought.
Afaik the way Heart did it was just add a tube stage after the existing output stage. Not impressive imho.
The extra tranny was a power xformer for the tube stage.

Quote:
My main issue now is, that my player - unlike the easier tunable CD63/67 - has two DACs(SM5872BS) instead of one. Each of them is fed by the same integrated circuit, that modifies the I2S signal for each DAC (L/R flag and DATA content). Each DAC has four analogue outputs - normally: L+L-R+R- (two balanced channels) - here in this case one DAC has got 1L+, 2L+, 1L-, 2L-, the other has 1R+, 2R+, 1R-, 2R-. Now Marantz uses two differential buffers to get again two single balaced signals (L+L-R+R-). One more buffer (HDAM) module per side convert this signal into SE signals. So far so complicated. I hope this is understandable and right. Here the sheme:

So in my eyes I have a few options:

++ use just one of the positive outputs, i.e. 1L+ and 1R+ right off the DACs and add Thorstens tube stage.
++ use one balanced pair of each dac, add a transformer to make it SE and a tube stage
++ use the first part of the players HDAMs to summarize all the doubled channels and add a tube stage
++ build four differential tube stages followed by 2 balanced2 SE tube stage (6 double triodes)

++ deeply modify this player, deactivate one DAC, deactivate the separating IC (TC160G11AU) and feed the other DAC with an external clock and the I2S stream right out of the TDA 1302T (drive logic). As an analogue stage I'd use a tube like Thorstens one.

My questions now are:

?? Is my analysis of the circuit right?
?? Which of the options might be the best?
?? Which would be the best simple one?

Thanks for reading. Please post just positive answers! I'm not skilled, I just read a lot and have a huge amount of dangerous quarter knowledge. I might be wrong and hope to get some help here.

Cheers Ernst
I'm curious about this also, I've been looking into adding a tube output to my Sony CDP which has dual balanced dacs as well.
I already designed a SRPP for a Philips TDA1541 with stunning results, I want more
  Reply With Quote
Old 11th October 2009, 12:40 PM   #4
diyAudio Member
 
Join Date: Oct 2007
Location: Cologne
Quote:
Originally Posted by Miniwatt View Post
Afaik the way Heart did it was just add a tube stage after the existing output stage. Not impressive imho.
The extra tranny was a power xformer for the tube stage.



I'm curious about this also, I've been looking into adding a tube output to my Sony CDP which has dual balanced dacs as well.
I already designed a SRPP for a Philips TDA1541 with stunning results, I want more
Hello Miniwatt,

I found this on: http://www.marantzphilips.nl/review_...cd6000_ose_le/

They explain:
Upgrades and Changes by Heart
Rik Stoet's Heart factory performs numerous modifications to the Marantz design. Most importantly in my opinion is that Heart disconnects the entire Marantz designed analog section (although it remains in the chassis) and substitutes a circuit using a pair of dual triode 6922 vacuum tubes. The pair is energized by two transformers, one powers the heaters, the other the anodes. The unit ends up with a total of 5 transformers, three for the power supply and 2 for the audio circuits. The printed circuit board is modified and relocated to reduce interference from the power supply.
The headphone jack has been disconnected to reduce the load on the audio section and increase performance. Additional Philips capacitors are installed in the high pass section. The transformer for 110-120 volt markets has been specially designed by Heart for use in those markets. Finally, additional damping to the chassis and transport has been added over and above that already installed by Marantz. Cosmetically, the only difference between the Marantz and the Heart unit is the addition of a small tasteful "Tubed by Heart" logo below the CD drawer.


Also I found this pic from the Heart CD6000:
There is no Opamp or HDAM visible. They route the signal out of the PCB far before the following opamp had been. And the added a clock.

I'd really like to know what they did precisely.

Yes, you probably could do the same or something similar, if we found out.
By the way, which 1541 CDP have you tubed?

Cheers Ernst
Attached Images
File Type: jpg cd6000tube.jpg (81.0 KB, 539 views)
  Reply With Quote
Old 11th October 2009, 10:39 PM   #5
diyAudio Member
 
Join Date: Oct 2006
Actually, the Heart 6000 is NOT entirely tube output. The signal taken from the SM5872 dac chips is voltage out, and therefore uses each dac chip's i/v converter opamp stages before going to the coupling transformers. The Heart guys did not take this into account, and thus did not add any extra filtering or HF bypassing to the analog supply lines to the SM5872's, which can make quite a difference. Personally, I have found, among the several Heart 6000's I've worked on and upgraded, a LOT of variability in sound quality, not attributable to the tubes. I had two Heart 6000's that had identical work done(including the same version, grey-cased transformers throughout), and, even with Dutch Amperex 6DJ8's installed, one sounded dismal & the other sounded quite nice. I did a lot of work on sorting out what was causing the difference, and either there's just a bad amount of variability in the coupling transformers, or there's a really unacceptable amount of batch variability in the SM5872 dac chips.
  Reply With Quote
Old 12th October 2009, 06:00 PM   #6
diyAudio Member
 
Join Date: Oct 2007
Location: Cologne
Hi Stephen (right?),

wow, so you know the Heart 6000CDP. Thats great. Thanks for your reply :-)

Very strange behavior. Might be the reason for some good reviews of the Marantz CD6000(OSE), even I wasn't happy at all with this CDP, when I bought it almost 10 years ago. It was a boring, dull sounding player. My expectations were much higher. So I did some mods, years ago, which resulted in better dynamics, more freshness and power. Even then I thought, this can't be the endÖ..

Yes I know, the SM5872 is a V-out DAC. Apparently it should be one of the better V-out types, referring to Thorsten Loesch, who wrote in his paper 'Thermionic Valvestages':

The generation before the latest designs is a bit more heterogeneous. We find many DACís with differential output but various different schemes of
on chip filtering and buffering and the occasional DAC with a single-ended output and really good on chip analogue silicon. In the latter case
(CS4328, certain OEM Products) the same as above holds true. For all other differential output DACís the circuit shown further down should/could
be used and should work under any conditions, offering good filtering and ensuring an easy load for the DAC Chip, keeping distortion low.
Finally, the original first-generation timeslicing DACís have no analogue stages. Some of these are early Philips chipís as used in the QED "Digit"
external DAC, also the NPC SM5872 (Marantz CD-53/57/63/67 among others) and NPC SM5864 (Arcam Alpha - lower Models) and surprisingly
many OEM chips even in current use. All these will require some buffering and due to the lack of on chip filtering a direct connection without any
additional lowpass filtering a direct connection even to high impedance inputs is not advisable. Again, the circuit shown below will work just fine.


Thatís why I started thinking about adding tubes to this player. Thorstens circuit includes filtering and puts load to the balanced signal too, although this signal is not used by the following stage. Almost perfect, if there wouldn't be the second channel per DAC.

You'd probably know, which outputs (of the 4 per DAC (2 pos and 2 neg )) the Heart guys used to feed the transformer with. It would be great if I could get this info. This would help me deciding how to move on.

Thanks a lot

Ernst
Attached Images
File Type: jpg TL.jpg (51.9 KB, 487 views)
  Reply With Quote
Old 12th October 2009, 07:58 PM   #7
diyAudio Member
 
Join Date: Oct 2006
I never got a schematic for the 6000 series, nor do I currently have one of the players on hand, but, as I recall, the 5872's were connected much like Pioneer does with their PD65/75 level 'pulseflow' stereo dacs, where they sum R+ & L+ via some small value resistors, and do likewise with the negative outs. This would certainly make sense with the 5872, given what Loesch says about output current(implies, at least) on these. BTW, the datasheet for the SM5872(which I googled easily) DOES show analog buffer(at least) stages before the chip's outputs, and I definitely can attest to the benefit of being very generous with the size & quality of supply caps next to the analog supply line pins.
As you may guess, I believe one must first find a 6000 that actually sounds OK to start with, as the dismal sounding ones tend to retain some of that no matter what upgrades one does to them.
  Reply With Quote
Old 12th October 2009, 09:44 PM   #8
Thomo is offline Thomo  United Kingdom
diyAudio Member
 
Thomo's Avatar
 
Join Date: Nov 2006
Hi guys, I don't know much about valve output for the cd6000, but in my experience, adding a really good, low-noise reg to power the dacs and separator chip is the only way I managed to remove the ragged treble sound, and this was after clocking too. There's a 7805 reg on the board that does the job already, but with a really good reg in there, the difference is amazing. I've done this mod two or three times now to different cd6000's.

Cheers, Lee.
  Reply With Quote
Old 13th October 2009, 12:33 AM   #9
ilimzn is offline ilimzn  Croatia
diyAudio Member
 
Join Date: Feb 2005
Location: Zagreb
SM5872 requires external subtractio of the DAC+ and DAC- outputs:
DACOUT=(DAC+)-(DAC-).
If differential out is to be preserved, you must do this twice:
DACOUT+=(DAC+)-(DAC-)
DACOUT-=(DAC-)-(DAC+)

The reason this operation must be done to gt the full DAC performance is twofold:

Firstly, the DAC outputs are 'quasi-differential' - this is the term used in the data sheet. The outputs are pulses 0 to 12 MCLK cycles in length occuring every 12 MCLK cycles (0% to 100% duty cycle, at 32FS). Odd lengths of pulses (1, 3, 5, etc) cannot be symetrically centered around the middle of each 12-pulse PWM modulation cycle, so if something is not done about that, each odd length PWM pulse is 1 MCLK cycle out of phase - i.e. a 1 MCLK cycle jitter correlated with the output level is generated. To prevent this, the DAC+ and DAC- outputs are out of phase by 1 cycle for odd length combinations, and recreating a single ended pulse from these results in a 'halftoned' pulse which is centered around the midde of each 12 MCLK PWM modulation cycle, preventing the introduction of jitter. Becasue this jitter is introduced at the low resolution end of the sigma-delta DAC, the noise and distortion generated when only one of the differential outputs is used, without subtraction, is very high (and very audible).

Secondly, becuase the rate of PWM pulses is quite low (32Fs, whereas 64, 96, 128 or 192 Fs or even higher is more common), it appears that quite deep dithering/noise shaping is used to get adequate performance. Based on measuring peak values of the signal at the DAC+ and DAC- outputs, vs RMS post filtering, it appears that as many as 4 levels out of the 13 possible PWM output levels can contain residual quantizing noise.
Even though this is pushed higher in the spectrum by noise shaping, the low rate of PWM pulses means the rise of the noise floor is noticeable while still inside the audio band, rising as you go to the high end of it. In order to considerably reduce this problem, the noise shaping is actually common mode across DAC+ and DAC- outputs, and some care was obviously taken to prevent colision of this scheme with the quasi-differential out-of phase PWM trick described above. When only one DAC output is used without subtraction, the noise is clearly audible for a digital zero at the DAC input (this largely prevents jitter artifacts described in the previous paragraph from appearing as added noise). When proper subtraction is used, the DAC stays dead silent within the audio band.

In other words, if you only use one output, you get a large loss in resolution, a large increase in residual noise and an even larger introduction of jitter - 1 MCLK correlated jitter in most implementations means 59ns peak to peak jitter, compare this with picosecond speciffications on MCLK source clocks needed to get the most out of the DAC. So, with all respect to Thorsten, the above schematic is NOT recommended for this particular DAC, even though it is electrically correct.
  Reply With Quote
Old 13th October 2009, 09:18 AM   #10
diyAudio Member
 
Join Date: Oct 2007
Location: Cologne
Quote:
Originally Posted by stephensank View Post
I never got a schematic for the 6000 series, nor do I currently have one of the players on hand
If you ever needed, I could support you with the manual, showing the scheme.
Quote:
Originally Posted by stephensank View Post
but, as I recall, the 5872's were connected much like Pioneer does with their PD65/75 level 'pulseflow' stereo dacs, where they sum R+ & L+ via some small value resistors, and do likewise with the negative outs.
Interesting Idea - I'll look after the manual.
Quote:
Originally Posted by stephensank View Post
BTW, the datasheet for the SM5872(which I googled easily) DOES show analog buffer(at least) stages before the chip's outputs, and I definitely can attest to the benefit of being very generous with the size & quality of supply caps next to the analog supply line pins.
As you may guess, I believe one must first find a 6000 that actually sounds OK to start with, as the dismal sounding ones tend to retain some of that no matter what upgrades one does to them.
Really strange but might be that way. After my early mods, good rectifiers - very good electrolytics - removed output elkos - removed mute transistors, slight damping of the chassis most people (the normal ones, not addicted ones ) would be happy with it. Unfortunately I've heard a few really good sounding CDPs.
Quote:
Originally Posted by Thomo View Post
Hi guys, I don't know much about valve output for the cd6000, but in my experience, adding a really good, low-noise reg to power the dacs and separator chip is the only way I managed to remove the ragged treble sound, and this was after clocking too. There's a 7805 reg on the board that does the job already, but with a really good reg in there, the difference is amazing. I've done this mod two or three times now to different cd6000's.
I thought about doing that, but never did so far. Good advice, thanks - will do since it is a rather easy mod.

Quote:
Originally Posted by ilimzn View Post
SM5872 requires external subtractio of the DAC+ and DAC- outputs:
DACOUT=(DAC+)-(DAC-).
If differential out is to be preserved, you must do this twice:
DACOUT+=(DAC+)-(DAC-)
DACOUT-=(DAC-)-(DAC+)

The reason this operation must be done to gt the full DAC performance is twofold:

Firstly, the DAC outputs are 'quasi-differential' - this is the term used in the data sheet. The outputs are pulses 0 to 12 MCLK cycles in length occuring every 12 MCLK cycles (0% to 100% duty cycle, at 32FS). Odd lengths of pulses (1, 3, 5, etc) cannot be symetrically centered around the middle of each 12-pulse PWM modulation cycle, so if something is not done about that, each odd length PWM pulse is 1 MCLK cycle out of phase - i.e. a 1 MCLK cycle jitter correlated with the output level is generated. To prevent this, the DAC+ and DAC- outputs are out of phase by 1 cycle for odd length combinations, and recreating a single ended pulse from these results in a 'halftoned' pulse which is centered around the midde of each 12 MCLK PWM modulation cycle, preventing the introduction of jitter. Becasue this jitter is introduced at the low resolution end of the sigma-delta DAC, the noise and distortion generated when only one of the differential outputs is used, without subtraction, is very high (and very audible).

Secondly, becuase the rate of PWM pulses is quite low (32Fs, whereas 64, 96, 128 or 192 Fs or even higher is more common), it appears that quite deep dithering/noise shaping is used to get adequate performance. Based on measuring peak values of the signal at the DAC+ and DAC- outputs, vs RMS post filtering, it appears that as many as 4 levels out of the 13 possible PWM output levels can contain residual quantizing noise.
Even though this is pushed higher in the spectrum by noise shaping, the low rate of PWM pulses means the rise of the noise floor is noticeable while still inside the audio band, rising as you go to the high end of it. In order to considerably reduce this problem, the noise shaping is actually common mode across DAC+ and DAC- outputs, and some care was obviously taken to prevent colision of this scheme with the quasi-differential out-of phase PWM trick described above. When only one DAC output is used without subtraction, the noise is clearly audible for a digital zero at the DAC input (this largely prevents jitter artifacts described in the previous paragraph from appearing as added noise). When proper subtraction is used, the DAC stays dead silent within the audio band.

In other words, if you only use one output, you get a large loss in resolution, a large increase in residual noise and an even larger introduction of jitter - 1 MCLK correlated jitter in most implementations means 59ns peak to peak jitter, compare this with picosecond speciffications on MCLK source clocks needed to get the most out of the DAC. So, with all respect to Thorsten, the above schematic is NOT recommended for this particular DAC, even though it is electrically correct.

So does this mean that best I leave the outputsage, as it is or, if I wanted a tubed one, I have to build right in the way the current silicium one is?

Thanks a lot. Even I have to think about that hard, I already have learned a lot.

Cheer Ernst
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Help with PSU for tube DAC output stage luvdunhill Tubes / Valves 1 6th May 2008 09:10 AM
Tube output stage for DAC ackcheng Tubes / Valves 10 14th January 2004 12:13 PM
dac tube output stage..pls help. tone Tubes / Valves 80 25th March 2003 06:13 PM
need help on tube output stage.. tone Digital Source 2 19th March 2003 04:52 PM


New To Site? Need Help?

All times are GMT. The time now is 06:51 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright ©1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2