Re: Power Supply
I did experiments with batteries several times and posted the results on the Audio Asylum. Please do a search over there, AA has an excellent searchengine.
My experience is that batteries kill the dynamics of the music making it flat sounding. As an added "benefit" batteries always seems to go dead in the weekend when shops are closed.
Shall I propose that other forum members email you the manual as I have a pretty slow connection to the internet and sending the 7MB file takes 45 minutes for me.
EUVL said:
I did experiments with batteries several times and posted the results on the Audio Asylum. Please do a search over there, AA has an excellent searchengine.
My experience is that batteries kill the dynamics of the music making it flat sounding. As an added "benefit" batteries always seems to go dead in the weekend when shops are closed.
Shall I propose that other forum members email you the manual as I have a pretty slow connection to the internet and sending the 7MB file takes 45 minutes for me.
Power supply first tests and listening…
I experiment quite a lot on the power supply… More precisely on a power supply for the output stage described in this thread and within the accuracy of my sound system. All the following comments on the settings and tunings apply on the power supply for this circuit within the XB940.
1) Transformer : I begun the test with a classic common toroidal transformer 2*12V no brand on it, the goal was to compare batteries to a power supply, and progressively tuning it to a battery level, keeping in mind to use later the sony “EI-analog” transformer.
2) Rectifiers : One of the most important parameters, I tried schottky, fast, ultrafast, soft recovery… After trying more than a dozen of them, what seems the most relevant to sound quality is the soft recovery and speed parameters, maybe the ta/tb in the trr. Schottky are not the best ones, like I thought, or even the amplitude of the reverse current. The ringing characteristic (frequency-amplitude) degrades considerably the quality. The best ones are the BYV27, MBR1100, RURD460, the last one sounded the best. I must say that the BYV27 are not bad, but depends enormously on the brand the batch… Philips are the best ones for this reference. Trying snubbers don’t solve the problem very much, and sometimes worse it. BTW if someone has an idea for SMD rectifier parts… I tried ES1D, BYG22D, MURS120T3, not so good performers as the preceding ones.
3) RC LC LCLC LCRC etc…: After various tests a simply RC sounds the best or almost equal to an LC filtering, in the best case L brings nothing more. ( I tried small L about 2.2mH DCR=0.5R.)
4) Active filtering : To achieve a final good filtering, and more specially noise filtering, I use a Mosfet “capacitor multiplier”.
The ground scheme is extremely important, maybe more, than the devices types or brands themselves and even the circuit. (Star ground / Decoupling caps position.)
In short:
Fast, ultra fast, hyper fast, soft recovery characteristic for rectifiers.
RC filtering, small R, 2 to 10R for followers preamplifiers (small currents/ class A)
Active filtering of noise. Ferrite beads sometimes well placed can help a lot… On this point maybe Charles Hansen or Jocko Homo could tell us more about the efficiency of ferrites over time?
This circuit doesn’t need an accurate regulation, the power supply must be stable and noiseless, that’s all.
The output is about a couple of volts lower than the input voltage (Essentially Vgs).
The second schematic use a TL431, for an adjustable V output.
After trying various designs, here the power supply schematics tuned by listening :
XB-940PS.
XB-940PS_Adj.
Ground scheme is quite difficult to optimize on this player and appears to be important on sound quality (air, emotion, softness) so as the pcb is not easy to modify, as just as ground wiring between pcboards, I wonder whether I will go for a floating output (no reference to player ground on the output stage). Just keep common ground for Clock and DAC, as the outputs are balanced… Just a little modification on the fet’s followers circuit…
I experiment quite a lot on the power supply… More precisely on a power supply for the output stage described in this thread and within the accuracy of my sound system. All the following comments on the settings and tunings apply on the power supply for this circuit within the XB940.
1) Transformer : I begun the test with a classic common toroidal transformer 2*12V no brand on it, the goal was to compare batteries to a power supply, and progressively tuning it to a battery level, keeping in mind to use later the sony “EI-analog” transformer.
2) Rectifiers : One of the most important parameters, I tried schottky, fast, ultrafast, soft recovery… After trying more than a dozen of them, what seems the most relevant to sound quality is the soft recovery and speed parameters, maybe the ta/tb in the trr. Schottky are not the best ones, like I thought, or even the amplitude of the reverse current. The ringing characteristic (frequency-amplitude) degrades considerably the quality. The best ones are the BYV27, MBR1100, RURD460, the last one sounded the best. I must say that the BYV27 are not bad, but depends enormously on the brand the batch… Philips are the best ones for this reference. Trying snubbers don’t solve the problem very much, and sometimes worse it. BTW if someone has an idea for SMD rectifier parts… I tried ES1D, BYG22D, MURS120T3, not so good performers as the preceding ones.
3) RC LC LCLC LCRC etc…: After various tests a simply RC sounds the best or almost equal to an LC filtering, in the best case L brings nothing more. ( I tried small L about 2.2mH DCR=0.5R.)
4) Active filtering : To achieve a final good filtering, and more specially noise filtering, I use a Mosfet “capacitor multiplier”.
The ground scheme is extremely important, maybe more, than the devices types or brands themselves and even the circuit. (Star ground / Decoupling caps position.)
In short:
Fast, ultra fast, hyper fast, soft recovery characteristic for rectifiers.
RC filtering, small R, 2 to 10R for followers preamplifiers (small currents/ class A)
Active filtering of noise. Ferrite beads sometimes well placed can help a lot… On this point maybe Charles Hansen or Jocko Homo could tell us more about the efficiency of ferrites over time?
This circuit doesn’t need an accurate regulation, the power supply must be stable and noiseless, that’s all.
The output is about a couple of volts lower than the input voltage (Essentially Vgs).
The second schematic use a TL431, for an adjustable V output.
After trying various designs, here the power supply schematics tuned by listening :
XB-940PS.
XB-940PS_Adj.
Ground scheme is quite difficult to optimize on this player and appears to be important on sound quality (air, emotion, softness) so as the pcb is not easy to modify, as just as ground wiring between pcboards, I wonder whether I will go for a floating output (no reference to player ground on the output stage). Just keep common ground for Clock and DAC, as the outputs are balanced… Just a little modification on the fet’s followers circuit…
What I felt about the ground layout and the ground distribution among the differents pcboards was right, so as it appears to me impossible to optimize the whole player on this point, It will take so much time to tune the ground distribution, listen, mod, listen, measure etc… taking off each time the pcboard modify it, re_putting it on the player, you see…
So, unless remaking all the pcboards, the solution, which not only solves the problem but moreover constitutes a very clear improvement, is to go to a floating mode at the output of the VC24+.
I tried this today and wow!!! The improvement is in the same order of magnitude than bypassing the current pulse converters. For those who have done it, they know what I mean.
I will post later the modifications to carry out on the output stage and power supply.
To test this mod I used the analog power supply trafo for the clock (like Sony did, but with my "reg" circuit:TL431/Mosfet).
To avoid common ground with the player, the easiest solution to test this mod was to re_use batteries, I don’t think the improvement is due to them, taking into account the tests carried out before, on the power supplies (no difference between them).
If I want to use the Sony analog trafo for the ouput stage, I must use the digital trafo (DAC supply) for the clock too, so ??? Maybe with a good supply circuit, it will be o.k. I’ll test this, tomorrow I hope.
If the results are not good, I will go for another trafo just for the clock.
M. Kohjin Yamada, find this link about the SACD test discs, many thanks.
Philips SACD test disc.
I need help on a TDS820 Tektronix scope, the model is discontinued and they don’t repair it any more, if someone in the forum can give me some advises or is related to Tektronix… The problem is in the VCO of the acquisition card, for some special measuring this model is very useful (6/8 Ghz).
So, unless remaking all the pcboards, the solution, which not only solves the problem but moreover constitutes a very clear improvement, is to go to a floating mode at the output of the VC24+.
I tried this today and wow!!! The improvement is in the same order of magnitude than bypassing the current pulse converters. For those who have done it, they know what I mean.
I will post later the modifications to carry out on the output stage and power supply.
To test this mod I used the analog power supply trafo for the clock (like Sony did, but with my "reg" circuit:TL431/Mosfet).
To avoid common ground with the player, the easiest solution to test this mod was to re_use batteries, I don’t think the improvement is due to them, taking into account the tests carried out before, on the power supplies (no difference between them).
If I want to use the Sony analog trafo for the ouput stage, I must use the digital trafo (DAC supply) for the clock too, so ??? Maybe with a good supply circuit, it will be o.k. I’ll test this, tomorrow I hope.
If the results are not good, I will go for another trafo just for the clock.
M. Kohjin Yamada, find this link about the SACD test discs, many thanks.
Philips SACD test disc.
I need help on a TDS820 Tektronix scope, the model is discontinued and they don’t repair it any more, if someone in the forum can give me some advises or is related to Tektronix… The problem is in the VCO of the acquisition card, for some special measuring this model is very useful (6/8 Ghz).
I believe someone asked about the service manual in pdf form. Download it from Sony here (7.00 MB PDF). As with all Sony service manuals, it includes block diagrams, schematics, parts lists, disassembly instructions, signal waveforms, etc.
Batteries supply
Yes I also tried Nicads. Caps do not help much.
Pan said:Elso,
you know they have invented these rechargable thingys made of lead and stuff?.
Seriously though, did you find the lack of dynamics being the case even when batteries were bypassed with decent capacitance?
/Peter
Yes I also tried Nicads. Caps do not help much.
Floating Ground
> So, unless remaking all the pcboards, the solution, which not only solves the problem but moreover constitutes a very clear improvement, is to go to a floating mode at the output of the VC24+.
Richard,
Thank you again for the many useful information you are posting. I do not quite get what you mean though with flaoting the output of the VC24. Maybe I just wait for your final modification instructions.
Patrick
> So, unless remaking all the pcboards, the solution, which not only solves the problem but moreover constitutes a very clear improvement, is to go to a floating mode at the output of the VC24+.
Richard,
Thank you again for the many useful information you are posting. I do not quite get what you mean though with flaoting the output of the VC24. Maybe I just wait for your final modification instructions.
Patrick
Hi, macboy,
Many many thanks, a lot of people here will be very happy with your XB940 pdf schematics, me the first! I have paper copied this manual at least 6 times!
Hi, Patrick,
What I mean by floating output is to not reference to ground, the balanced outputs of the VC24+ (just get the common mode), because between the ground and the audio outputs +/- of this DAC appears out, unwanted “noises” (power supplies, clock, digital and RF signals) who interferes on the audio signal, the only way to get rid of this, is a perfect ground or not ground at all.
As it appears really very boring, difficult and fastidious to redraw all the ground scheme of this player unless redesigning all the pcboards, the easiest way, is to not use the player ground.
The solution in my design (audio-trafo output) is to reference the input fet’s at half the power supply by a resistor dividing network between V+ and V-, and connecting this power supply middle point to the gates of the input fet’s with 2 high value resistors (1 MEG). Of course the power supply is not connected to the player ground, for this, I use the analog-trafo of the player, providing a few mod’s, with a schematic it will be clearer, I’ll post it soon. (c.f. below)
Tested this last days:
- Digital trafo for DAC and Clock. One reg per “function” (no difference with a complete separate supply just for the clock! Even with batteries//capacitor, curious ???
- Analog power supply trafo for output stage (with no common ground with the player).
Always as good as on my last post, and same quality, with my power supply, but now that I have increased the “resolution” by just using common mode, I can hear a very little difference on batteries vs. my reg power supply on the analog output stage, it’s a little better (more involved, more on the stage with the musicians), with batteries. I will try to improve that the next days…
Updated schematics, when I end up with the analog power supply.
Many many thanks, a lot of people here will be very happy with your XB940 pdf schematics, me the first! I have paper copied this manual at least 6 times!
Hi, Patrick,
What I mean by floating output is to not reference to ground, the balanced outputs of the VC24+ (just get the common mode), because between the ground and the audio outputs +/- of this DAC appears out, unwanted “noises” (power supplies, clock, digital and RF signals) who interferes on the audio signal, the only way to get rid of this, is a perfect ground or not ground at all.
As it appears really very boring, difficult and fastidious to redraw all the ground scheme of this player unless redesigning all the pcboards, the easiest way, is to not use the player ground.
The solution in my design (audio-trafo output) is to reference the input fet’s at half the power supply by a resistor dividing network between V+ and V-, and connecting this power supply middle point to the gates of the input fet’s with 2 high value resistors (1 MEG). Of course the power supply is not connected to the player ground, for this, I use the analog-trafo of the player, providing a few mod’s, with a schematic it will be clearer, I’ll post it soon. (c.f. below)
Tested this last days:
- Digital trafo for DAC and Clock. One reg per “function” (no difference with a complete separate supply just for the clock! Even with batteries//capacitor, curious ???
- Analog power supply trafo for output stage (with no common ground with the player).
Always as good as on my last post, and same quality, with my power supply, but now that I have increased the “resolution” by just using common mode, I can hear a very little difference on batteries vs. my reg power supply on the analog output stage, it’s a little better (more involved, more on the stage with the musicians), with batteries. I will try to improve that the next days…
Updated schematics, when I end up with the analog power supply.
Clocks on an XB940
Elso,
I remember you mentioning in one of the threads (can't find it now) that there are two clocks that need synchronising in a typical Sony, at 45mHz and 16.9 MHz. And you mentioned the Swoboda clock.
I had a look at the schematics of the XB940 yesterday (Thanks Richard), and saw 2 clocks at 45mHz, 27MHz and 12.5MHz, but they don't seems to be synchronised in any way.
Would you be kind enough to shred some light on the topic ?
Thanks,
Patrick
Elso,
I remember you mentioning in one of the threads (can't find it now) that there are two clocks that need synchronising in a typical Sony, at 45mHz and 16.9 MHz. And you mentioned the Swoboda clock.
I had a look at the schematics of the XB940 yesterday (Thanks Richard), and saw 2 clocks at 45mHz, 27MHz and 12.5MHz, but they don't seems to be synchronised in any way.
Would you be kind enough to shred some light on the topic ?
Thanks,
Patrick
Re: Clocks on an XB940
Hi Patrick,
There is a non-synchronised 45.1584MHz clock at the digital filter/DA converter (IC301)
There is a 27 MHz clock at the servo digital signal processor (IC607), also not synchronised.
The 12.5 MHz clock is for the system controller.
In my opinion it is worthwhile to replace the 45.1584 and 27 MHz clocs for sound improvement.
My comment about the synchronised clocks was for much older Sony models.
Please contact me by email for a working solution with my clock at these frequencies.
EUVL said:
Hi Patrick,
There is a non-synchronised 45.1584MHz clock at the digital filter/DA converter (IC301)
There is a 27 MHz clock at the servo digital signal processor (IC607), also not synchronised.
The 12.5 MHz clock is for the system controller.
In my opinion it is worthwhile to replace the 45.1584 and 27 MHz clocs for sound improvement.
My comment about the synchronised clocks was for much older Sony models.
Please contact me by email for a working solution with my clock at these frequencies.
> Please contact me by email for a working solution with my clock at these frequencies.
Elso,
Many thanks for the reply.
I already have a copy of your 45MHz version, so I guess I only need the 27MHz one. Am on business trip (in Eindhoven) so it won't be a few days. But surely will get in touch.
Cheers,
Patrick
Elso,
Many thanks for the reply.
I already have a copy of your 45MHz version, so I guess I only need the 27MHz one. Am on business trip (in Eindhoven) so it won't be a few days. But surely will get in touch.
Cheers,
Patrick
One year on, it is finally finished. Thanks to Richard for his insirations and motivations.
Sounds great to my ears (reference Meridian 508-24). No plans for further changes, at least for some time.
Great value for money at less than 500 Euros, including a second hand player from Ebay.
Patrick
Sounds great to my ears (reference Meridian 508-24). No plans for further changes, at least for some time.
Great value for money at less than 500 Euros, including a second hand player from Ebay.
Patrick
Attachments
The Perez buffer at 9mA bias, with a different filter (3rd order 100kHz passive). Dual mono, separate PSU's (current sourced discrete shunt regulators, +/- 15V). Balanced output.
Clock is Tent (gift from Richard with sincere thanks, otherwise no other reason), 8.4V NiMH battery powered with a separate current sourced shunt regulator in between. House in separate nickel plated aluminium housing (top left hand side).
Signal cables from DAC teflon twinax 75 ohm shielded, from clock 50 ohm telfon coax.
Clock is Tent (gift from Richard with sincere thanks, otherwise no other reason), 8.4V NiMH battery powered with a separate current sourced shunt regulator in between. House in separate nickel plated aluminium housing (top left hand side).
Signal cables from DAC teflon twinax 75 ohm shielded, from clock 50 ohm telfon coax.
Attachments
Hello,
Compliments, your modifications sure looks professional.
I'm about to install the Tent-clock in my XB940 and I read about this nice thread. Would you be willing to pass information on the output section that you've made? Or may be you would even have a spare set of pcb's left?
If you want to respond off-list you can mail me at: vrhff04_at_xs4all.nl (please replace _at_ by @).
Best regards, Arjen.
Compliments, your modifications sure looks professional.
I'm about to install the Tent-clock in my XB940 and I read about this nice thread. Would you be willing to pass information on the output section that you've made? Or may be you would even have a spare set of pcb's left?
If you want to respond off-list you can mail me at: vrhff04_at_xs4all.nl (please replace _at_ by @).
Best regards, Arjen.
PSU of Perez buffer
Hello Patrick,
In post 37 you write:
The Perez buffer at 9mA bias, with a different filter (3rd order 100kHz passive). Dual mono, separate PSU's (current sourced discrete shunt regulators, +/- 15V). Balanced output.
Would you mind sharing the schematic of the current sourced discrete shunt regulators, +/- 15V?
Best regards, Arjen
you can mail me off list when you prefer:
vrhff04_at_xs4all.nl
NB. please replace _at_ by @
Hello Patrick,
In post 37 you write:
The Perez buffer at 9mA bias, with a different filter (3rd order 100kHz passive). Dual mono, separate PSU's (current sourced discrete shunt regulators, +/- 15V). Balanced output.
Would you mind sharing the schematic of the current sourced discrete shunt regulators, +/- 15V?
Best regards, Arjen
you can mail me off list when you prefer:
vrhff04_at_xs4all.nl
NB. please replace _at_ by @
Someone posted this a while ago, a link to sony schematics:
" Post #26
I believe someone asked about the service manual in pdf form. Download it from Sony here (7.00 MB PDF). As with all Sony service manuals, it includes block diagrams, schematics, parts lists, disassembly instructions, signal waveforms, etc."
I am trying the link, but it asks for login, etc and won't let me access anything. Really like to get ahold of some sony manuals, anyone help me out?
" Post #26
I believe someone asked about the service manual in pdf form. Download it from Sony here (7.00 MB PDF). As with all Sony service manuals, it includes block diagrams, schematics, parts lists, disassembly instructions, signal waveforms, etc."
I am trying the link, but it asks for login, etc and won't let me access anything. Really like to get ahold of some sony manuals, anyone help me out?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.