Upgrading Philips DVD 963 SA

Jump to Latest
#103
The AD8032 output OP AMP has a maximum rated power supply range of +/- 6V.
Philips is running it already at +/- 8V.
The +8V line is also providing bias current for the I/V stage.
I was not ready to replace this yet, so I kept the same operating voltages.

Also, the large new BG caps around the 8032 now make it difficult to replace...

EG
 
#105
Hi Eganz,

In the past days I did some of the same modifications as you did.

They were:

- Removing the van Medevoort caps at the front channel outputs (two cheap 470 uF caps with 0.22 uF MKT and a small Silver Mica), and I substituded them by two Blackgate N 220 uF 6.3 Volt in super-e configuration.

- In the SPSU I exchanged the 100 uF 400 V Philips cap for a Blackgate VK 150 uF 350 V cap. I also replaced #2235 with 220 uF 16V BG-FK, and #2285 with 100 uF 16V BG-N

- On the DAC pcb I swapped #2635 for a 100 uF 16V BG-N.


My impressions are:

- Much better sound than with the van Medevoort mods. With the van Medevoort mods the player sounds rather fat in the bass, it lacked midrange and the trebel was irritatingly predominant and fatiqueing. Now there is mor midrange, a well controlled bass and a much more liquid and clean treble without any harshness. Details come much more natural and within a coherent soundstage. Although terribly expensive, these Blackgates do a terrific job. I am curious how the sound will further develope when they start to break in...

The only thing i would wish for is a further increase of midrange fullness. I hope that this will be achieved by adding an extra linear PSU for the analogue output...

So far I did not replace #2312 (DAC bypass) and #2442, #2443, #2444, #2445 (op amp bypass) with blackgates. I noticed that there is very little space for thicker caps. What BGs did you use here? Van Medevoort had added the 0.22 uF MKTs here as well. Do you think I should swap these for BG's? Does it matter much here?

There is one other important difference now: Before, I preferred to listen to CD without upsampling (it sounded a bit less thin and artificial). Now after the BG mods, the upsampling to 24 bit 96 kHz sounds much better than without upsampling. More ambience and dynamics! Isn't that very strange indeed?

Does upsampling indeed influence the importance of a jitterfree clock, like those Danish state in their ridiculous advertisement? They did not give any scientific argument, but maybe they had this same experience?

I do understand whether people like Guido Tent find upsampling highly suspicious, but when things start to sound so much better than without upsampling, it becomes interesting to wonder what is actually achieved by those mysterious chips like the AD 1895?

Anyway, I finally begin to like my DVD 963 SA. It is getting closer and closer to the (van Medevoort-) modded SACD 1000.

Regards,

Lucas
 
#106
Hi Lucas,

I also prefer the upsampling. Smoother, warmer, more ambience, and less residual harshness, closer to the wonderful SACD sound.

Replacing #2312 (DAC bypass) is useful.

Replacing #2442, #2443, #2444, #2445 (op amp bypass) with blackgates will have a large effect, as these are located right on the op amps. For me, these replacements produced great coherence and solidity in the sound. I used 470 uF BG STD 16V caps here. You may need to bend (dogleg) some leads to squeeze these in. Also, if you are planning to replace the output op amps, you might choose to do it before you install these large capacitors. On the other hand, once you install these capacitors, you might not need a new output op amp. If you are already adding a new power supply, you could probably get away with smaller value BG caps here.

And yes, I agree that a new linear power supply for the analog stage will provide a huge step up... this should be your first priority. The output voltage will of course depend on the output op amps that you want to use.

Eric
 
#107
OK, to give the 5V regulator for the DAC its own 8v reg instead of sharing with the 8v OP AMP power:

I replaced the surface mount regulator #7601, and removed resistor #3659 and removed jumper #4606 to isolate the +8V supply from the 5V regulator. Piped in the unregulated +13V to +8V regulator #7601, and reconnected the +8V and GND directly to the op amps.

Also removed unused parts #7400, 7420, 2412, 2413.
 
#108
DAC master clock

For SACD play, the clock is generated on the AV board (DAC master clock). The crystal oscillator circuit is fed by the +3V3_osc (+3.3V)supply, which has an LC filter. The clock then runs through several gates to the DAC. I replaced the stock 47 uF filter capacitor #2610 with a 100 uF BG to provide quieter power for the crystal oscillator. This improves the musicality.

I am currently running two grounds to the AV board. The new +- 8V supply brings an analog ground, and I also connect to the tab at the edge of the board which provides a digital ground. These both run back to the star ground at the IEC connector.
 
#109
Philips DVD-963SA audio mods summary

Power supply board:

Replace capacitors
# 2116 with 150uF, 350V BG VK
# 2215 with 100 uF, 16V BG STD (+12V)
# 2235 with 220 uF, 16V BG STD (+3.3V)
# 2255 with 470uF, 16V BG STD (+5V)
# 2285 with 100 uF, 16V BG STD (-12V)

Install Three-pronged filtered AC plug

Remove the progressive scan board and cable.

AV board:

Replace output capacitors #2421, 2429 with nonpolar 100uF 16V BG N.

Bypass caps #2205, 2206, 2213, 2301, 2307, 2309, 2606, 2608, 2610, 2614, 2616, 2619, 2625, 2626, 2627, and 2628 with 0.33 uF metallized polyester or PP film caps.

Replace
#2312 with 100 uF, 16V BG STD (DAC bypass)
#2442, 2443, 2444, 2445 with 470 uF, 16v STD BG (op amp bypass).
#2635 with 100uF, 16V BG STD (+5V).
#2610 with a 100 uF, 16V BG STD (filter for crystal oscillator).
#2612 with 100 uF, 16V BG STD
#2613 with 47uF, 16V BG STD
#2620 with 100 uF, 16V BG STD

Add new analog supply: +/- 9 VAC, 1.1 A toroidal transformer (digikey #TE62041-ND), diode bridge, and 0.25 A fuse wired to 2x 1000uF BG STD caps, NJM7808FA and NJM7908FA +8V and -8V regulators, and then 2x 470uF BG STD caps.

Remove jumper #4604, resistor #3659, and jumper #4606 to isolate the original +/- 8V supplies. Bring in the unregulated +13V to input of +8V regulator #7601, and connect the new regulated +/-8V and GND directly to the op amps.

Remove unused parts #7400, 7401, 7420, 2412, 2413.


Use two grounds to the AV board. The new analog supply brings an analog ground. Also connect to the tab at the edge of the board which provides a digital ground. These both run back to star ground at the IEC connector.


Main Board:
Add 47 uF BG and 1uF ceramic at #2910 (+5V)
Add 100 uF BG at #2912 (+12V) on the top of the board.
 
#111
Re: check out this mods !

Jaap said:
Highly interesting guys:

http://www.audioasylum.com/forums/hirez/messages/166420.html

Gurus, Guido Tent, Eganz, give us earthlings advice: buy or not buy, true or false.
Wait for vacuumstate or act now:dodgy:
Click to expand...

Hi

On the analogue supply: OK, but the SMPS is still in (I am doing a linear equivalent, form fit and function). No place for SMPS in audio

On the piggy backing: Scarry, but it could work (ever parallellled 2 car batteries ?)

On the clock: They still think ppm has to do with jitter. The clock architecture in the 963 is a mess, and I will fix it with a dedicated upgrade. Be patient, can't handle all developments in a short timeframe.....

Guido
 
#113
In the mentioned Forum there was an argument going about the use of the 3 different clock-frequencies:

"You are wrong. Please read pg 22 of AD1895 data specs - Tze Guan 10:48:51 01/22/04 (5)

In Reply to: Sorry, I really meant the SRC AD1895/96, not the DAC AD1955... posted by Alex Peychev on January 22, 2004 at 10:00:35:

It specifically says that if your masterclock to the sample rate converter is less than 30Mhz, you can only operate up to 96Khz upsampling.
For 192Khz sample conversion, the only mode supported is the slave mode, and you need a masterclock > 30Mhz to do that.

In 963SA, the 24.576Mhz clock is used for

1. 96Khz upsampling master clock for AD1895sample rate convertor in Master mode.

2. Bit clock and LRCK generation in 192Khz mode (thru the syncronous counter IC, LVC161 I think.)

The 33.8688Mhz then is used as master clock for 192Khz mode only for AD1895A as it needs to operate in Slave mode and hence needs a master clock greater than 30Mhz for this.

In any case I have already measured the master clock pin of the AD1896A (mine is modded) of my 963SA in both 96Khz and 192Khz mode. I am correct. "


"No, you read pg 22 of the AD1895 or AD1896...... - Alex Peychev 11:53:13 01/22/04 (0)

In Reply to: You are wrong. Please read pg 22 of AD1895 data specs posted by Tze Guan on January 22, 2004 at 10:48:51:

....The clock CAN NOT be more than 30M. Since the AD1895/96 CAN NOT provide master clock in 192K mode (because it MUST work in "slave mode"), the Data Sheet calls for additional (optional) 24,567M clock to provide MCLK and LRCLK (divided) for the DAC. That is why the 963SA sounds worse in 192/24 mode. It is because the AD1985 switches to "slave mode" and the MCLK and LRCLK are provided by dividing the 24,567M clock.
I am not sure how to be more clear with you!? 33,8688 CAN NOT clock 192K as in slave mode as the AD1895/96 are SYNCHROUNOUS to the MCLK not asynchronous. You do the math, 33,8688 can not be divided by 128 in order to clock 192K.

128 x 192 is 24,576 it's THAT SIMPLE!

This is my last post here on this subject, but I highly recommend that you update your friend's site with the RIGHT information. "



Guido, who is right here?

Furthermore some other forummembers have serious doubts on the sincerity of the modder; sending in a player and cash in advance seems risky indeed. The price is quite low ($ 590) for all these mods, but if one looses both money and player it can become very expensive indeed... :devilr:

Personally I hesitate in soldering two extra DACs on top of the existing one. It is all so terribly small! Are there more folks with experience on soldering such tiny chips?

Regards,

Lucas.

P.S. The Blackgate orgy is doing wonderful things with the 963...🙂
 
#114
Lucas,

The whole DAC master clock thing comes straight out of the manual, but the actual story appears more complicated.

Tze Guan also wrote

"they change the architecture to:
27Mhz oscillator on the monoboard -----> ICS MK2703 PLL---->11.28Mhz for CD/SACD. ..."

This generally agrees with measurements posted by Jos v E
"I have done some measurements on the clock circuit of my DVD963....

A high-quality AD1955 audio DA-converter is used for the left and right Front channels. The 'master clock' of this chip (pin 27)
is 16.9MHz for CD-audio playback, 12.28MHz for DVD 5.1 audio, and 11.28MHz for MP3 and SACD playback.
This latter frequency is not generated with a nice crystal oscillator on the audio board, but seems to originate from a PLL circuit at the main driver-and-video board, passing through connectors and selection gates."


and with a post by Guido
"The clock architecture of 963 is a mess

The SACD clock stms from a LL chip driven by the clock from the drive (yuck). Furthermore, 2 other crystal clocks are present, and too many gates control all the clocks..."
 
#115
Second stage of modding

Hi Eganz,

A couple of days ago I did the second bunch of upgrades:


On the SPSU I did this time:

# 2215 with 100 uF, 16V BG N (+12V)
# 2255 with 470uF, 16V BG N (+5V)

I also added 0.33 uF MKT bypasses to all the other remaining caps on the SPSU.

On the previous occasion I had already done:

# 2116 with 150uF, 350V BG VK
# 2235 with 220 uF, 16V BG N (+3.3V)
# 2285 with 100 uF, 16V BG N (-12V)


On the DAC-board I did this time:

#2312 with 100 uF, 16V BG N (DAC bypass)
#2442, 2443, 2444, 2445 with 470 uF, 16v N BG (op amp bypass).

Space was narrow so I did have to dogleg the large 470 uF BG-N's indeed.

On the previous occasion I had already done:

#2635 with 100uF, 16V BG N (+5V).


On the main Board I added this time:
a 47 uF BG N at #2910 (+5V)
a 150 uF Oscon at #2912 (+12V) on the top of the board.


I also installed a Netfilter at the IEC-mains-input.


I find that these mods are very worthwile. The sound has gained considerably in weight and there is much more detail in the upper midrange.

It seems that these Blackgate Nonpolars give the excellent AD dac and AD opamps the silent powersupply that they need to sing and shine.

The sound is now such that I wouldn't like to exchange the CD-performance of the 963 to that of a Meridian 588, and that had been my reference for some time😀

Furthermore the sound of SACD has become really fabulous: What a natural resonance and ambience there is! Sometimes almost frightening real and immediate :smash:

I find it difficult to imagine how a seperate powersupply could further improve things now that the SPSU has been upgraded so much. But of course I want to try! 😉
I have already got the parts to build one.

However in my haste I could not find the jumper #4604, resistor #3659, and jumper #4606 that you mention in your post. :bawling:
Eganz, where should I look? Are they on the SPSU-board? Close to what other component?

Thanks for all your suggestions so far!

Regards,

Lucas
 
#116
Lucas,

Jumper #4604 is tiny and is next to the negative side of cap #2619 on the top of the AV board. When you remove this, you isolate the -8VA line for the op amps. The original -8V regulator is still used for various digital pulldowns.

Jumper #4606 is also tiny and is on the bottom of the board ~1 cm from resistor #3659 (which is ~ under cap#2612) in the direction towards part #7502 . This will isolate the +8V op amp supply. You can check with a DVM that you have the right ones.

The new +8V and GND were wired onto the bottom of the board at the op amps, and to the bypass capacitors. The new -8V was wired to cap #2445.

E
 
#118
OK, here is a very detailed discussion of the actual Phillips 963 clock scheme from Effective Audio Mod

http://effectiveaudiomod.com

So the 963 was designed in Singapore :bigeyes:

E

Philips 963SA Detailed Clocking Scheme and History
There has been some debate over the web on the audio clocking scheme used by 963SA. After reading thru the overall schematic of 963SA, measuring on the actual set and talking to the actual Philips designer of 963SA who is based in Singapore (the 963SA is designed in Singapore), the correct information is as below:

Clock Frequency Usage

24.576Mhz Audio Master Clock for CD-Upsampling
27Mhz Video Clock. Audio Master Clock for SACD/CD non-upsampling/DVD
33.8688Mhz Processor Clock for SR C AD1895A in 192Khz Upsampling.

In order of importance for front channel audio quality:

24.576Mhz (crystal based) on AV Board

Very important for CD Upsampling Mode 96Khz and 192Khz audio quality

Audio Master Clock for AD1955 DAC during 96Khz and 192Khz Upsampling mode

Processor Clock for Sample Rate Converter (SRC) AD1895A during 96Khz (AD1895A in master mode

Left/Right Clock and serial bit clock generation for SRC AD1895A and AD1955DAC during 192Khz Upsampling mode (SRC AD1895A in slave mode)


27Mhz oscillator on motherboard

Very Important for SACD, DVD, CD non-upsampling audio quality and video quality

Originally not used (in fact, for the 1st production run of maybe 2000pcs, it is not there at all), as the 27Mhz clock is supposed to be generated from the 16.9344Mhz clock from the AV Board thru PLL ICS612G. However, due to problem described in the section (33.8688Mhz on AV Board point 3), later used to improve performance.

Video processor Clock

Farouja P-scan encoder Clock

Base frequency for audio clock generation (SACD, DVD, CD non-upsampling) thru PLL MK2703. This clock is fed to PLL MK2703. PLL MK2703 will, depending on the material, output audio master clock (256FS) for SACD, CD(non-upsampling) and DVD 48Khz/96Khz


33.8688Mhz on AV Board:

Not really important at all

Processor Clock for SRC AD1895A during 192Khz Upsampling. Because the SRC AD1895A is in slave mode, all the data would be clocked out by the serial bit clock and left/right clock generated by the 24.576Mhz clock, which is used by the AD1955DAC as well (hence the data clocked out by the serial bit clock and Left Right Clock would be in sync for the DAC, as it is using the 24.576Mhz clock). There is no need for this processor clock to be in sync with the left right clock and serial bit clock for the SRC AD1895A. They could of course save cost and used the 24.576Mhz clock and need not use another clock, but using a higher speed clock greater than 30Mhz in this case gives a better performance for the SRC AD1895A accordingly measured by the designer of 963SA.

Originally intended to be Audio Master clock for AD1955DAC (384 FS by 33.8688Mhz divided by 2 using a flip flop to get 16.9344Mhz, and this frequency 16.9344Mhz to be fed to motherboard for 27Mhz clock generation thru the PLL on the motherboard ICS612G) for SACD and CD non-Upsampling. However, at the very last stages of development, it was discovered that the AD1955DAC has a problem with SACD (DSD format) for this clock frequency. Out of 100 play back, once or twice the audio output would be noisy. The problem is traced to the PLL of the AD1955DAC. Its internal clock is 256FS, and the 384FS is not a integer of the internal clock. As a result, sometimes the DAC internal PLL will miss-sync, creating typical audio master clock miss sync phenomenon (noisy audio output). Hence they have to abandon to use the use of this clock for SACD/CD non-upsampling and instead only use for processor clock for the SRC during 192Khz operation.
Click to expand...
 
#119
eganz said:
OK, here is a very detailed discussion of the actual Phillips 963 clock scheme from Effective Audio Mod

http://effectiveaudiomod.com

So the 963 was designed in Singapore :bigeyes:

E
Click to expand...



Hi Eganz,

In this thread: http://www.audioasylum.com/scripts/t.pl?f=hirez&m=166420

there is a long discussion going on with respect to what the man behind EffectiveAudioMod claims.
As it stands, it seems that he is totally wrong about the clock architecture of the 963. :xeye:

Furthermore he seems to be the same person as the socalled Mr Tze Guan, who first claims being an ordinary customer, but later in the thread seems so much involved that he likely is the owner of EffectiveAudioMods.

Personally I cannot judge who was right here. Maybe guys like Guido Tent and others who know more about digital clocking can explain who is right: Tze Guan or Alex Peychev?:smash:

I am sure Guido Tent will come up one day with a good solution for the 963 clocking.🙄

Regards,

Lucas


P.S. Eganz, thanks for explaining the position of the jumpers for feeding the extra PSU into. I will do that probably next week.
 
#120
Lucas_G said:


Personally I cannot judge who was right here. Maybe guys like Guido Tent and others who know more about digital clocking can explain who is right: Tze Guan or Alex Peychev?:smash:

I am sure Guido Tent will come up one day with a good solution for the 963 clocking.🙄
Click to expand...

Hi

What we develop is based on the full documentation and measurements

Once our prodcution is run I can describe the architecture

Believe me, the way I intend to do it is to inject the cleanest clocks at the points where they matter: At the DAC chips. This is not making the specific sollution easy, but it is much better than " just" substituting some clocks and leave the rest of buffers, dividers and NAND gates in the clock signals......

regards
 
You must log in or register to reply here.
Top Bottom