OK so I modded my CD723 - but it's still mediocre

[Nuuk]

Nuuk - Yes, thanks, I had already seen your mods. I would prefer to keep it as a front loader - fits in my rack better.

I know what you mean Dave. Having just added a DAC, I have virtually nowhere to put it.

I would suggest that with mechanical mods to a CD player or DAC that empirical methods are probably as good as any (providing you have the time to experiment).

Did you ever read about the 'string suspension' that was popular in Germany for a while?

 

[ThorstenL]

Konnichiwa,

I'm really looking for some scientific explanation of why it should or should not be mass loaded / damped / rigid / sprung etc.

To keep it short and sweet. The lase system will "misread" more easily if the drive is exposed to vibrations. Many components are microphonic in varying frequency ranges and to varying degrees.

One can deal with vibrations and resonances in many ways. I personally have my favourite approaches which center around lossy, rigid natural materials (eg. Wood) combined with suspension methodes around string/spring etc.

Clear enough?

Sayonara

 

[Bricolo]

when you mention wood, do you mean as a support, or to modify the case with wood?

 

[ThorstenL]

Konnichiwa,

when you mention wood, do you mean as a support, or to modify the case with wood?

Both actually. I don't want to (or have time to) trot out all my philosophy and experience on the subject, however as a rule, conventional damping or excessive massloading does not to me seem to result in all that good sound.

Keeping things fairly light, lossy and rigid seems to help. Suspending vibration sensitive items helps too. If resonances cannot be avoided those from wood "sound" less offensive (and hence better) than those from metal or plastic.

Sayonara

 

[Bernhard]

The secret of perfect CD sound:

Upgrade to a transport that has been assembled by elfs.

 

[Fin]

More on the power supplies.

Hi Thorsten

Please see Post #96 of this thread, if you have time.

I have tried to digest your power supply recommendations and adapt them to my application, and the parts I already have.
It would be great if you could have a quick look at my understanding of how to implement them.
I am unsure about the arrangement of the components and what to use as a final regulator.

1: Analogue output stage:- one dual opamp.
Use one 2*22Vac 25VA PCB torroidal TF. Use Schottky diodes to build recifiier bridge. This produces +/- 31Vdc. Pass this through 3R3 - 470uF - LM317/337 pre-reg - 10R - 4700uF - ????? final reg???

Use capacitors like Rubycon ZA/ZL throughout this supply.
Would Panasonic FC or Rubycon YXF be OK here for the higher values?
Use Audio Grade capacitors (Elna Silmic or Starget) at the output of the reg ie. the 4700uF cap??
Is Nichicon Muse KS also good for this?


2: Analogue side of the DAC chip:- four +5V supplies.
Use one winding of the 2*15Vac 25VA PCB torroidal TF. Use Schottky diodes to build recifiier bridge. This produces +21Vdc. Pass this through 3R3 - 470uF - LM317 pre-reg -10R - 4700uF - ????? final regs (maybe 4 regs)???

Same type of caps as above??


3: Digital side of the DAC(including filter) and Decoder:- three +5V supplies.
Use second winding of the 2*15Vac 25VA PCB torroidal TF. Use Schottky diodes to build recifiier bridge. This produces +21Vdc. Pass this through 3R3 - 470uF - LM317 pre-reg - 10R - 4700uF - ????? final regs (maybe 3 regs)???

Same type of caps as above?

4: Supply for the Servos: +/-10V
Use 2*12Vac 30VA PCB torroidal TF. Use Schottky diodes to build recifiier bridge. This produces +/- 17Vdc. Pass this through 3R3 - 6800uF - LM317/337 Reg 10R - 6800uF - to provide one +10V supply and one -10V supply. This is connected to the main pcb traces feeding the servo section.

Would Rubycon YXF be OK here as they have this value?
Does each servo need a separate regulator/supply?


5: Supply for Master Clock:- Tent XO2 clock module.
Using dedicated Tent XO Supply.


6: Supply to the rest of CDP:- Display, uP, remote control...
Use original internal TF. Change existing capacitors in the original +/-5V/10V supplies from 6800uF and 1000uF to 470uf caps. Increase the size of the caps after the regs???

Have I understood any of what you have said or have I got it totally wrong?

In your article on modifying the CD63/67 you built the following:-

Say two 330 Watt Toroidal Transformers (one 2 X 18 Volt AC and one 2 X 12 Volt AC) with 12,000uF (12 X 1,000uF Nichicon PL or panasonic FA Series) Filtering per Voltage and some Chokes and polypropylene Cap's thrown in for good measure with LM317/337 Pre-Regulators.
This would give us two separate "raw" 12Volt Voltages for the two 5Volt Supplies and the "raw" +/- 18 Volt Voltages for the +/- 12 Volt Supply. In this case the +/- 12 Volt regulators should be really replaced with Linear Technologies LT1033/1085.

Is the move away from the 12,000uF filtering mainly due to space inside the CDP?

Thanks again for all your help.

 

[ThorstenL]

Re: More on the power supplies.

Konnichiwa,

1: Analogue output stage:- one dual opamp.
Use one 2*22Vac 25VA PCB torroidal TF. Use Schottky diodes to build recifiier bridge. This produces +/- 31Vdc. Pass this through 3R3 - 470uF - LM317/337 pre-reg - 10R - 4700uF - ????? final reg???

Nope, I'd go rectifier - 3R3 470uF 10R 4,700uF - Reg

Would Panasonic FC or Rubycon YXF be OK here for the higher values?

Sure.

Use Audio Grade capacitors (Elna Silmic or Starget) at the output of the reg ie. the 4700uF cap??

Large values on the output of LM317/1085 do not bring that much improvement, try 100uF each on adj to negative and on the output.

Is Nichicon Muse KS also good for this?

Maybe, never really worked with it. I like Elna Silmic best among "Audio" electrolytics and they are neither expensive nore hard to get.

2: Analogue side of the DAC chip:- four +5V supplies.
Use one winding of the 2*15Vac 25VA PCB torroidal TF. Use Schottky diodes to build recifiier bridge. This produces +21Vdc. Pass this through 3R3 - 470uF - LM317 pre-reg -10R - 4700uF - ????? final regs (maybe 4 regs)???

Same as above. Stick to RCRC filtering and one Reg. Also, not so much voltage. Keep around 3 - 5V headroom, no more. So no more than 10V DC for the 5V supplies. All else equally applies.

3: Digital side of the DAC(including filter) and Decoder:- three +5V supplies.
Use second winding of the 2*15Vac 25VA PCB torroidal TF. Use Schottky diodes to build recifiier bridge. This produces +21Vdc. Pass this through 3R3 - 470uF - LM317 pre-reg - 10R - 4700uF - ????? final regs (maybe 3 regs)???

See two previous posts, use Sanyo Os-Con in the Local decoupling and around the regulators.

4: Supply for the Servos: +/-10V
Use 2*12Vac 30VA PCB torroidal TF. Use Schottky diodes to build recifiier bridge. This produces +/- 17Vdc. Pass this through 3R3 - 6800uF - LM317/337 Reg 10R - 6800uF - to provide one +10V supply and one -10V supply. This is connected to the main pcb traces feeding the servo section.

I would omit any resistors, just rectifier and maybe 0.33R, then a big Cap (10,000uF ballpark), 1A (or better 3A) fixed voltage regulators and a large value cap on the output of the regulator.

Would Rubycon YXF be OK here as they have this value?

Sure.

Does each servo need a separate regulator/supply?

Never tried it, I think it's mostly important to have a high current supply separated from the rest of the circuitry.

6: Supply to the rest of CDP:- Display, uP, remote control...
Use original internal TF. Change existing capacitors in the original +/-5V/10V supplies from 6800uF and 1000uF to 470uf caps. Increase the size of the caps after the regs???

Leave the original supply as is, large cap's only increase peak currents and reduce conduatcion angle and to only drive "administrative" functions of a CDP there is no benefit.

Is the move away from the 12,000uF filtering mainly due to space inside the CDP?

No, change in Philosophy.... ;-)

Sayonara

 

[Fin]

More on the power supplies.

Thanks again Thorsten

Originally posted by Kuei Yang Wang Konnichiwa,
Nope, I'd go rectifier - 3R3 470uF 10R 4,700uF - Reg

Large values on the output of LM317/1085 do not bring that much improvement, try 100uF each on adj to negative and on the output.

I like Elna Silmic best among "Audio" electrolytics and they are neither expensive nore hard to get.

OK - For the analogue stage (opamp) I'll use:
3R3(Carbon Film) - 470uF (Pana FC) - 10R(Carbon Film) - 4,700uF(Pana FC) - LM317/337 - 100uF(Silmic*2). Should I forget about the Pre-Reg/Final Reg arrangement?
LM6172/OPA2604/OPA2134 - Decoupled with 100uF(Silmic) - 100uF DC Blocking/Coupling(Nichicon Muse ES Bi-Polar). Are Silmics good for DC blocking as well?
Any other comments?

Originally posted by Kuei Yang Wang
Same as above. Stick to RCRC filtering and one Reg. Also, not so much voltage. Keep around 3 - 5V headroom, no more. So no more than 10V DC for the 5V supplies. All else equally applies.

OK - For the analogue side of the dac I'll use:
Same RCRC as above and one Reg. Does "one Reg" mean:- one reg for all four analogue supplies to the dac; or:- forget about Pre-Regs and still consider one reg per suppliy pin on the dac?
The reason I went for the higher voltage tf is to allow for Pre Regs - but I gather you don't like that idea???
Should I use something like Elna Silmics for decoupling on the analogue side of the dac?

Originally posted by Kuei Yang Wang
See two previous posts, use Sanyo Os-Con in the Local decoupling and around the regulators.

OK - For the digital side of the DAC and Decoder I'll use:
Same as above with a separate Reg for the DAC and a another for the Decoder.
Is it definitely OK to group these together on the same winding?
What value Os-Cons would you suggest? There are 33uF electrolytics with 100nf smd ceramics in place originally. Should the 100nF's be removed? Would Rubycon ZA with the ceramics left in place be nearly as good as the Os-Cons?

Originally posted by Kuei Yang Wang
I would omit any resistors, just rectifier and maybe 0.33R, then a big Cap (10,000uF ballpark), 1A (or better 3A) fixed voltage regulators and a large value cap on the output of the regulator.

OK - For the supply to the Servos I'll use:
3A Schottky Diodes - 0.33R - 8,200uF/16V(Pana FC - the biggest they have) - 3A Reg - 8,200uF.
Would it be better to use 2*4,700uF or 5*2,200uF instead of 1*8,200uF?
Should the Regs set the voltage at +/-12V or +/-10V?

Will any of the servos or laser current need any adjustment after this?

Originally posted by Kuei Yang Wang
Leave the original supply as is, large cap's only increase peak currents and reduce conduatcion angle and to only drive "administrative" functions of a CDP there is no benefit.

So - they only need something like 470uF??

Originally posted by Kuei Yang Wang
No, change in Philosophy.... ;-)

Change in Philosophy away from banks of capacitors or large capacitance in analogue supplies or both?

 

[ThorstenL]

Re: More on the power supplies.

Konnichiwa,

Should I forget about the Pre-Reg/Final Reg arrangement?

To me a pre/post reg arrangement only makes if the post reg is optimised for low and constant with bandwidth output impedance with the pre reg providing the bulk of ripple rejection. Just cascading multple 317's does not seem to give much advantages.

LM6172/OPA2604/OPA2134

I do not recommend the LM6172 in this application. USe LM6182 or AD812.

Decoupled with 100uF(Silmic)

Keep the original decoupling arrangement and add local at the Chip one Cap between +V & -V, be it a 1uF Epcos stacked film or a Kemet 0.1uF "Gold" Ceramic.

100uF DC Blocking/Coupling(Nichicon Muse ES Bi-Polar). Are Silmics good for DC blocking as well?

DC blocking? What for? You can cancel the offset and direct couple.

Is it definitely OK to group these together on the same winding?

In what sense? Will it be no worde than using seperate windings? I feel seperate windings are ALWAYS better.

What value Os-Cons would you suggest?

Whatever fits PCB and Budget, I usually keep bags of 22uF/6.3V SH Series around for Digital Mods.

Should the 100nF's be removed?

Yes, but replace by local SMD Capacitors 10nF/0805 X7R and 100nF/0805 Z5U/Y5U.

Would Rubycon ZA with the ceramics left in place be nearly as good as the Os-Cons?

Nope.

Would it be better to use 2*4,700uF or 5*2,200uF instead of 1*8,200uF?

Usually No.

Should the Regs set the voltage at +/-12V or +/-10V?

Measure whatever is in place originally and replicate, or check in the datasheets what the nominal supply is for the Chips used in the player.

Will any of the servos or laser current need any adjustment after this?

In modern players - no.

So - they only need something like 470uF??

Probably. Why not check the Datasheets, total up worst case supply current and use Duncanamps Powersupply Designer to see how low you can go?

Change in Philosophy away from banks of capacitors or large capacitance in analogue supplies or both?

I never used banks of capacitors nor do I feel they have merit, other than to lower costs in commercial production.

Sayonara

 

[Bricolo]

"in modern players - no"

and in my CD-62? (CDM4)




One thing: some people like to parallel caps (the 100uF+smd ceramic is seen quite often, the triplet is also a well spread technique), but Guido Tent's article abous PS decoupling nearly convinced me that it's not a good idea.
Thorsen, all your help here is based on a good technical explanation (either here or in your articles). Could you tell us why the use of the paralleled caps? Aren't the oscons effective enough at high frequencies?

 

[ThorstenL]

Konnichiwa,

"in modern players - no"

and in my CD-62? (CDM4)

Maybe, I would think not if you match the original voltages.

One thing: some people like to parallel caps (the 100uF+smd ceramic is seen quite often, the triplet is also a well spread technique),

For very good reasons too.

but Guido Tent's article abous PS decoupling nearly convinced me that it's not a good idea.

How so? I would not think that Guido is arguing against using local bypass capacitors of suitable nature, which invariably means multiple value capacitors in parallel FOR EACH SUPPLY NODE.

Could you tell us why the use of the paralleled caps?

In order to retain a low impedance PSU nore for each and every given PSU into the high 2 digit MHz range.

Aren't the oscons effective enough at high frequencies?

Well, normal 'lytics crap out usually around 100KHz, Os-Cons have a lower ESR and remain capaictive to 300...500KHz, neither one is sufficient for the kind of switching frequencies for the logic in modern gear. So further local bypassing is essential, plus Os-Con's are physically quite large, meaning they will remain placed at a considerable distance from the actual PSU node, adding a lot of leadlength and PCB inductance. Thus the combination of multiple capacitors with different ESL and C to spread and dampen any LC tank circuit resonances and to keep the PSU node impedance low.

A reasonable overview of the topic is found here:

http://fibbla.ce.chalmers.se/undergraduate/E/EDA232/2002/datablad/elektorpaper.pdf

Sayonara

 

[Peter Daniel]

One thing: some people like to parallel caps (the 100uF+smd ceramic is seen quite often, the triplet is also a well spread technique), but Guido Tent's article abous PS decoupling nearly convinced me that it's not a good idea.
Thorsen, all your help here is based on a good technical explanation (either here or in your articles). Could you tell us why the use of the paralleled caps? Aren't the oscons effective enough at high frequencies?

I was almost convinced about that too, but nevertherless, I tried adding ceramic bypasses and the sound improved, while using BG N type caps (10u and 100u). Some other people also observed improvement with adding ceramics SMD caps (not all of them are the same, a proper choice is important)

 

[Bricolo]

Something like this?

 

[Fin]

More on the power supplies.

Hi Thorsten

I'm sure your are getting tired of our non-stop questions!

Unfortunately, I'm not an electronics student and I'm struggling to keep up with the rest of you.

Originally posted by Kuei Yang Wang
To me a pre/post reg arrangement only makes if the post reg is optimised for low and constant with bandwidth output impedance with the pre reg providing the bulk of ripple rejection. Just cascading multple 317's does not seem to give much advantages.

Would it be worthwhile using the pre/post reg arrangement that you used in the CD63/67 TNT article - quoted below?

Taken fron the TNT CD63/67 article.
Say two 330 Watt Toroidal Transformers (one 2 X 18 Volt AC and one 2 X 12 Volt AC) with 12,000uF (12 X 1,000uF Nichicon PL or panasonic FA Series) Filtering per Voltage and some Chokes and polypropylene Cap's thrown in for good measure with LM317/337 Pre-Regulators. This would give us two separate "raw" 12Volt Voltages for the two 5Volt Supplies and the "raw" +/- 18 Volt Voltages for the +/- 12 Volt Supply. In this case the +/- 12 Volt regulators should be really replaced with Linear Technologies LT1033/1085.

Originally posted by Kuei Yang Wang
I never used banks of capacitors nor do I feel they have merit, other than to lower costs in commercial production.

Maybe my use of the term; "banks of capacitors", was incorrect.

Originally posted by Kuei Yang Wang
Keep the original decoupling arrangement and add local at the Chip one Cap between +V & -V, be it a 1uF Epcos stacked film or a Kemet 0.1uF "Gold" Ceramic.

For us "Dummies", this is between pins 4 and 8 on a dual opamp - and not from supply to ground?

Originally posted by Kuei Yang Wang
DC blocking? What for? You can cancel the offset and direct couple.

Maybe I have used the wrong term here again! The original arrangement in the CD624 has a 100uF/16V Nichicon Muse Bi-Polar electrolytic directly in each output path from pins 1 and 7 from the NJM5532 opamp.
I need to find out more about cancelling offset.

Originally posted by Kuei Yang Wang
In what sense? Will it be no worde than using seperate windings? I feel seperate windings are ALWAYS better.

I suppose what I am trying to ask here is: Which of the following two arrangements would be better, if I want to power the SAA7310 Decoder IC and the SAA7321 Bitstream DAC/Filter IC from one dual secondary transformer?

Option 1. The four +5V supplies of the analogue side of the dac powered from one winding, and;
The two +5V supplies for the digital side of the dac and the +5V supply for the decoder powered from the second winding.

OR

Option 2. The four +5V supplies of the analogue side of the dac and the two +5V supplies for the digital side of the dac powered fron one winding, and;
The +5V supply for the decoder powered fron the second winding.

Originally posted by Kuei Yang Wang
Whatever fits PCB and Budget, I usually keep bags of 22uF/6.3V SH Series around for Digital Mods.

You obviously feel that the SG (audio) Series is not necessary here?

Originally posted by Kuei Yang Wang
Yes, but replace by local SMD Capacitors 10nF/0805 X7R and 100nF/0805 Z5U/Y5U.

So - for decoupling around the DAC - I should remove the existing 33uF/16V electrolytics and the 100 nF SMD's. These should be replaced with Os-Cons (around 22uF - whatever fits neatly) and 10nF/0805 X7R and 100nF/0805 Z5U/Y5U??

Originally posted by Kuei Yang Wang
Measure whatever is in place originally and replicate, or check in the datasheets what the nominal supply is for the Chips used in the player.

Is there any point in changing any of the opamps (NJM4560) in the servos?

Originally posted by Kuei Yang Wang
In modern players - no.

If the CD624 falls into this category - then I'm happy as I don't fancy messing with this.

Originally posted by Kuei Yang Wang
Probably. Why not check the Datasheets, total up worst case supply current and use Duncanamps Powersupply Designer to see how low you can go?

I will definitely be looking at this!

 

[Fin]

Optimum Decoupling

One thing: some people like to parallel caps (the 100uF+smd ceramic is seen quite often, the triplet is also a well spread technique), but Guido Tent's article abous PS decoupling nearly convinced me that it's not a good idea.
Thorsen, all your help here is based on a good technical explanation (either here or in your articles). Could you tell us why the use of the paralleled caps? Aren't the oscons effective enough at high frequencies?

I tried to address this in another thread.

Optimum Decoupling in Digital IC's

 

[ThorstenL]

Konnichiwa,

Something like this?

Not really.

First, I would always use pre-regulators (317/1085) with LM/TL431 Shunt Regs.

Secondly, the shuntreg and bypass capacitors for each IC should combine at a local star, the ground pin of the IC.

One final "refinement", R1/R2 and R3/R4 can be split and distributed across the two supply lines, preferablyt using inductive wirewound resistors with added ferrite beads.

Sayonara

 

[Bricolo]

Konnichiwa,



Not really.

First, I would always use pre-regulators (317/1085) with LM/TL431 Shunt Regs.

Secondly, the shuntreg and bypass capacitors for each IC should combine at a local star, the ground pin of the IC.

One final "refinement", R1/R2 and R3/R4 can be split and distributed across the two supply lines, preferablyt using inductive wirewound resistors with added ferrite beads.

Sayonara

one pre reg per shunt, or one per secondary?

do you mean, the shunt+cap's ground should be placed at the chip's ground? so no direct connection to the bridge?

I don't understand the last one

 

[ThorstenL]

Konnichiwa,

one pre reg per shunt, or one per secondary?

One per secondary.

do you mean, the shunt+cap's ground should be placed at the chip's ground? so no direct connection to the bridge?

I don't understand the last one

Remember my mantra - look where current will flow....

You have your IC's, their grounds are joined via a Bus/Groundplane System. You have several pre-regulated, "clean" supplies you wish to feed to various Supply Nodes, in most cases individual per IC, meaning the negative supply to ground pin and the positive to the supply pin via a resistor/ccs, assuming the shunt regulator is locat to the Chip, where it belongs (otherwise it's > 100KHz bandwidth is wasted).

If you have several IC's to supply look at the one that pump's the most junk into the supply/ground loop and/or which requires the cleanest "ground" and attach the negative wire to the relevant chip's (hard to say which is more relevant) ground pin, then fan out individual resistors/ccs per chip to the positive supply pin's.

Allways thinbk of the circuit (chip) as posessing one positive and one negative power pin and as having one signal ground and several I/O Pin's. Sadly for us the signal ground and negative pin have usually been combined, be it in logic circuits or in Op-Amp's and single ended audio circuits.

However, you still CAN separate the supply current and signal current loops by having seperate supplies (galvanically seperate) per chip.

Clearer?

Sayonara

 

[ThorstenL]

Re: More on the power supplies.

Konnichiwa,

Would it be worthwhile using the pre/post reg arrangement that you used in the CD63/67 TNT article - quoted below?

Maybe, probably not. In those days I failed to understand what really went on in supplies and hence followed the (arguably pretty reliably applicable) "bigger transformers, capacitors and more regs are good" credo. I have since learned that once you understand the goings on you can get "more with less".

Maybe my use of the term; "banks of capacitors", was incorrect.

I understand it as using several smaller value capacitors in parallel instead of using one suitable value part of sufficient quality.

For us "Dummies", this is between pins 4 and 8 on a dual opamp - and not from supply to ground?

You. Just try to think where the Op-Amp's ground pin is. Can't see it? It usually is the negative supply pin, but sometimes the positive supply pin. Only this current loop is essential for stability at high frequencies.

Maybe I have used the wrong term here again! The original arrangement in the CD624 has a 100uF/16V Nichicon Muse Bi-Polar electrolytic directly in each output path from pins 1 and 7 from the NJM5532 opamp.
I need to find out more about cancelling offset.

It implies to introduce a direct current/voltage into the circuit in such a mannner that at the Audio Stages output the Output is held reliably at or close (enough) to 0V.

I suppose what I am trying to ask here is: Which of the following two arrangements would be better, if I want to power the SAA7310 Decoder IC and the SAA7321 Bitstream DAC/Filter IC from one dual secondary transformer?

One supply for DAC digital, one for Decoder. Tap of the 5V supplies for the Analogue side of the DAC from the +/-15V supply of the Output Stage, using seperate regs directly of the raw rectified supply.

To be absolutely certain what should be consideredn analog and what digital and what pin's are noise sensitive one would have to go in detail through datasheet and apps notes, which I cannot find on the quick.

You obviously feel that the SG (audio) Series is not necessary here?

Or apropriate.

So - for decoupling around the DAC - I should remove the existing 33uF/16V electrolytics and the 100 nF SMD's. These should be replaced with Os-Cons (around 22uF - whatever fits neatly) and 10nF/0805 X7R and 100nF/0805 Z5U/Y5U??

If SMD's are in place, you may find they are okay to leave and to add a 10nF SMD directly at the chip Pin's (as close as you can get). Replace the existing 'lytics by Os-Con's.

Is there any point in changing any of the opamps (NJM4560) in the servos?

Maybe, but doubtful. One would have to see the full schematic and details on the other integrated circuits to be certain.

If the CD624 falls into this category - then I'm happy as I don't fancy messing with this.

I think yous has the old swingarm mechanism, a very nice drive but rather aged and yes, these do tend to require adjustment eventually, as not yet all adjustments where on-chip dynamically.

BTW, a final warning. What you are trying to do should IMHO really only be done by people with both substantial theoretical grounding in electronics and substantial practical experience. If you are in doubt about something - don't do it. Any of my comments and suggesting are basically aimed at qualified and experienced EE's, as I cannot see how people without this can truely understand what they are doing. If this limits the extend of your modifications, tough luck.

Sayonara

 

[Fin]

More on the power supplies.

Thanks again, Thorsten, for the good advice.

Originally posted by Kuei Yang Wang
You. Just try to think where the Op-Amp's ground pin is. Can't see it? It usually is the negative supply pin, but sometimes the positive supply pin. Only this current loop is essential for stability at high frequencies.

The "Us Dummies" comment was only referring to myself and anyone else reading this thread who might find themselves on the same learning curve.

Also - I never said the op-amp had a ground pin - but thanks for the clarification.

Originally posted by Kuei Yang Wang
One supply for DAC digital, one for Decoder. Tap of the 5V supplies for the Analogue side of the DAC from the +/-15V supply of the Output Stage, using seperate regs directly of the raw rectified supply.

To be absolutely certain what should be consideredn analog and what digital and what pin's are noise sensitive one would have to go in detail through datasheet and apps notes, which I cannot find on the quick.

I can send them to you if you like............................

Originally posted by Kuei Yang Wang
Or apropriate.

So - where would one use Os-Con SG series?

Originally posted by Kuei Yang Wang
If SMD's are in place, you may find they are okay to leave and to add a 10nF SMD directly at the chip Pin's (as close as you can get). Replace the existing 'lytics by Os-Con's.

I was hoping you would say that the existing SMD might be OK!

If the Vdd and Vss pins are not next to each other - but are on adjacent sides of the IC - what do you think of using an axial 10nF ceramic directly across the pins?

Originally posted by Kuei Yang Wang
Maybe, but doubtful. One would have to see the full schematic and details on the other integrated circuits to be certain.

Again, I can send them to you if you like............................

Originally posted by Kuei Yang Wang
I think yous has the old swingarm mechanism, a very nice drive but rather aged and yes, these do tend to require adjustment eventually, as not yet all adjustments where on-chip dynamically.

Yes - it is a CDM4/19 - in mint condition - hardly used.

Originally posted by Kuei Yang Wang
BTW, a final warning. What you are trying to do should IMHO really only be done by people with both substantial theoretical grounding in electronics and substantial practical experience. If you are in doubt about something - don't do it. Any of my comments and suggesting are basically aimed at qualified and experienced EE's, as I cannot see how people without this can truely understand what they are doing. If this limits the extend of your modifications, tough luck.

Thanks for the warning - I have taken note of it. Yes - it is my tough luck that I studied Chemisry instead of Electronics!