Marantz CD63 & CD67 mods list

Member
Joined 2006
Paid Member
Re: Re: Re: Marantz CD63 & CD67 mods list

peufeu said:


So what ? OK, it has a $0.2 VHCMOS Hex inverter. The description is pretty accurate.

As shown from the schematic, this inverter acts as a buffer for a crystal oscillator of completely unknown specifications. Such an oscillator can cost anywhere from $0.5 to $50 depending on manufacturer, and specs you might not care about (like accuracy), and specs you might care about (like phase noise and jitter). The power supply also looks pretty high-noise to me. LM317 is not known as a low-noise part. So, as far as I can tell from your data the performance of this clock is unknown.

As for regulators, try MIC5205 or LP3985, look at the datasheets, compare the noise to good old LM317, you will know what I mean. Also those regulators cost less than $1, so you can't go wrong ! You can use a LM317 as a pre-regulator, for instance LM317 to make +6V and MIC5205 to make +5V, you will get low thermal dissipation in the extremely tiny MIC5205 and lower noise at the end.

Tent sells what is basically a not very expensive XO at an inflated price because Tent spent lots of time testing lots of XOs to find the right one. Tent also had a deal with the manufacturer to get better specifications on his XO. FYI someone I know asked for something like that from a reputable manufacturer and they were quoted a $2500 design fee because "the frequency is not common" (this manufacturer makes low-jitter parts for telecom gear so they have no stock for audio frequencies).

Personally I find it completely worthwile to pay the price for the Tent oscillator since the time and gear necessary to sort the good one from the bad ones is worth much more that what Tent asks for his oscillator.

If you consider you need a $50K jitter analyzer and a few weeks of lab work to select the right part, well, you get the idea, $50 for a preselected part is a good deal. Also, I like mister Tent, so I have no problem helping him to make a living, lol.

Also, for your power supply, please note that if you feed the clock from the supply in your player, you will get a very nasty ground loop. Since the clock will travel in this ground loop it will radiate like crazy.

Using a separate low-power transformer (pretty cheap) and rectifier bridge like they do in this clock is a very good idea, I encourage you to do that. You should keep the wires from the clock source to the point it is used on the PCB as short as possible, and insert ferrite beads like VK200 in both transformer wires at the PCB to block the RF from the supply.

Also please note that CD63 uses a complete POS DAC so IMHO you'd better tweak a player with a real DAC inside.


rowemeister said:
I agree, the cost of buying audio freq xtals is quite high.

But why are some people on the 'DACs make all the difference' band wagon. Mostly new dacs are used in newer models of cdp due to marketing because it sells cd players. OK I understand there are different quality dacs, but i would rather have very clean power rails on a pants dac than the other way round

Hi Brent

So my experiments with Chinese unknown clocks ends here.

Maybe I will try one of Sercal clocks after I install the sreg on analog dac.

Can you comment the difference between the Sercal unit and the Tent XO + sreg ?

Regards

Ricardo
 
Hey, it could be good (well the power supply still sucks though) but it could as well suck, you got no guarantee. e-bay stuff from China is cheap but not known for the quality of customer service if you know what I mean, lol. If you're going to the trouble of gutting your CDP, adding this, plus an extra power transformer etc, you might as well get a good one for your trouble ;)
 
Hi Ricardo

The C1 clock consists of a ultra precise low noise 3.3V regulator to allow the oscillator to run with a noise free supply. Of course all the capacitors on the circuit board were designed to match both the regulator and oscillator perfectly. Also being smt the size is compact and the tracks are much shorter.

The Tent will be similar in many ways, the C1 will perform as well as if not better than the Audiocom superclock 2. I have not done an exact comparison with the Tent to the C1 so I would not like to say. But like the Tent many hours of testing various voltage regs and oscillators have gone into the C1.

Brent
 
peufeu said:
My bad, there is no IV, I spent too much time with those other DACs, lol.

I was referring to the fact of using an opamp with 13 MHz GBW to filter the modulator output which is basically spread spectrum noise up to tens of MHz. Well, stuff like THS4131 didn't exist at that time...

The player does sound poor with the standard op-amps, power supplies and clocking arrangements in place! Fix these 3 things and it is very impressive indeed.
 
Member
Joined 2006
Paid Member
rowemeister said:
Hi Ricardo

The C1 clock consists of a ultra precise low noise 3.3V regulator to allow the oscillator to run with a noise free supply. Of course all the capacitors on the circuit board were designed to match both the regulator and oscillator perfectly. Also being smt the size is compact and the tracks are much shorter.

The Tent will be similar in many ways, the C1 will perform as well as if not better than the Audiocom superclock 2. I have not done an exact comparison with the Tent to the C1 so I would not like to say. But like the Tent many hours of testing various voltage regs and oscillators have gone into the C1.

Thank you Brent

This was just the info I needed. ;)

Can the C1 be powered by just one Spower reg ? I will build the dedicated psu myself.

Regards

Ricardo
 
Member
Joined 2006
Paid Member
SimontY said:


Brent, on Ebay LOL.

He was clearing out the last of his Super Regs to make way for S Powers, as the S are his own product and better (higher current delivery and with better parts).
Brent !!!

Any leftovers ?

I love the Spower I recently got.... Wonderfull miniature... (In the pictures it looked much bigger)

Work of art indeed :wiz:

I did not install it yet because I want to build a dedicated psu for the dac analog and digital supply.

I must do everything at once because I am afraid the next time I disasemble the CD53, it will stop playing (dangerous ribbon :hot: ).

After soldering the new ribbon I got from a friend, I will not be able to operate on the board easily.

Ricardo
 
RCruz said:


After soldering the new ribbon I got from a friend, I will not be able to operate on the board easily.

Ricardo


That's not as bad as it seems, you can remove the entire drive mechanism along with the main-board fairly well.

ps - my dac appears to have stopped outputting any signal. Any ideas? It was working the last time I played it, all feeds into the discrete read 0v

Lee.
 
RCruz said:

Brent !!!

Any leftovers ?

I love the Spower I recently got.... Wonderfull miniature... (In the pictures it looked much bigger)

Work of art indeed :wiz:

I did not install it yet because I want to build a dedicated psu for the dac analog and digital supply.

I must do everything at once because I am afraid the next time I disasemble the CD53, it will stop playing (dangerous ribbon :hot: ).

After soldering the new ribbon I got from a friend, I will not be able to operate on the board easily.

Ricardo

Thanks Ricardo

Lots of care was taken when designing the reg, one of those smt caps is actually 10uF lol.

Brent
 
Caps on Super Rayregs

Hi guys.

C1 master clock now installed and very impressed, Brent.

Even just fed from U163 (5v from C815), the extra detail and timing is significant.

470u/16v BG Std from Thomo settling in nicely, 4 on opamps, 1 each on Decoder ana/dig and 2 on DAC dig at CD05 and CD06.

220u/16v BG FK on DAC ana at CD12, CD13.

Now to redo the regs.

I have been using the basic Rayregs and want to upgrade them to SuperRayregs by adding the VBE before the reg (still LM317 based)

I have quite a few 220u/16v Cerafines, and wondered if it would be ok to put these on the regs between output/ground.

Is it better to stick with the smaller 10u/16v tants, which I think may be faster?

Any thoughts on this?

Thanks for any advice.

Jim
 
If you put the caps on the regulator board and not at the chips where the power is used, the inductance and resistance of your wire will be much higher than the capacitor's own parasitics.

In that case Panasonic low ESR types (I think it's the FA-FK series) can be useful. However, remember that the capacitor on the regulator board, the wire inductance, and the capacitor at the chip where the power is used form a LC resonant circuit, so you actually want some ESR on your regulator board to damp the resonance. It's the same as snubbing a rectifier. Twist your wires.

Not a good idea to pay a premium for a 20 mOhm ESR cap and stick a few tens of nH of wire inductance after it, which will give you more than 100 mOhm impedance at 1 MHz... your cap is wasted. In that case a solid tantalum (the orange SMD ones, you saw them already I bet) is a better choice because it will have higher ESR which will dampen the resonance.

If your tantalums are "drops" type they probably suck at high frequencies. Cerafine is likely to be better. Tantalum drops also have high distortion.

If you want really high-performance bypass, at the chip, for small values 0805 SMD ceramic is good (X7R) but ESR is so low that it might make LC oscillations with trace inductance. Paralleling caps can be dangerous if you overdo it. Obviously adding 2 cm of trace to a ceramic capacitor makes it worse than useless since you get a LC circuit which will ring. Those little Murata BLM ferrite beads are your friend in digital circuits ;)

For larger values you can try the various "special polymer" like this :

http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail?name=PCE3605CT-ND

Or OSCON (a classic) or Panasonic low ESR types (I think it's the FA-FK series). Those are intended for switching supplies so in low voltages they're pretty cheap thanks to the computer inductry...

Also ceramic caps and crystal oscillators are microphonic so you should control the vibration of your PCB. A good method I have applied is to use anti-static foam (like the one that came in your computer's motherboard box) and stick it below the PCB, so it will support the PCB and prevent it from slapping around. You could use other stuff if you like.
 
Hi Peufeu

I've got 470u/16v BG's at the chips on the main pcb.

As I always solder the reg output pin into the fuse resistor hole (RD01, RD04, R508,R511, etc, the reg is as close as I can get to the chips' reservoir caps.

I also always twist the supply wires to the regs together, so that's good also.

The LM317 regs I'm using at the moment just have 16v/10u tantalum beads between output/adjust(ground) with metal film resistors to set the voltage.

The regs I'm wanting to use BC547,1k5 metalfilm and 16v/4.7u tant beads in the VBE before the reg to convert, and have seen people advise using something like the cerafines on the reg itself.

As I already have the Cerafines, it seems a shame to waste them?

Jim

PS thanks for the tip on the PCB. I have some sound deadening foam from a PC soundproofing kit I never used, so I might give that a try once I've finished all my mods (after I've checked it's non-conductive!!)