Marantz CD63 & CD67 mods list

Hi,

Basically, you are bypassing a large section of the analog low-pass filter with this mod. That means lots of high-frequency noise from the DAC can enter your system. This is not a good thing, but the effect of it may vary from system to system. Some have reported a better sound because the extra opamp stage is bypassed, but it can also get worse and cause more distortion. So YMMV... and watch your tweeters!

Ray
@Ray: You are most probably right but since I will use +/- 15V low noise voltage regulators for the analog rail and the low jitter clocks are having their low noise regulators I think noise might not be a big issue here even if the LPF is taken away. Someone has also told me that the SQ would be more neutral and non colored if the LPF is removed. I will test it out and see if that was the case.

BTW I have assembled the two Fleas from you and this will be used in my 63SE and the CDP will be used in the bed room for my headphone amp. I will start the new mod project after Xmas.

One thing I want to seek your view, Ray. In my modded CD63 KI since I have removed everything after the opamps and when I switch off the CDP there is always cracking noise feeding to the amplifier. Everytime I need to switch off the amp. first then turn off the CDP to avoid this disturbing noise. Do you have any simple cure for this and not affecting the SQ ?
 
Last edited:
All those low-noise regulators will not help against the high-frequency square wave that is output by the DAC itself... it's the nature of the beast, so to speak.

The analog filter is needed to create a sinewave shaped signal out of this pulse-width modulated output, because it's nowhere near analog. If you bypass the filter, all kinds of higher-order harmonics will enter the system and this will create all sorts of distortion and intermodulation products. Not good...
But, depending on the system, you may or may not hear the negative side-effects of this.

The solution for the switch-off noise is a muting relay. You can find more info on this on my CD57-XXL page.

Good luck with The Flea's :)

Ray
 

Attachments

  • PWM.jpg
    PWM.jpg
    66.8 KB · Views: 215
Thanks Ray for the valuable advise and I will try to construct the muting circuit when I have the time.

In my KI mod I tried to feed the 16 MHz clock signal directly from the clock to the Decoder via C521. Unfortunately the CD was unable to read any tracks and started to spin by itself. I have no idea what has gone wrong :(. I used the same clock output to both DAC and Decoder (in parallel). Even though I added a 100ohm resistor to the cable connected to the Decoder the result was the same. Do you have any advise on this please Ray ?
 
It should work with a small capacitor in series. IIRC, the decoder's clock-input has some DC offset that may screw things up. Did you disconnect the PCB trace between the DAC's clock output and decoder properly?
Thanks Ray. BTW what is IIRC ? What I did was to connect the clock cable to one leg of C521 on top of the PCB and cut its leg on the side connecting the DAC. See the uploaded diagram.
 

Attachments

  • CD63KI_doc.pdf
    143.6 KB · Views: 54
I just fixed my buzzing problem by moving the centre tap of the output stage transformer from U309 to here:

Capture.jpg


I can't rule out the possibility that it was just a bad connection before. Thanks for spurring me to take another look, higlander - that's been bugging me for a while.
Hi, it is really nice to hear that eventually yo have solved your long lasting problem. I am sure your CD63 is now very quiet. Actually I am now enjoying my KI while typing :p
 
Thanks Ray. BTW what is IIRC ? What I did was to connect the clock cable to one leg of C521 on top of the PCB and cut its leg on the side connecting the DAC. See the uploaded diagram.

:) No, it has nothing to do with caps :D

Looks o.k. to me, it should work this way... Did you also connect a GND wire from the Flea to the decoder?

Maybe C522 and 523 should be removed. It's the last thing I can think of. Especially that 10nF is quite a capacitive load.
 
Last edited:
Looks o.k. to me, it should work this way... Did you also connect a GND wire from the Flea to the decoder?

Maybe C522 and 523 should be removed. It's the last thing I can think of. Especially that 10nF is quite a capacitive load.
I connected the clock as in the diagram but it did not work :(

The ground wire of the clock was connected and the DAC worked nicely but only not on the Decoder :(

May be you are right I will try to remove C522 C523 as well as R521 (I think I need to remove R521 otherwise CR OUT will have clock signal via R521). With this arrangement Pin 14 CR OUT will be totally isolated i.e. free from any connection and the clock signal goes directly to Pin 13 CR IN. I hope it will work.
 
Hi,
i'm using the 63 as a transport (internal is dac disconnected) and i ran into some trouble clocking the decoder (alone) as well.
What i found is the following:
-removing all parts (c521,c522,c523, etc) and feed the clock signal straight in did NOT work. CRIN obviously needs a series resistor to the clock signal, i took what i had at hand, 100R. 390R seemed a bit high to me.
-removing R521 also caused the player to fail, the decoder seems to need a connection between CRIN and CROUT, i took 500kohm

What worked when i clocked the dac and the decoder from the same clock, was indeed a series-capacitor of 10pF.

Mickie
 
Hi,
i'm using the 63 as a transport (internal is dac disconnected) and i ran into some trouble clocking the decoder (alone) as well.
What i found is the following:
-removing all parts (c521,c522,c523, etc) and feed the clock signal straight in did NOT work. CRIN obviously needs a series resistor to the clock signal, i took what i had at hand, 100R. 390R seemed a bit high to me.
-removing R521 also caused the player to fail, the decoder seems to need a connection between CRIN and CROUT, i took 500kohm

What worked when i clocked the dac and the decoder from the same clock, was indeed a series-capacitor of 10pF.

Mickie
@Mickie: Am my understand is correct that I should remove everything connected to CR IN and CR OUT and then connect the clock signal to CR IN using a 10p cap in series ??
 
..sorry, i was a bit unclear. I think you can / should leave everything in place, just lift the 390R (clock out from the dac is now disconnected) solder a 10pF the the clock cable and connect it to the 390R resistor. This is basically you original configuration but you have bypassed the dac and therfore have cleaner signal. Anyway this should work. You can experiment with removing the other small caps later.
Mickie
 
..sorry, i was a bit unclear. I think you can / should leave everything in place, just lift the 390R (clock out from the dac is now disconnected) solder a 10pF the the clock cable and connect it to the 390R resistor. This is basically you original configuration but you have bypassed the dac and therfore have cleaner signal. Anyway this should work. You can experiment with removing the other small caps later.
Mickie
Thanks Mickie. Is this what you mean as shown in the diagram ?
I connect the clock as in the diagram but it did not work.
 

Attachments

  • CD63KI_doc.pdf
    143.6 KB · Views: 64
Right.
Unfortunately i don't know why the clock-signal can't be fed straight into CRIN by removing all other parts (C521,522,523, R521, U193) -as you can do with the dac- and what those caps/resistors actually do.
Maybe the SAA7345 needs a bufferd clock-signal when its synchronously reclocked.

You may also do some considerations about the connection-cables you use; while a shielded coax for the dac is highly recommended, it adds some capacitance to your 10pF cap (i used a silver mica, btw.) which is not wanted.
Anyway, reclocking the decoder seems tricky.

Maybe someone else out there can offer a more elegant solution?

EDIT: I'm sorry i skipped your comment that its not working for you this way - it definitely has to, thats strange.....hmmmmm.
What i can think of, is the fact that there are issues with the ribbon cable (from mainbord to drive pcb). Connections can fail when the cable is often removed/reconnected. Symptoms can be very similar to a bad clock-connection.....so please check

Regards,
Mickie
 
Last edited:
Clocking Decoder

Thanks Mickie for your valuable comments and as you said theoretically speaking feeding a clock signal via the 10p cap should work but unfortunately it did not. Anyway I will try other ways and connections and see if I can resolve the problem.
BTW when you have the independent clock signal feeding to the decoder what are the improvements you observed ? Do you think it is worthy to carry out this mod ?
 
Do you think it is worthy to carry out this mod ?

..well, not if you spend days resolving issues.

I did it out of curiosity and because i felt uncomfortable looping the low jitter clock-signal through the dac...

It's audible, yes, but as you can't feed the signal - clean, with a shielded cable - straight forward to the decoder, and, in addition to that, have to consider cable capacitances, HF-noise beeing emitted and picked up, etc., benefits of this mod are easily "reversed" i'm afraid.

I don't know which clock you are using, but there are some that offer a buffered (also divided by 2 or 4) frequency output - that might be a different story.

On the CD67 for example, a "parallel" reclocking of dac and decoder from the same clock is not a problem, but it uses a different decoder-chip.

I think at this point its better to put everything back together as it was and reclock the dac only (for now).

Reclocking the servo-chip / transport is probably more effective.

Regards,
Mickie

PS: Everbody please feel free to correct or complete my statements.
 
Last edited:
..well, not if you spend days resolving issues.

I did it out of curiosity and because i felt uncomfortable looping the low jitter clock-signal through the dac...

It's audible, yes, but as you can't feed the signal - clean, with a shielded cable - straight forward to the decoder, and, in addition to that, have to consider cable capacitances, HF-noise beeing emitted and picked up, etc., benefits of this mod are easily "reversed" i'm afraid.

I don't know which clock you are using, but there are some that offer a buffered (also divided by 2 or 4) frequency output - that might be a different story.

On the CD67 for example, a "parallel" reclocking of dac and decoder from the same clock is not a problem, but it uses a different decoder-chip.

I think at this point its better to put everything back together as it was and reclock the dac only (for now).

Reclocking the servo-chip / transport is probably more effective.

Regards,
Mickie

PS: Everbody please feel free to correct or complete my statements.
Thanks Mickie again for the reply.
Actually I have used 2 independent clocks - one for the servo and the other for the DAC. Clocking the Decoder is the last mod I would like to try but if it does not work I will leave it. My clock to the DAC has no buffer and I can only tap the signal out in parallel with the DAC signal - that may have caused the problem :(