Marantz CD63 & CD67 mods list

[martin clark]

I would hesitate to comment on one of Guido's products, especially unseen/unheard, but they tend to be extremely well thought out and realised Here teh obvious things are low output impedance and very quiet, too: 20nV/SQRT(hz) at 5V out; for the sake of comparison, the best you can get out of a LM317 is about 330nV/SQRT(hz)...

I must have a play one day - those look like a very neat drop-in replacement for a three-pin reg if the current output is sufficient for the task at hand. Neat.

 

[SimontY]

I would hesitate to comment on one of Guido's products, especially unseen/unheard, but they tend to be extremely well thought out and realised Here teh obvious things are low output impedance and very quiet, too: 20nV/SQRT(hz) at 5V out; for the sake of comparison, the best you can get out of a LM317 is about 330nV/SQRT(hz)...

I must have a play one day - those look like a very neat drop-in replacement for a three-pin reg if the current output is sufficient for the task at hand. Neat.

Thanks for the comment, Martin. I would naturally use Brent/Sercal's excellent S-Power regulators but they are a fair bit more expensive (and good for several amps' current)!

The applications I have in mind (eventually) will be discrete output stage +/-15v and +5v for DAC analogue, then maybe digital too. Possibly others eventually. Clock and clock divider each have their own Audiocom Super Reg.

Simon

 

[6h5c]

Ray,

Are those what you used on your discrete output stage?

Cheers,
Simon

Yep, those are the ones. I use them in my SA8400 on the PCB of the output stage and as 3.3V and 5.6V regs for the DAC. I like 'em very much, they gave lot's of improvement over the standard regs.
I did modify them though

Ray

 

[SimontY]

Yep, those are the ones. I use them in my SA8400 on the PCB of the output stage and as 3.3V and 5.6V regs for the DAC. I like 'em very much, they gave lot's of improvement over the standard regs.
I did modify them though

Ray

Excellent work, did you put that chip cap on the back??

Simon

 

[6h5c]

Yes, that's my work, it's a 100n PPS cap. The main reason to replace that was to get the shunts to fit on the outputstage PCB. The red 100n Wima's were fitted on the other side originally, where they interfered with the output caps. I also fitted a few different SMD resistors to lower the shunt-current a bit. These things were drawing 80...100mA at no load, where I needed only 30...40mA. So, to keep the dissipation low, I altered the shunt-current to twice the load current required. Seemed like a good ratio to me.

Regards,

Ray

 

[iTiberius]

Wow what a huge thread !!!

Have anyone tried to change TCA0372 with LM3886 ?
This should work at +/-12v voltage supply quite well.

I plan to implement such mod into an AvantGarde CDI and I wonder if someone have some experience.

Thanks/Regards,
Tibi

 

[martin clark]

Why would you want to do that? It's a servo, the audio /distortion qualities don't matter at all; and I suspect the 3886 will draw a lot more power to do, at best, the same job. The TCA part only draws 5mA at idle... and is characterised to do the job.

 

[SimontY]

Why would you want to do that? It's a servo, the audio /distortion qualities don't matter at all; and I suspect the 3886 will draw a lot more power to do, at best, the same job.

I don't really know the ins and outs of it but it's already been established (Brents, Glenns and partly my mods) that improving the servo section improves the sound greatly.

Simon

 

[martin clark]

Ah, but does that servo improvement really come from the servos action, or from removing the noise it puts on the PSU / radiates from interfering with the rest of the player?

In CD servos the actual loop bandwidth and slew rate requirement are entirely defined by the decoder's hardware/software. Apart from feeding this bit decent clock signal, and isolating its PSU, thereafter it's best from an interference and loop stability POV to use the slowest, most stable and lowest-power driver parts that will do the job.



EDIT - If you're going to stick a 3886 anywhere in a modded 63, do it at the output and makes yourself a Linn Classik beater

 

[SimontY]

Ah, but does that servo improvement really come from the servos action, or from removing the noise it puts on the PSU / radiates from interfering with the rest of the player?

In CD servos the actual loop bandwidth and slew rate requirement are entirely defined by the decoder's hardware/software. Apart from feeding this bit decent clock signal, and isolating its PSU, thereafter it's best from an interference and loop stability POV to use the slowest, most stable and lowest-power driver parts that will do the job.

EDIT - If you're going to stick a 3886 anywhere in a modded 63, do it at the output and makes yourself a Linn Classik beater

I like your suggestion Martin. The 3886 is very good indeed. I'm about to begin making an über version of mine.

About the servo... Brent isolated every 5v rail in the player and he reported additional improvements powering the servo from Super-regulators.

My servo is still running on the original 7805! But at least it's been reclocked, with LARGE gains.

Simon

 

[iTiberius]

I already replaced, in Marantz CD56/65/75/40/50/60 L272 and TCA0372 with L2720/2/4 series and was a clear improvement.

In fact I'm currently testing this mod (TCA0372 with LM3886) with separate and regulated supply @ +/-24V for LM's in a Marantz CD40.

I can tell you that the servo is much much faster ...

Regards,
Tibi

 

[martin clark]

Well in that case - try it and let us know what you find!

 

[jksmurf]

Also, I have made a 12V superraygulator, but have had problems building a -12V version. Is the positive reg the most important of the two, or are they equally important? Simon (Oz)

In hindsight, if I tried to send +15V through the input maybe that was the problem. Yep, Reviewing Rays specs shows me that I needed to feed it with a negative voltage. Simon (Oz)

OK, my wife has made the +12V super-rayg no probs (+11.96v) . However, like Simon(Oz), is becoming frustrated on her third attempt at the -12V version (-8V...). Simon(Oz) mentions above about "needing to feed it with a negative voltage.". Wife asks "how do you guys TEST it with a -ve Voltage?" She doesn't want to test it using the main PCB in case it all goes to pot. She tested the +12V using the Tx she built for the clock flea with +18V. Can you give me some hints before she clobbers me with a frying pan?

Cheers,

k.

 

[martin clark]

If you have an isolated raw supply, say 0-20v or something, you just hook up the negative regulator (+20v side to the 0v input on the reg, 0v supply to -ve input) and verify the output voltage.

No risk of blowing it up, its just the way you look at it - a differential measurement.

 

[jksmurf]

If you have an isolated raw supply, say 0-20v or something, you just hook up the negative regulator (+20v side to the 0v input on the reg, 0v supply to -ve input) and verify the output voltage.

No risk of blowing it up, its just the way you look at it - a differential measurement.

Cheers Martin, apprecaite that; she's got a DC Power supply unit at work, but (now says...) she was scared to switch it like that in case she blew something ... which she did

k.

 

[SimontY]

as long as the Kiwames fit on the board, i'm the last person that's going to say anything about madness....

The 2w Kiwames are quite small, so following a test in my preamp, I will order those. The 1w Holco H2 are actually bigger.

I love what the 5w Kiwames are doing in my speakers! I'm using two in each for tweeter L-pad duties and it's great, very musical and easygoing.

Just ordering the filter caps and coils and the diodes from Farnell now...

Simon

 

[Glenn2]

I regulated the servo amps in a CD63 using LM317/337 (they are unregulated, unlike the CD67) and the bass got a whole lot more tuneful.

I expect I was just stopping the junk from these entering back into the raw supply for the other reg(s).

Andy Weekes posted some plots on here back then, showing nasty LF spikes in the noise spectrum when he hit PLAY.

I had an Arcam Alpha (TDA1541A) player hanging around at the time which I had just fixed for a friend. It had much tighter bass... until I did this. I have a hunch that the servo noise translates directly into LF jitter.

 

[jksmurf]

Hi Ozzy Simon

They are equally important. The op amps have no ground reference on the supply rails and are actually running on 24V (across the supply rails). So you need both regs the same to get full benefit.

Also take note that the pin outs of -ve regs are different to +ve regs

+ve = I G O

-ve = I O G

Brent

OK, I think I figured it out.

With reg facing towards me, legs down, I had them as follows:

| | |
I GO

in, ground, out

I think that's correct for the +12V reg, but the -12V should be:

| | |
GI O

Shouldn't have rushed it.

Is the Upper one for LM317/LM337 regs and kower 79M12? Is that the reason for the difference for pinout of +12V/-12V regs?

k.

 

[Malefoda]

As I am (un?)famous here now, I've a important question wich needs an answer if I want to sleep at night.
This question is also written in antoher thread wich is only seen by Simon and few chinese...

It's this one:
So MKP are metalized popyPropylene.
They do better than other in audio path.
I've found some from Epcos and Vishay wich seems prefered beside Epcos. But they are X2 rated, what does that mean and is it still good for audio?
Epcos are far cheaper anyway if still sounds good for MKP.

It can be for headphones part as DC blocking caps after main opamps. But this is not or maybe...

 

[SimontY]

they are X2 rated, what does that mean and is it still good for audio?

X2 means they can safely go across live and neutral on the mains supply and if they fail it won't cause a disaster. You might be better off buying caps designed for audio. Or at least one that somebody can recommend as good in the signal path.

Simon

ps - caps, diodes 'n' coils from Farnell came today, they're so fast!