Marantz CD63 & CD67 mods list

[Fidelity Audio]

mmmm interesting. I'll put that on my to do list.

[davidsrsb]

If you look at the servo data sheet, the various +5V supplies are each fed from a small resistor with an e-cap decoupler. The CD6000 actually has 4R7 resistors on the regulators (defeats any benefit from better regulators) and just 47n decoupling.

[6h5c]

Originally posted by davidsrsb
If you look at the servo data sheet, the various +5V supplies are each fed from a small resistor with an e-cap decoupler. The CD6000 actually has 4R7 resistors on the regulators (defeats any benefit from better regulators) and just 47n decoupling.

Well, at least they split the 5V line in +5VA and +5VB, it's better then nothing...

But that's why we replace the 4,7R resistors by low-ohmic inductors, and replace the 47n ceramics by a large el-cap + 100n.

The best thing is to get rid of anything in series with the supply-line of course, and use a regulator for every pin. This is not as hard as it seems, because you can just leave a series resistor in there, and put a shunt regulator behind it. But then again, unless you want to use TL431's, you would still need a PCB with parts to build a good shunt...

Ray.

[Fidelity Audio]

Tomorrow I am moving the two audiocom super regs (+/- 12V) from the usual place (Q801/Q802) and placing them right next to the op amps where the feed resistors go.


I will post pics tomorrow after fitment.

I wont bother replacing Q801/Q802 as they only power HDAM , muting and op amps because the HDAM and muting is removed.

[6h5c]

Originally posted by rowemeister
Tomorrow I am moving the two audiocom super regs (+/- 12V) from the usual place (Q801/Q802) and placing them right next to the op amps where the feed resistors go.

I will post pics tomorrow after fitment.

I wont bother replacing Q801/Q802 as they only power HDAM , muting and op amps because the HDAM and muting is removed.

Are you going to jumper Q801/802 and use the PCB traces for raw supply?

Ray.

[Fidelity Audio]

Probably not. I will probably run a wire direct from the I/P of Q801/Q802 to the regs.

I will be able to route it nicely then too

[guido]

Originally posted by 6h5c





Hi rowemeister!

Some mod eh? I think you will like the OPA627 + OPA132 even better. Make sure you put the 627 after the DAC and 132 in the filter.

What Guido means is that the regs are o.k. obviously, but the noise is not coming from them, it's coming out of the DAC!

Take a look at the datasheet. The SM5872 has separate +5V lines for the digital part (pins 15 & 16), the analog part (pins 17, 19, 21, 22, 24, 26) and the clock part (pins 2 & 27). Since the supply lines of the DAC are not separated completely now, some (digital...) noise will be coupled to other (analog...) stages. If you tie the clock part to the supply of your external clock and the digital part to the +5V you use for the decoder, you can fit a third (clean) +5V for the analog part. Have I 'translated' this correctly Guido?

The key is to separate these lines. You can do that by looking at the board layout.
The clock part is just two internal gates that used to form the oscillator, but they still carry the 16,9MHz. It's easy: it is already separated, it get's +5V through RD01. You should tie it to the external clock supply to keep all the noise in one place.

The digital AND the analog part get +5V through U164. Then through RD04 to the digital part, and U200 to the analog part. If you take these two out, you have separated these parts too

Ray.

I did not try or measure much, just did what i thought was right.
Since the clock is generated by the dac i assume that the output stages are latched with that and not with the incoming i2s signals (i need to read the datasheet to verify )

So jitter on incoming i2s (from decoder) I ASSUME is not that important. Feeding the clock circuit in the dac with a clean powersupply should be to minimize jitter on the latching.

So i used the same 5V for clock and dac clock circuits (there is a limit to the regs i was willing to use, only 8 left... :->. This should minimize jitter. Then i used a second reg for the analog supply to minimize noise going into the dac's analog outputs.

I left the digital powersupply of the dac and the decoder and the rest of the stuff (controllers etc) to the players +5. Just 'upgraded' the filter components as usuall.

took out the hdam too recently. really worth it!

[6h5c]

Hi Guido,

Well, I didn't measure it either... Just followed my own common sense.

I use exactly the same configuration, starting from the same viewpoints.
Seems to work fine and with some regulators to spare

I looked into the DAC datasheet, it's input data is latched: "isolating them from jitter or skew on the input circuit clocks"

Ray.

[Fidelity Audio]

whooo my brown dogs have arrived

So I put OPA627 on IC1
and OPA132 on IC2

[6h5c]

And, and, and, and, and, and ???????
:n ote: