Simple, good quality DAC

[PMiczek]

Was the consensus in favor of making modifications to what is already a pretty good CDP w/o replacing or adding a DAC chip, or was it in favor of adding a board with a TDA1543 or similar simple DAC? (I realize there is no one RIGHT answer here, but I am curious if there is agreement among people who have done both, in this one particular case).

[Peter Daniel]

I would take a good transport and add (if not there already) TDA1541 or TDA1543 DACs, omitting filter and using I2S directly from decoder chip. It is a simple mod, and the results might be great, if done right.

[miguel2]

Hi Kuei

This weekend I opened my marantz cd48 to do the modifications you mentioned. But instead of a TDA1545 I found a surface mount 16 pin TDA1549. With 16 tiny tiny legs I really dont' want to mess with it. But then I though I could get the I2S signal from it by soldering some wires to the respective (sooo tiny) legs, and then use an external dac chip. Would this work, I mean, would it be ok to connect the I2S signals to 2 dac chips? This way I would keep the cdp working in its original state and would have an external dac chip to play with passive conversion, like the circuit you have mentioned.

Miguel

[ThorstenL]

Koinichiwa,

Originally posted by miguel2


This weekend I opened my marantz cd48 to do the modifications you mentioned. But instead of a TDA1545 I found a surface mount 16 pin TDA1549.

Seems you are somewhat unlucky.

Seems Marantz "upgraded" the CD - 4X after the version i encountered to be based on the Philips CD-75X series, instead of the (in some ways as "DIY Upgrade Platform" superior CD71X/72X).

The TDA1549 is internally very similar to the TDA1545 in some ways, but it replaces the passive divider with a "Bitstream/Delta-Sigma" modulator and places the Op-Amp's on board of the chip.

For best sound from this take the output directly of the chip outputs (pin4/7) and take care to improve the quality of Cext (pin 8) and of the output coupling Cap's.

Originally posted by miguel2

Would this work, I mean, would it be ok to connect the I2S signals to 2 dac chips? This way I would keep the cdp working in its original state and would have an external dac chip to play with passive conversion, like the circuit you have mentioned.

Sure. The I2S input pins on the TDA1549 are:

Data - 16
LCRCK/WS - 1
Bitclock - 2

The data of course is still 4 * Oversampling.

Sayonara

[rbroer]

Miguel2,

With a trick you can even bypass the TDA1549's internal opamps.

Replace the external filter capacitors (Cext1 and Cext2, see datasheet) with shortcuts.
Now you have a current output and can use your favourite I/V opamp or discrete circuit by feeding the TDA1549 outputs again into the -input and connecting the Vref to the +input of your new I/V stage (I always liked LM6172 at that location).

Compared to the TDA1545A and feeding it with the same 4*fs signal from the SAA737X servo/decoder, it sounds less digital IMHO.

For starters I would get the maximum performance from your Marantz CD48 player first; feed the DAC from three individual supplies, use direct voltage output from DAC or bypass internal opamp and add "better" output stage, and, assuming this player has 8.X MHz clock like the Philips CD75X series, replace it with a Tent clock or Kwak clock, but also connect it directly to the BCK input of the TDA1549 as well.

KYW,
Does the Philips 72X series have 1*fs I2S signals ? How's the space inside these players, like the newer 75X series ?

Regards,

[miguel2]

Hi Kuei and Rudolf,

I never worked with SMD chips before and I dont want to ruin this cdp. So for now I am buying a TDA1543 and connect it to the I2S signals taken from the TDA1549 pins. I want to hear this first before atempting something more complicated with those small pins. I saw Peter Daniels circuit on this cheap, with 1K on the output and 47k to ground and I will try that. I will use batteries for power the chip, as Kuei suggested. As I am not buying expensive caps for now, I was thinking in using back to back regular elecytrolytics (4.7uF), but dont know if this back to back arrangement has any advantage.

Miguel

[ThorstenL]

Konichiwa,

Originally posted by rbroer
With a trick you can even bypass the TDA1549's internal opamps.

Replace the external filter capacitors (Cext1 and Cext2, see datasheet) with shortcuts.
Now you have a current output and can use your favourite I/V opamp or discrete circuit by feeding the TDA1549 outputs again into the -input and connecting the Vref to the +input of your new I/V stage (I always liked LM6172 at that location).

Did you try this actually?

I would expect this not to work, as with Cext shorted the inverting input is still a virtual ground, sinking all the current from the DAC and delivering non externally. But I'm happy to stand corrected and would like to have a way to circumvent these sections in the various Philips DAC's...

Originally posted by rbroer
Does the Philips 72X series have 1*fs I2S signals ?

No.

Originally posted by rbroer

How's the space inside these players, like the newer 75X series ?

The 71X/72X is basically empty, apart from X-Port and mains TX.

Originally posted by miguel2

I never worked with SMD chips before and I dont want to ruin this cdp. So for now I am buying a TDA1543 and connect it to the I2S signals taken from the TDA1549 pins.

If you trace the pin's and their PCB traces you can probably find a point where there is a jumper wire or a resistor to slow down the edge of the signal. Solder to there, it's much easier.... ;-)

Use miniature PTFE Coax cable, ground screens at the DAC only and use a piece of copper foil (insulated) or silverfoil to make a low inductance ground return from DAC to Player PCB.

Originally posted by miguel2

I saw Peter Daniels circuit on this cheap, with 1K on the output and 47k to ground and I will try that. I will use batteries for power the chip, as Kuei suggested.

Actually, for the TDA1543 I would recommend 6 or 7pcs "D" Cell rechargables with a suitably scaled I/V resistor (I hear 1K5 should work well). Then put a 2K2 adjustable at the Ref pin and adjust the resistor for lowest distortion (preferably using a 0dbfs test cd and and a 'scope) with the batteries near "empty" (1V per battery).

To simulate empty batteries when having full batteries use only 4 or 5 pcs "D" Cells to set the reference Voltage, then connect the other cells to go.

The TDA1543 draws a lot more current than the TDA1545, so powring it from non-recharables will drain your batteries and pockets fast.

Originally posted by miguel2

As I am not buying expensive caps for now, I was thinking in using back to back regular elecytrolytics (4.7uF), but dont know if this back to back arrangement has any advantage.

Not in this case. Use one Cap, plus towards the DAC.

Sayonara

[rbroer]

Originally posted by rbroer
With a trick you can even bypass the TDA1549's internal opamps.

Replace the external filter capacitors (Cext1 and Cext2, see datasheet) with shortcuts.
Now you have a current output and can use your favourite I/V opamp or discrete circuit by feeding the TDA1549 outputs again into the -input and connecting the Vref to the +input of your new I/V stage (I always liked LM6172 at that location).

Did you try this actually?

I would expect this not to work, as with Cext shorted the inverting input is still a virtual ground, sinking all the current from the DAC and delivering non externally. But I'm happy to stand corrected and would like to have a way to circumvent these sections in the various Philips DAC's...

Yup.
The "virtual ground" or in fact "virtual Vref" is just an abstract result of the feedback loop; there's no way the opamp input stage sinks or sources current. One only looses some really minor bias current (pA ~ nA) for the input stages, but the current can only flow through Riv (made 0 ohms in this case)
A few years ago when still using opamps, I used the LM6172 which gave a nice improvement over the DAC's internal opamps.

[Elso Kwak]

Hi,
I found this old post on the Audio Asylum:
http://db.audioasylum.com/cgi/m.pl?f...erter&session=
With The TDA1543 I found the OPA604 performs very well. Connection is as in the datasheet of the TDA1543. The resistor from Vref to ground is 3k3 as used by Philips in there players (CD380 etc.). Offset at the output is not too high but not zero.
I also tried the "looped cascode" by Wildmonkeysects and the one transistor scheme by Thijs Schrama but I was not impressed.
I feel if the first pole of the analog lowpass filter is in the IV-converter [cap across the feedbackresistor] the demands on the opamp are not that high i.e no need for ultrahighspeed opamp like AD817, LM6172.

[tschrama]

I also tried the "looped cascode" by Wildmonkeysects and the one transistor by Thijs Schrama but I was not impressed.

Hi Elso,

Just incase anybody cares: the one-transistor I/V stage was just a very very simple circuit to increase the dynamics of my TDA1541A DAC compaired to a passive resistor I/V. Clearly Jocko's simple I/V would be the obvious next step as it would be a clear improvement if only from an enginering point of view.

Regards,
Thijs