It's all in the planning
I like to change only one thing at a time, so the result is clear.The clamp is a leftover from a home-made toploader cd drive I once bought, it's 90 grams.
It does sound better. The whole tray mechanism seems to resonate quite a bit and thus deteriorate the sound. It was quite obvious: the tubes in the output stage are microphonic as hell. With the tray in place and a cd paused I could clearly hear noise from the disc spinning. That's gone now, all I can hear is a thump when it stops and the laser goes back to the initial position
It was quite obvious: the tubes in the output stage are microphonic as hell. With the tray in place and a cd paused I could clearly hear noise from the disc spinning. That's gone now, all I can hear is a thump when it stops and the laser goes back to the initial position
Have you tried cryo treated valves? Cryo valves are much less microphinic than non treated IMHO. I used to have a modded copy MF valve buffer post I/V on the TDA. TBH the TDA is so warm and smooth once it has decent psu & regulation you don't don't need a tube. The characteristics are inherrent to the DAC. But then again it depends on the tube stage configuration.
Again IMHO, if you go for the sharpest most detail revealing dryest sounding components on this DAC (ZA de-couple, LME opamp I/V etc) you'll still have the warmth and depth most people associate with a valve sound.
the TDA is so warm and smooth once it has decent psu & regulation you don't don't need a tube. The characteristics are inherrent to the DAC. But then again it depends on the tube stage configuration.
Again IMHO, if you go for the sharpest most detail revealing dryest sounding components on this DAC (ZA de-couple, LME opamp I/V etc) you'll still have the warmth and depth most people associate with a valve sound.
I generally agree with this Ian. I have to say that with a Pass D1 class A FET output stage the sound is as warm and cuddly as anyone could need, only with full detail and clarity at the same time. You could call that nirvana, but I know better than to say something's finished - there's always more
It can be done in these players too. I did it in my Marantz CD94 and got a nice increase in detail, separation etc. Much like in the cd63, but to a slightly lesser exten, which I think is because the servo section in the cd94 is already pretty sorted.
Cheers, Lee.
Fair enough.
On my CD624 (I think that's the same as a Marantz CD40 or 41, can't remember), and CD930/940, and possibly the CD610 as well, the decoder chips simply run off the 11MHz MCLK. I assumed the seperate clock for the decoder was quite unique to the chipset in the CD63 - for Philips transports anyway.
On my CD624 (I think that's the same as a Marantz CD40 or 41, can't remember), and CD930/940, and possibly the CD610 as well, the decoder chips simply run off the 11MHz MCLK. I assumed the seperate clock for the decoder was quite unique to the chipset in the CD63 - for Philips transports anyway.
Hi Phil, I'm pretty sure that the chip that controls the servo in these old players is simply a programmed eprom. I reckon it'd be worth trying to separate that from the main audio clock lines.
On my ancient Philips I had with cdm2, the oscillator was on the board mounted underneath the mech.
Simon, one is for the dac/decoder etc, one for the servo control and one for the display I think.
Cheers, Lee.
The chip that controls the servo is a microcontroller (much more than an eprom), called servo uP. Which is on the main pcb or sub assembly above it (depending on version) and not on cdm2 pcb. And it has it's own xtal. There are cdm2 versions with the servo uP on the cdm2 pcb, but not in this case. Then there's the control pcb with the main contol processor with an xtal. There's another processor for the FTS memory, running on the clock of the control uP. Both xtals are 6MHz.
The audio clock is connected to the 7720 filter, the 11.2896 MHz xtal. The decoder receives a clock signal from the filter, the dac is clocked on the I2S clock.
Wouldn't hurt to have a look at the service manual...
Yes, I2S is the datastream. The TDA1541 non A has a separate clock input (pin4) which clocks the output registers. Different from pin2 which clocks the I2S data into the chip. So the idea was to use low jitter clock for pin 4 (see picture in 7220 datasheet to get the idea. and i'm not saying anything low jitter can come out of a 7220 anyway).
In reality they used the same clock on 2 and 4, they are always connected in philips players (and nearly all others). So with the 1541A they didn't bother anymore. 4 has no function in I2S mode. But still they connected 2 and 4 in the players, even in the CD7.... I've had a 1541A dac with 4 grounded, worked fine.
CDM2 - ah, that explains it. I've only played with 4s, 9s and 12s.
PMD100 - well, the information is all elsewhere, and all the investigation was done by Guido, but this is what I did....
I sniffed off the I2S from the 7210 then used Guido's conversion circuit to convert the I2S into the format required by the PMD100 (although I think I omitted an inverter as I couldn't get it to work with it - can't fully remember now). I think Guido built his with a GAL, and used a PIC to program it into different modes. I used TTL logic instead, and used a set of DIP switches to set the limited modes available with standalone mode.
Then the output (which can be configured - unlike the input) needs to interface the TDA1541A. Because this is 8x oversampling, the DAC chip needs to be running using LE instead WS, and using DATAL and DATAR as inputs (basically doubling the data input for the same clock frequency as used for 4x OS).
I can't remember the exact details to hand, i.e. which pins, but you need to use all for input pins independently (BCLK, LE, DATAL and DATAR) and you need to change the voltage on the input select pin (it accepts +5v, 0v and -5v). You really need to check the schematics and datasheets.
Also, as rfbrw has pointed out, it is true that the PMD100 is outputting data in bursts that are out of spec for the TDA1541A, but I've had no issues with it myself with several DAC chips- although I don't measure things much.
Buffering I2S, well, You could use D type flip flops. I've done this in the past, although don't do it currently. I don't think there's much point buffering the output of the 7210 though. I think, if you were to do it at all, do it just before it enters the DAC (and add filtering accoridngly - e.g. series resistors on the data lines). If you do it on the output of the decoder, you could well add jitter later on, so I suspect it might be fruitless.
Cheers,
Phil
PMD100 - well, the information is all elsewhere, and all the investigation was done by Guido, but this is what I did....
I sniffed off the I2S from the 7210 then used Guido's conversion circuit to convert the I2S into the format required by the PMD100 (although I think I omitted an inverter as I couldn't get it to work with it - can't fully remember now). I think Guido built his with a GAL, and used a PIC to program it into different modes. I used TTL logic instead, and used a set of DIP switches to set the limited modes available with standalone mode.
Then the output (which can be configured - unlike the input) needs to interface the TDA1541A. Because this is 8x oversampling, the DAC chip needs to be running using LE instead WS, and using DATAL and DATAR as inputs (basically doubling the data input for the same clock frequency as used for 4x OS).
I can't remember the exact details to hand, i.e. which pins, but you need to use all for input pins independently (BCLK, LE, DATAL and DATAR) and you need to change the voltage on the input select pin (it accepts +5v, 0v and -5v). You really need to check the schematics and datasheets.
Also, as rfbrw has pointed out, it is true that the PMD100 is outputting data in bursts that are out of spec for the TDA1541A, but I've had no issues with it myself with several DAC chips- although I don't measure things much.
Buffering I2S, well, You could use D type flip flops. I've done this in the past, although don't do it currently. I don't think there's much point buffering the output of the 7210 though. I think, if you were to do it at all, do it just before it enters the DAC (and add filtering accoridngly - e.g. series resistors on the data lines). If you do it on the output of the decoder, you could well add jitter later on, so I suspect it might be fruitless.
Cheers,
Phil