Re: Re: Re: Re: D1V3 in Progress...
The Marantz CD 10 is a disaster to be modified. I could put a better clock in it but.... Moreover, a have met Guido today and he is not so fond of the CD 10......
The CD72-SE, is that also Marantz? I cannot find any documentation on: 'Marantz CD72 SE'.....
Before I promise to modify the player (or transport as it is called here) I must have some insight in the circuits.....
By the way, do you use a separate DAC as well, and if so, which one?
irgendjemand said:
The Marantz CD 10 is a disaster to be modified. I could put a better clock in it but.... Moreover, a have met Guido today and he is not so fond of the CD 10......
The CD72-SE, is that also Marantz? I cannot find any documentation on: 'Marantz CD72 SE'.....
Before I promise to modify the player (or transport as it is called here) I must have some insight in the circuits.....
By the way, do you use a separate DAC as well, and if so, which one?
Re: Re: To All,
Can anybody tell me, or refer to, the arguments for "8 x PCMxx parallel" ?
Bernhard said:
Very silently I want to mention plans for production of a 4 layer PCB with 8 x PCM56 parallel, 2 PCBs per DAC channel (balanced), probably gold plated traces. Total 32 x PCM56 per DAC.
Start to save your
Can anybody tell me, or refer to, the arguments for "8 x PCMxx parallel" ?
Re: Re: Re: Re: D1V3 in Progress...
Guess your CD10 is better, right? CD72 is a bitstream machine with plastic cdm4, CD10 has a metal cdm4. Now i don't think the material matters as such, but what could matter is that the plastic cdm has a 'toy' motor for the spindle and the metal has the hall motor.
Next to that, it's wheight, power supplies etc. I think both have a 7310 is the decoder and a pcfxxxx chip for spdif out. But i don't have schematics to back that up.
Fixed a cd10 one day (laser transplant) and it's a nice machine. From what i can remember, the dac section is on a separate pcb. Handy for modding.
irgendjemand said:
Guess your CD10 is better, right? CD72 is a bitstream machine with plastic cdm4, CD10 has a metal cdm4. Now i don't think the material matters as such, but what could matter is that the plastic cdm has a 'toy' motor for the spindle and the metal has the hall motor.
Next to that, it's wheight, power supplies etc. I think both have a 7310 is the decoder and a pcfxxxx chip for spdif out. But i don't have schematics to back that up.
Fixed a cd10 one day (laser transplant) and it's a nice machine. From what i can remember, the dac section is on a separate pcb. Handy for modding.
Re: Re: Re: Re: Re: D1V3 in Progress...
The CD 10 is a bit stream machine ! I wonder if the CD72 is also a bit stream one. Modifying bit streem DAC's is not so easy....
I thought a CDM4 is a CDM4. There is NO plastic CDM4 AND a metal CDM4....
If the I2S in the system is well reclocked and the 'back end' behind it is powered from a SEPARATE power supply (from 220 V AC on) you build a real brick wall between the turn table, the servo system and, and, and, so the system becomes INDEPENDENT of the front end....
guido said:
The CD 10 is a bit stream machine ! I wonder if the CD72 is also a bit stream one. Modifying bit streem DAC's is not so easy....
I thought a CDM4 is a CDM4. There is NO plastic CDM4 AND a metal CDM4....
If the I2S in the system is well reclocked and the 'back end' behind it is powered from a SEPARATE power supply (from 220 V AC on) you build a real brick wall between the turn table, the servo system and, and, and, so the system becomes INDEPENDENT of the front end....
Re: Re: Re: Re: Re: Re: D1V3 in Progress...
ahum......
http://img123.imageshack.us/img123/8456/CDM4.jpg
cheers
PA0SU said:
ahum......
http://img123.imageshack.us/img123/8456/CDM4.jpg
cheers
Bernhard said:
Your chip definitions seem to be rather fluid Bernhard. One days mediocre is the next days garbage or so it seems:
Note: Just looked at the PCM56 data sheets and see the K grade -60dB THD+N quoted as -35dB typ. vs the PCM63 at -48dB. I don't quite understand the massive difference between Bernhard's figures and the factory spec.
Seriously, I don't see how it is possible to directly compare absolute figures given the variables in play:
- Measuring device
- FFT settings
- Test Signal frequency
- DAC supporting electronics
I had thought it had been established very early in the thread that data was only comparable between tests done in the same rig, otherwise we are comparing apples and avocados.
fwiw there is a fantastic article on theortical dac SNR and FFT noise floors here:
http://www.analog.com/en/content/0,2886,761%5F795%5F88014%5F0,00.html
There is roughly a 20dB difference between the FFT noise floor of a 4096 FFT and the 323,xxx FFT I was using in RMAA. The theoretical noise floor of a 20bit dac is something like -122dB, and BB specified the typical idle noise floor as -120dB. The theoretical noise floor of a 16bit dac - like pcm56 - is -98dB, so the -67dB quoted by Bernhard sits something like 24dB below the theoretical noise floor.
- Measuring device
- FFT settings
- Test Signal frequency
- DAC supporting electronics
I had thought it had been established very early in the thread that data was only comparable between tests done in the same rig, otherwise we are comparing apples and avocados.
fwiw there is a fantastic article on theortical dac SNR and FFT noise floors here:
http://www.analog.com/en/content/0,2886,761%5F795%5F88014%5F0,00.html
There is roughly a 20dB difference between the FFT noise floor of a 4096 FFT and the 323,xxx FFT I was using in RMAA. The theoretical noise floor of a 20bit dac is something like -122dB, and BB specified the typical idle noise floor as -120dB. The theoretical noise floor of a 16bit dac - like pcm56 - is -98dB, so the -67dB quoted by Bernhard sits something like 24dB below the theoretical noise floor.
PA0SU said:
Please see the nice & supportive text of GuidoB about the CD10, here: Fixed a cd10 one day (laser transplant) and it's a nice machine. From what i can remember, the dac section is on a separate pcb. Handy for modding
To your information – it is being used by me as a transport, only.
I am using a separate DAC: It is a strongly modified AVM DAC 1.2 DAC/pre, with CS8414/DF1704/PCM63P-Y, OsCons in the Digital section; 4x Burson's OpAmps in I/V + Black Gate, Class A biased OpAmps (8x BB 627 BP) in the preamp section.
Next thing on my own check list: XO in the DAC and the SPDIF-Out of the CD10. Am I wrong?
spzzzzkt said:Seriously, I don't see how it is possible to directly compare absolute figures given the variables in play:
- Measuring device
- FFT settings
- Test Signal frequency
- DAC supporting electronics
I had thought it had been established very early in the thread that data was only comparable between tests done in the same rig, otherwise we are comparing apples and avocados.
The theoretical noise floor of a 16bit dac - like pcm56 - is -98dB, so the -67dB quoted by Bernhard sits something like 24dB below the theoretical noise floor.
I tested PCM63s with the same analyzer settings and they had no chance.
The noise floor goes down because of averaging.
irgendjemand said:
Yes, you should put the XO in the DAC (near the reclocking circuits just before the PCM63's) lead back the XO-signal to the CD10 and enhance the SPDIF output in the CD10. This is in short.
If there is no reclocking circuit in the DAC, it should be inserted!
My offer is:
update the SPDIF output from the CD10,
insert the reclocking circuits in the DAC (if they are not already there),
place a real good XO in the DAC, and
make an extra B&C-connection between DAC and CD10 for the 'clock-back'.
I'm as curious as you to hear the differences.....
irgendjemand said:
As a retired man i take a looooooooong holiday..... and 'the job' will take more than one day.....
Another important question: are the clock frequencies in the DAC and the CD10 equal? The CD10 runs on 11.2896 MHz. If the frequency in the DAC is different, then there is a lot more work to do....
16 or 20 bits?
Gentlemen,
In the article mentioned before Eq. 10 gives an SNR of 99,25 dB for 16 bits and 134,62 dB for 20 bits with 8x oversampling.
The source for the measurements is a CD.
I'm wondering which figure counts in our case. The digital filter and the DAC (in my case) are both 20 bits, but the signal source is 16 bits.....
Gentlemen,
In the article mentioned before Eq. 10 gives an SNR of 99,25 dB for 16 bits and 134,62 dB for 20 bits with 8x oversampling.
The source for the measurements is a CD.
I'm wondering which figure counts in our case. The digital filter and the DAC (in my case) are both 20 bits, but the signal source is 16 bits.....
PA0SU said:
You might have missed my earlier Post, no. 285...
Nevertheless, I was thinking of having a "Tent-Link" which uses XO and VCXO 
. One of the reasons for this - the AVM DAC 1.2 and the Marantz CD10 Transport are having indeed different Quartz frequencies... 
May be I will do it myself at home; will take not more then one week of work ...
but I will have fun 
.
PA0SU said:
Bernhard: read the last sentence in the above article.
Averaging a number of FFTs does not further reduce the noise floor, it simply reduces the variations between the individual noise spectral component amplitudes.
It does not average the FFTs but the input signal before the FFT.
It shall reduce the noise floor of the signal, not that of the analyzer.
spzzzzkt said:
without oversampling and without FFT processing gain: 10.log (M/2)
again: without oversampling and without FFT processing gain
so the -67dB quoted by Bernhard sits something like 24dB below the theoretical noise floor.
so, without specifying the times of oversampling and the xxx-point FFT, you can't judge the numbers of Bernhard.
8 times oversampling yields 12,46 dB,
a 4096-point FFT yields 33 dB,
a 2048-point FFT yields 6 dB, and
a 323,xxx-point FFT yields 3 dB.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.