Re: Re: Re: Re: Re: Re: Re: D1V3 in Progress...
ahum
That's a cdm1mk2. In my view a metal cdm2. The cdm4/11 has hall too, like cdm1mk2 and cdm2.
Attached a cdm4/19 and 4/36 (the metal one) during laser transplant of cd10 player. Here the cdm4/19 is without hall, easy to spot the 'toy' motor. Btw, the player worked afterwards. Much to the surprise of some friends which were present...
The cd72 is bitstream: dac7, but only a single one. cd10 has two. Transport according to marantz folder is cdm4MM vs cdm4MD, where MD is the 4/36 metal one.
Back to the thread topic.
Guido Tent said:
ahum
That's a cdm1mk2. In my view a metal cdm2. The cdm4/11 has hall too, like cdm1mk2 and cdm2.
Attached a cdm4/19 and 4/36 (the metal one) during laser transplant of cd10 player. Here the cdm4/19 is without hall, easy to spot the 'toy' motor. Btw, the player worked afterwards. Much to the surprise of some friends which were present...
The cd72 is bitstream: dac7, but only a single one. cd10 has two. Transport according to marantz folder is cdm4MM vs cdm4MD, where MD is the 4/36 metal one.
Back to the thread topic.
Attachments
irgendjemand 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
What i meant with nice is that it's well build and it looked like the dac pcb can be easily replaced by something else. Can't remember though if the filter/dac feeding the tda1547's was also on that pcb. There might be enough room in there to put the dac inside. So no comment on sound
And one more picture
Attachments
Re: Re: Re: Re: Re: Re: Re: Re: D1V3 in Progress...
Both are bitstreamers, Okay, thanks.
An interresting site:
http://www.marantzphilips.nl/index.asp?strPage=Daclist&strBrand=Various#top
In this table I find:
CD10: CDM4/36MD
CD72: CDM4/19
CD72a: CDM-4
CD72 Mk II SE: CDM-4/28
Don't ask me for the differencies......
Back to the thread topic.
guido said:
Both are bitstreamers, Okay, thanks.
An interresting site:
http://www.marantzphilips.nl/index.asp?strPage=Daclist&strBrand=Various#top
In this table I find:
CD10: CDM4/36MD
CD72: CDM4/19
CD72a: CDM-4
CD72 Mk II SE: CDM-4/28
Don't ask me for the differencies......
Back to the thread topic.
Originally posted by guido
as far as I can see in the doc, the I2S-interface is exactly there!
yes, the filtering is also there, so that the PCB in question could be replaced by a digital filter, reclocking circuit (with the XO close to it), DAC's (PCM) and IV-convertor. In case you get an exellent player !!!!
That's what I did with my CD624.........
irgendjemand said:
as far as I can see in the doc, the I2S-interface is exactly there!
yes, the filtering is also there, so that the PCB in question could be replaced by a digital filter, reclocking circuit (with the XO close to it), DAC's (PCM) and IV-convertor. In case you get an exellent player !!!!
That's what I did with my CD624.........
PA0SU said:
Gentlemens, Indeed, I didn't want to overload this thread with my CD10 (or 72SE). But OK, there are some excellent pictures of the CD10 on the page: http://www.marantzphilips.nl/index.asp?strPage=Info&strBrand=Marantz&strType=cd10
It is a real excellent Transport; Funny, I NEVER listened to it as a CD Player - and I have no idea how it sounds "as it is".
Herb, modifying it as you already did in your 624 - should be here indeed as much succesfull, and even more: The CD10 is a very well build device.
PA0SU said:
You really didn't read that very carefully did you. 🙄
The oversampling effect you refer to is related to the analog to digital conversion.
I've reread the page and some other links, and I was wrong to suggest that the processing gain only applied to ADC's.
That said I don't think you are correct to link the processing gain from oversampling with the processing gain from FFT transform?
Decreasing the quantisation noise floor of the DAC by the use of oversampling is not going to decrease the the processing gain of the FFT. Bear in mind we are measuring the analog output of the DAC, so any effect of oversampling will be apparent in the measurements without calculating processing gain. You seem to be conflating the processing gain relating to oversampling with the processing gain of the FFT. They are not the same thing.
Processing gain for the FFT is:
4096 = 33.11
2048 = 30.10
I checked the correct number of points I've used, which is actually slightly lower:
262,144 = 51.17
This corresponds with other information I've quoted that doubling the number of sampling points increases the FFT gain by 3dB.
That said I don't think you are correct to link the processing gain from oversampling with the processing gain from FFT transform?
Decreasing the quantisation noise floor of the DAC by the use of oversampling is not going to decrease the the processing gain of the FFT. Bear in mind we are measuring the analog output of the DAC, so any effect of oversampling will be apparent in the measurements without calculating processing gain. You seem to be conflating the processing gain relating to oversampling with the processing gain of the FFT. They are not the same thing.
Processing gain for the FFT is:
4096 = 33.11
2048 = 30.10
I checked the correct number of points I've used, which is actually slightly lower:
262,144 = 51.17
This corresponds with other information I've quoted that doubling the number of sampling points increases the FFT gain by 3dB.
The problem I have with Bernhard's measurements is they are divorced from the noise floor of the DAC.
According to the PCM56 datasheet:
Julian Dunn describes a procedure for making this measurement in Audio Precision App Note #5 (Measurement Techniques for Digital Audio):
This measurement takes in to account all the harmonics plus the noise floor of the DAC, rather than an isolated harmonic, and this is pretty much the way figures are calculated for at least the PCM63. The exact setup is described in the datasheets.
According to the PCM56 datasheet:
Julian Dunn describes a procedure for making this measurement in Audio Precision App Note #5 (Measurement Techniques for Digital Audio):
This measurement takes in to account all the harmonics plus the noise floor of the DAC, rather than an isolated harmonic, and this is pretty much the way figures are calculated for at least the PCM63. The exact setup is described in the datasheets.
irgendjemand said:
Try it, you will be surprised.
Sure, but it will give NO beter results because with my technique you build 'a brick wall' between transport and DA-conversion.
PA0SU said:
I was just trying the Marantz CD10’s Analogue-Out (with a real good signal cable - Ortofon N7-N8). It was very easy to make a 1:1 comparison – through a selector on the AVM DAC/Pre, so I can switch between analogue-in and the digital-in while listening (dynamic is adjusted, no gain differences).
Well, the sound of the original CD10's DAC is very good - no colouring whatsoever; dynamic was right. Feels good. It is a reliable DAC “as it is” - no doubts!
However, the AVM's DAC (with all its “goodies”) sounds more as if the music is being played "in the room". There is a real stage, there are more vibrations, there is more room between instruments, etc.
I was listening to Jazz, Classics & my own orchestral recordings, Piano solo, Accordion (Richard Galliano & Michel Portal…), all truly excellent recordings: The difference are all the time there. One can't miss it.
So when listening to the CD10's (which is already slightly modified) own sound, you will feel that you are looking at a picture, from the front, but you can't say that you are part of the colours. The sound of the CD10 remains somehow “flat”. Of course, your modification would have change all this.
irgendjemand said:
I could not say it better..... and I think your separate DAC needs an update as well, so....🙂
Another point is: how are your speakers positioned in your room? I quote from my web site:
but this has little to do with the suject of this thread.....
PA0SU said:
I well bring the DAC with me... 😉
In fact, this somehow belongs to this thread, and I will tell you the reason for it later here. Before I will do this -and just in short: My speakers are good one meter in front of the wall, even though they are 2-Way bassreflex (but with active Subwoofer). It took me about 4 month to find the place for them...
Back to this thread: In my room, once I let the HiFi play "Libertango" (from the CD "Blow Up" by Galliano/Portal) or if I will perform myself some of the solos on my Double Cassotto 120 Bass (+58 MIII freebass) Accordion, you will not know which is the real accordion and which is the HiFi. The sound is accurate
So you will surely understand why I am not in a rash to give up about the wonderful sounding PCM63P-Y, even if they are fake... may be…
What shell we do about the reclocking & the DAC? May be I should leave it simply like it is
irgendjemand said:
I well bring the DAC with me... 😉
What shell we do about the reclocking & the DAC? May be I should leave it simply like it is
you mean: 'reclocking IN the DAC'
I can't say from a distance. Could you mail me some documentation (block diagrams and detailed diagrams) before to decide...
spzzzzkt said:
Why THD+N ?
Maybe one reason it is measured like this:
The way a conventional distortion analyzer does it, always results in THD+N.
The question is: Are we interested in the noise ? I am not because I have never recognized differences in noise floors between PCM56 chips.*
So I remove a part of the random noise by averaging the signal, that allows to see things hidden in the real world noise floor.
If harmonics and noise have the same level = harmonics touching the noise floor, not visible yet, then you will hear noise and distortion equally. Multiple signals mixed together.
To hide the harmonics and hear a pure signal, the harmonics have to be below the noise floor.
* And here starts a problem with 20 bit chips:
Because of the lower noise floor, the distortion is even more noticeable and linearity requirements for pure sound are much higher and more or less unachievable.
Another reason why I stay with 16 bits, besides nonos.
Bernhard said:
If the source is a 16-bit signal (as from a well generated CD with dithering) the last four [perhaps five!] (LSB) bits are consequently 1 at the input of the 20-bit DAC.
If a 20-bits DAC would be not so good (say, bit 19 and 20 are not always correct) they will be overwelmed by the dither noise of the 16-bit input signal.
Is this a correct statement?
PA0SU said:
Many thanks for this Herb. I will look for the AVM block diagrams and will come back to you soon.
I've ordered an Echo Audiofire2 firewire interface so I can redo the tests, and hopefully eliminate the issues I seem to have with the HDP2. I'll post to another thread when I've done the new tests, but for my interests this current thread has strayed so far from the original question of whether A'af's PCM63P-Y's were fakes or not that it seems utterly pointless to continue posting here.
spzzzzkt said:
That would be a pitty! If the Y's are fake or not.... in my opinion they have been produced in another plant, but is that a fake?
I THINK Ti has sold the rest of the PCM63-silicon-pieces of BB during take over with the permission to carry the name BB on it. The other plant continued production (on a different molding machine [so that they look different]) and added a Y to it.
This kind of scenario is not so strange in IC-land !
On this thread people tried to find performance differencies to support their idea if the Y's could be fake or not.....
Then the environment in which the IC's have been measured came forward [reclocking etc.]
Now we are so far to measure and listen again in 'the new' environment.....
This is about the history of this thread..... I think.....🙂 so keep an eye on it......
PA0SU said:
Herb, I agree with you and could not say it better myself...
In case you have sockets build on the DAC's PCB, would it be possible for you to describe sound differences while using a pair of Y PCMs (let’s say - in your modified CD624)?
I don't remember that you did this before, but I do remember you swapping PCMs between 2 devices, etc. What I am still not sure about is if you did found any matched-pair of Y‘s, either by measurements nor by listening.
Greetings!
- Status
- Not open for further replies.