Re: Optimizing audio interface
Hi EC,
It appears that SPDIF IP is very important, many people have
same experience.
Have you thought about making a discrete differential receiver
without transformer. The RXP and RXN IP's of CS8412/4 can have
up to +-12V on them due to inbuilt attenuation network. Hence
these IP's could be connected directly to collectors of discrete
dif pair. The devices can be RF devices with sub 1pF capacitance
and FT in 1GHz range.
This should have extremely good rejection to CM noise and better
wave form integrity than a transformer if designed carefully. It
will also have very resistive 75ohm load for the SPDIF
transmission. With transformers there are many conflicting
parameters at play. To get low C then leakage inductance
increases and wave form integrity suffers.
I might do a bit of dabling myself with this.
Cheers,
Terry
-ecdesigns- said:
Hi EC,
It appears that SPDIF IP is very important, many people have
same experience.
Have you thought about making a discrete differential receiver
without transformer. The RXP and RXN IP's of CS8412/4 can have
up to +-12V on them due to inbuilt attenuation network. Hence
these IP's could be connected directly to collectors of discrete
dif pair. The devices can be RF devices with sub 1pF capacitance
and FT in 1GHz range.
This should have extremely good rejection to CM noise and better
wave form integrity than a transformer if designed carefully. It
will also have very resistive 75ohm load for the SPDIF
transmission. With transformers there are many conflicting
parameters at play. To get low C then leakage inductance
increases and wave form integrity suffers.
I might do a bit of dabling myself with this.
Cheers,
Terry
clock and I2S
Hi EC, its a pitty you have not enough space for a KC7 upgrade.
Another option: a Tent XO, with external PS, could fit in. Whats the frequency? 16Mhz i guess.
The non-os player with discrete I/V needed a longer warm up period then the modified 304. After 15 min. it warms up indeed, the soundstage opens up a little. Warm up in the CD640 Nonos player showed a bigger difference in sound.
I have been thinking RS422 interface too with I2S
Hi EC, its a pitty you have not enough space for a KC7 upgrade.
Another option: a Tent XO, with external PS, could fit in. Whats the frequency? 16Mhz i guess.
The non-os player with discrete I/V needed a longer warm up period then the modified 304. After 15 min. it warms up indeed, the soundstage opens up a little. Warm up in the CD640 Nonos player showed a bigger difference in sound.
I have been thinking RS422 interface too with I2S
I'm starting to get worried (just ordered the parts as well, d'oh!).
I've just acquired a philips CD624 in order to tap off the I2S for a dac.
It has (can't remember the part no to hand) an saa7310 cmos decoder and the datasheet states the bclk output is 2.8224Mhz.
This is 44100*64, which seems okay to me, there are 32 bclk cycles per word (from what I can tell from the datasheet) which allows reproduction at the correct sample rate.
I assumed this would be okay for the TDA1541A (thankfully haven't ordered those yet), but its not so clear in the data sheet.
Surely it should be able to cope with a bclk frequency this low, in nos applications? Or is it not quick enough to present the sample for any reasonable duration?
I thought if I stuck to Philips/Marantz bits to experiment it would give me the best shot at this. Perhaps I'll have to use SPDIF after all? Or design some fifo of sorts and clock it out quicker with some blank data.
Cheers,
Phil
Hi rfbrw,
thanks for your reply [post#461]
See you had a good laugh
, you are right, I totally forgot about the different sony I2S specifications.
So my question is, can the Sony format be converted to the Philips format without adding new problems like addittional jitter or synchronization problems. Are there schematics for this? or can a sample rate converter be used? I need 64 cycles for the direct interpolation timing
thanks for your reply [post#461]
See you had a good laugh
, you are right, I totally forgot about the different sony I2S specifications.So my question is, can the Sony format be converted to the Philips format without adding new problems like addittional jitter or synchronization problems. Are there schematics for this? or can a sample rate converter be used? I need 64 cycles for the direct interpolation timing
Hi Ecdesigns,
found this:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=58564&highlight=
and using it for TDA1545, till i could built 8x TDA1541 DAC.
regards
Bostjan
found this:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=58564&highlight=
and using it for TDA1545, till i could built 8x TDA1541 DAC.
regards
Bostjan
I2S formats
Hi philpoole,
thanks for your reply [post#464]
I would be surprised if the Philips CD624 didn't use the Philips I2S format. BCK rate of 2.8224 MHz is ok, this would be no problem for the TDA1541A. The TDA1541A pin 27 also provides different mode selections. I have set the CS8412 interface to output a time multiplexed signal, two's complement, MSB first. Pin 27 of the TDA1541A is connected to +5V to match these settings. The TDA1541A is also not a problem.
The sony format uses 48 cycles instead of 64 among other differences, so the sample rate is still 44.1 KHz but only 24 bits for each channel are transmitted instead of 32. That's the reason why BCK is running at 2.1168 MHz (48 * 44.1 KHz). This is a problem since the octal D-I DAC needs 64 cycles for the timing chain to function correctly. But there must be a way to convert formats if necessary.
cheers,
John
Hi philpoole,
thanks for your reply [post#464]
I would be surprised if the Philips CD624 didn't use the Philips I2S format. BCK rate of 2.8224 MHz is ok, this would be no problem for the TDA1541A. The TDA1541A pin 27 also provides different mode selections. I have set the CS8412 interface to output a time multiplexed signal, two's complement, MSB first. Pin 27 of the TDA1541A is connected to +5V to match these settings. The TDA1541A is also not a problem.
The sony format uses 48 cycles instead of 64 among other differences, so the sample rate is still 44.1 KHz but only 24 bits for each channel are transmitted instead of 32. That's the reason why BCK is running at 2.1168 MHz (48 * 44.1 KHz). This is a problem since the octal D-I DAC needs 64 cycles for the timing chain to function correctly. But there must be a way to convert formats if necessary.
cheers,
John
Hi John,
You've put my mind at rest. Thankyou.
It was because you specified an although not exact but similar frequency as too low that I supposed you meant a frequency several times more (not just a few hundred kHz).
I've ordered the bits to do RS422 fed by the I2S to an external DAC box. I'll start simply with the DAC (this is a voyage of discovery for me) and work up from there.
Looking forward to it.
Cheers,
Phil
You've put my mind at rest. Thankyou.
It was because you specified an although not exact but similar frequency as too low that I supposed you meant a frequency several times more (not just a few hundred kHz).
I've ordered the bits to do RS422 fed by the I2S to an external DAC box. I'll start simply with the DAC (this is a voyage of discovery for me) and work up from there.
Looking forward to it.
Cheers,
Phil
Re: Listening session
I have 8 TDA1541A (11118 HSH8910 2) I would be glad to start working on the project . I will use saa7310 from a Philips CD player ( it works pretty good in I2S nos ) . Then it would be nice to have a design to carry out the signals from the player in a good way ... maybe balanced .
The oscillator choice : I work with TENT Xo from years and it is an " non replaciable " tool for me in every kind of machine , DVD CD , Video and sound cards . I care less if there is jitter at the end of the chain , but at the start I want a very good oscillator. Alwayis good improvements and correct sense of perspective
with it . Some ECC83 are around here so they just wait to be telsted lol
But the Pass D1 output stage is another nice possibility for those ( like me) who enjoy the lack of feedback in a nice and simple circuit .
-ecdesigns- said:
I have 8 TDA1541A (11118 HSH8910 2) I would be glad to start working on the project . I will use saa7310 from a Philips CD player ( it works pretty good in I2S nos ) . Then it would be nice to have a design to carry out the signals from the player in a good way ... maybe balanced .
The oscillator choice : I work with TENT Xo from years and it is an " non replaciable " tool for me in every kind of machine , DVD CD , Video and sound cards . I care less if there is jitter at the end of the chain , but at the start I want a very good oscillator. Alwayis good improvements and correct sense of perspective
with it . Some ECC83 are around here so they just wait to be telsted lol
But the Pass D1 output stage is another nice possibility for those ( like me) who enjoy the lack of feedback in a nice and simple circuit .
-ecdesigns- said:Hi rfbrw,
thanks for your reply [post#461]
See you had a good laugh
So my question is, can the Sony format be converted to the Philips format without adding new problems like addittional jitter or synchronization problems. Are there schematics for this? or can a sample rate converter be used? I need 64 cycles for the direct interpolation timing
The Sony format can be converted to the I2S format with a shift register. I2S doesn't care about BCK so long as there is at least one BCK cycle per bit. 64 cycles is more an artifact of a digital receiver in master mode. A poster by the name of cm put forward such a circuit a while ago.
So you have three choices. Change the shift register taps to suit the convert to the I2S format and change the taps to suit 48Fs(easy). Build the logic needed to move data from the 48Fs domain to the 64Fs domain(resource intensive). Use SPDIF. Left out the use of an ASRC as I would thought it defeat the object of the excercise.
Rfbrw, dear, if the term "audiophile" was ment to be pejorative, I , for instance, would prefer to be called "amateur DIYer" or "melomaniac".
We are talking about I2S (just one format amongst many) because it is supported by the DAC chip in question and also by many nice transports as my very good and affordable CDPRO-2M, so a more direct signal path can be achieved.
My non-technically educated/intuitive mind thinks that the fewer the chips involved in signal path the better, no matter how good they measure. Am I wrong?
That is, to fully obtain the best sound from this clever filterless DAC.
Ecdesigns, what if you invested on a I2S transport?
(Joko sais SPDIF is flawed and I believe him)
Best wishes,
M
maxlorenz said:
Rfbrw, dear, if the term "audiophile" was ment to be pejorative, I , for instance, would prefer to be called "amateur DIYer" or "melomaniac".
I wouldn't say pejorative. To use an analogy. In order to get from A to B, there are two approaches. One can memorise a favoured route or one can learn to read a map. Audiophiles tend to dogmatically stick to one route and invariably crash and burn when faced with a diversion.
Have you tried the D-I dac with the CD Pro2? Let me know if it works, you might be in for an inpleasant surprise.
There is a lot more to electronics than counting chips. Assuming that is, you actually need more chips.
Ecdesigns, what if you invested on a I2S transport?
Tell me, M. If you bought a loaf of uncut bread and the slices you cut wouldn't fit in the toaster, would you buy a bigger toaster or cut thinner slices ?
Dear rfbrw,
Thanks for your interesting reply.
As I said, music lover is a nicer term than audiophile. Anyway, I am trully wanting to learn how to read maps and that is the reason to be in this forum in the first place. I can't afford to go back to school to study EE and you will agree that digital maybe the hardest part of it.
I don't consider myself to be dogmatic
No. Have you?
Yes. With my level of skills it is a strong possibility. That is why I want to gently push ecdesigns to it.
If you know of any reason why this approach would not work we will be all ears!
Thanks for your ever enlighting (though criptic) posts
Best regards.
M
Thanks for your interesting reply.
As I said, music lover is a nicer term than audiophile. Anyway, I am trully wanting to learn how to read maps and that is the reason to be in this forum in the first place. I can't afford to go back to school to study EE and you will agree that digital maybe the hardest part of it.
I don't consider myself to be dogmatic
No. Have you?
Yes. With my level of skills it is a strong possibility. That is why I want to gently push ecdesigns to it.
If you know of any reason why this approach would not work we will be all ears!
Thanks for your ever enlighting (though criptic) posts
Best regards.
M
Hi Rfbrw,
i tried TDA1541, TDA1545, TDA1543 (1x, 4x) NOS DAC using direct from tansport its I2S.
To my ears CS8412 receiver quite degrade sound compare to direct connection from I2S.
It is also easier to make one box cd player - less cables, shorter connections... better quality sound.
I like Ecdesigns way of doing: measuring then listening.
just my opinion.
regards, Bostjan
i tried TDA1541, TDA1545, TDA1543 (1x, 4x) NOS DAC using direct from tansport its I2S.
To my ears CS8412 receiver quite degrade sound compare to direct connection from I2S.
It is also easier to make one box cd player - less cables, shorter connections... better quality sound.
I like Ecdesigns way of doing: measuring then listening.
just my opinion.
regards, Bostjan
As a been pointed out, many of you are non-technical people. Now if I was was venturing into an unfamiliar arena, I would take the time to read at least the most relevant information available.
This, it would seem, is not the way of the audiophile. Why, in this very forum one of audiophiledoms most exalted gurus has dismissed the idea of reading datasheets as unnecessary as it is all about the ying and the yang of the design, concepts alien to ee types. OK so I may be over-egging the pudding a little but you get the drift.
Back to the CD Pro2. No, I haven't connected one to the D-I dac, but I have thought of buying one at some point and as such have read a lot of the available literature. True, the CD Pro2 has an I2S option but BCK is 48Fs.
This, it would seem, is not the way of the audiophile. Why, in this very forum one of audiophiledoms most exalted gurus has dismissed the idea of reading datasheets as unnecessary as it is all about the ying and the yang of the design, concepts alien to ee types. OK so I may be over-egging the pudding a little but you get the drift.
Back to the CD Pro2. No, I haven't connected one to the D-I dac, but I have thought of buying one at some point and as such have read a lot of the available literature. True, the CD Pro2 has an I2S option but BCK is 48Fs.
