Atom Feed - Twisted Pear Audio Support - Topic:DAC board inside a CDP (I2S - 16/44.1) - 20Twisted Pear Audio Support - Atom Feedurn:twistedpearaudio-com:AtomFeed:TwistedPearAudioSupport:Topic:DACboardinsideaCDP(I2S-16/44.1)-20:1Copyright 2024 Twisted Pear Audio Support2024-03-28T22:28:58Zhttp://www.twistedpearaudio.com/forum/Images/YAFLogo.pngForum Adminhttp://www.twistedpearaudio.comfeedback@twistedpearaudio.comGSUMhttp://www.twistedpearaudio.com/forum/profile/444-GSUMGSUMhttp://www.twistedpearaudio.com/forum/profile/444-GSUMGSUMhttp://www.twistedpearaudio.com/forum/profile/444-GSUMRuss Whitehttp://www.twistedpearaudio.com/forum/profile/9-Russ-WhiteGSUMhttp://www.twistedpearaudio.com/forum/profile/444-GSUMRuss Whitehttp://www.twistedpearaudio.com/forum/profile/9-Russ-WhiteGSUMhttp://www.twistedpearaudio.com/forum/profile/444-GSUMYetAnotherForum.NETurn:twistedpearaudio-com:ftPosts:st1:meid1249:1DAC board inside a CDP (I2S - 16/44.1)<table class="content postContainer_Alt" width="100%"><tr><td>Just to tidy this up, page 24 of the NPC SM5872BS DAC chip data sheet has example interfaces with decoders:<br /><br />CKO (system clock) = 16.9344MHz<br />LRCI (left/right clock) = 44.1kHz<br />BCKI (bit clock) = 2.1168MHz<br /><br />2116.8 / 44.1 = 48<br /><br />Therefore bit clock is 48fs. Not the required 64fs.</td></tr></table>2008-07-21T15:25:26-07:002008-07-21T15:25:26-07:00GSUM<table class="content postContainer_Alt" width="100%"><tr><td>Just to tidy this up, page 24 of the NPC SM5872BS DAC chip data sheet has example interfaces with decoders:<br /><br />CKO (system clock) = 16.9344MHz<br />LRCI (left/right clock) = 44.1kHz<br />BCKI (bit clock) = 2.1168MHz<br /><br />2116.8 / 44.1 = 48<br /><br />Therefore bit clock is 48fs. Not the required 64fs.</td></tr></table>urn:twistedpearaudio-com:ftPosts:st1:meid1248:1DAC board inside a CDP (I2S - 16/44.1)<table class="content postContainer" width="100%"><tr><td>Cool, thanks for all the info!<br /><br />The S/PDIF route sounds like a winner then, especially as I may move over to an SB3 as a transport at some stage.<br /><br />Cheers,<br />Glenn</td></tr></table>2008-07-21T06:32:22-07:002008-07-21T06:32:22-07:00GSUM<table class="content postContainer" width="100%"><tr><td>Cool, thanks for all the info!<br /><br />The S/PDIF route sounds like a winner then, especially as I may move over to an SB3 as a transport at some stage.<br /><br />Cheers,<br />Glenn</td></tr></table>urn:twistedpearaudio-com:ftPosts:st1:meid1247:1DAC board inside a CDP (I2S - 16/44.1)<table class="content postContainer_Alt" width="100%"><tr><td>Hi Glenn,<br /><br />I was surprised too. I am not sure why many chip designers went the 64fs bitclock only route. but I have a good guess, and that is because almost all of them are designed to support more modern multi-channel DVD/Blueray type playback, not so much CDP.<br /><br />The Buffalo (well the Sabre chip) also requires a bit clock at 64fs, if you want to use I2S in from your player you are likely going to need the Metronome no matter which of the DACs in our lineup you choose.<br /><br />I would not hesitate to use the SPDIF directly into the Buffalo. It sounds fantastic.<br /><br />Cheers!</td></tr></table>2008-07-21T06:17:09-07:002008-07-21T06:17:09-07:00Russ White<table class="content postContainer_Alt" width="100%"><tr><td>Hi Glenn,<br /><br />I was surprised too. I am not sure why many chip designers went the 64fs bitclock only route. but I have a good guess, and that is because almost all of them are designed to support more modern multi-channel DVD/Blueray type playback, not so much CDP.<br /><br />The Buffalo (well the Sabre chip) also requires a bit clock at 64fs, if you want to use I2S in from your player you are likely going to need the Metronome no matter which of the DACs in our lineup you choose.<br /><br />I would not hesitate to use the SPDIF directly into the Buffalo. It sounds fantastic.<br /><br />Cheers!</td></tr></table>urn:twistedpearaudio-com:ftPosts:st1:meid1246:1DAC board inside a CDP (I2S - 16/44.1)<table class="content postContainer" width="100%"><tr><td>Thanks for such a speedy and helpful response Russ!<br /><br />Ok, the one missing piece of information then is the bit clock.<br /><br />Looking at the built-in DAC, I think it's 32fs unless I'm misunderstanding something.<br />The block diagram shows the input to first block (IIR filter) as 16bits@fs and the input to the PWM modulator (the last block) as 32fs. <br /><br />The clock frequency of the player is 384fs (16.9344MHz). This is at the DAC itself and is buffered by an internal gate on the DAC and fed on to the decoder IC (SAA7372) still at the same frequency. If I eventually remove the internal DAC I would have to clock the decoder directly.<br /><br />Does this back up your theory that for correct operation I would need:<br /><br />I2S --> Metronomone -> COD/OPUS<br />or<br />I2S or S/PDIF --> BUFFALO<br /><br />I am a little surprised that I can feed the original I2S signal into a TDA1543 but I can't feed it straight to an all-singing, all-dancing modern DAC chip!<br /><br />Thanks,<br />Glenn</td></tr></table>2008-07-20T23:49:05-07:002008-07-20T23:49:05-07:00GSUM<table class="content postContainer" width="100%"><tr><td>Thanks for such a speedy and helpful response Russ!<br /><br />Ok, the one missing piece of information then is the bit clock.<br /><br />Looking at the built-in DAC, I think it's 32fs unless I'm misunderstanding something.<br />The block diagram shows the input to first block (IIR filter) as 16bits@fs and the input to the PWM modulator (the last block) as 32fs. <br /><br />The clock frequency of the player is 384fs (16.9344MHz). This is at the DAC itself and is buffered by an internal gate on the DAC and fed on to the decoder IC (SAA7372) still at the same frequency. If I eventually remove the internal DAC I would have to clock the decoder directly.<br /><br />Does this back up your theory that for correct operation I would need:<br /><br />I2S --> Metronomone -> COD/OPUS<br />or<br />I2S or S/PDIF --> BUFFALO<br /><br />I am a little surprised that I can feed the original I2S signal into a TDA1543 but I can't feed it straight to an all-singing, all-dancing modern DAC chip!<br /><br />Thanks,<br />Glenn</td></tr></table>urn:twistedpearaudio-com:ftPosts:st1:meid1245:1DAC board inside a CDP (I2S - 16/44.1)<table class="content postContainer_Alt" width="100%"><tr><td>Hello Glenn,<br /><br />That sounds like a fun project.<br /><br />I don't know enough about the devices in your CDP to give you concrete answers, but what I can do is tell you what the DACs can support. :)<br /><br />Both the COD and the Opus and be configured for 16/44.1 easily, but you need to be aware that almost all modern DACs(actually all I know of) require a 64fs bitclock. Many older (and some not so old) CDPs have a 32fs or 48fs bitclock. You can not feed this directly to the Opus,COD or even Buffalo. You would need to reclock it (Metronome does this nicely).<br /><br />So assuming your CDP has a bitclock at 64fs then you can take I2S directly, if not, then I would just add the metonome set either for 24/96 or 24/192, your choice.<br /><br />The Buffalo treats the two (SPDIF and I2S) exactly the same once they are initially clocked in (then later asynchronously re-clocked internally). From an absolute performance perspective I2S will tend to be better simply because you can leave the DPLL bandwidth at the lowest possible setting (which helps reduce jitter) without any fear that your SPDIF will not stay locked, since you would not be using SPDIF. :) I actually have found that I can leave the DPPL at the lowest setting even for SPDIF, but others have had to move it up to "MEDIUM" to keep a good lock because of poor SPDIF signals.<br /><br />Cheers!<br />Russ</td></tr></table>2008-07-20T15:59:47-07:002008-07-20T15:59:47-07:00Russ White<table class="content postContainer_Alt" width="100%"><tr><td>Hello Glenn,<br /><br />That sounds like a fun project.<br /><br />I don't know enough about the devices in your CDP to give you concrete answers, but what I can do is tell you what the DACs can support. :)<br /><br />Both the COD and the Opus and be configured for 16/44.1 easily, but you need to be aware that almost all modern DACs(actually all I know of) require a 64fs bitclock. Many older (and some not so old) CDPs have a 32fs or 48fs bitclock. You can not feed this directly to the Opus,COD or even Buffalo. You would need to reclock it (Metronome does this nicely).<br /><br />So assuming your CDP has a bitclock at 64fs then you can take I2S directly, if not, then I would just add the metonome set either for 24/96 or 24/192, your choice.<br /><br />The Buffalo treats the two (SPDIF and I2S) exactly the same once they are initially clocked in (then later asynchronously re-clocked internally). From an absolute performance perspective I2S will tend to be better simply because you can leave the DPLL bandwidth at the lowest possible setting (which helps reduce jitter) without any fear that your SPDIF will not stay locked, since you would not be using SPDIF. :) I actually have found that I can leave the DPPL at the lowest setting even for SPDIF, but others have had to move it up to "MEDIUM" to keep a good lock because of poor SPDIF signals.<br /><br />Cheers!<br />Russ</td></tr></table>urn:twistedpearaudio-com:ftPosts:st1:meid1244:1DAC board inside a CDP (I2S - 16/44.1)<table class="content postContainer" width="100%"><tr><td>Hi all,<br /><br />I have two old Marantz players; a Marantz CD67SE (with Tent XO2+XO supply + LC Audio Zapfilter2) and a stock CD63.<br /><br />Whilst the DAC in these players is well regarded, I fancy playing around with some modern alternatives.<br /><br />The DAC (NPC SM5872BS) has the 8x oversampling digital filter in it, so the I2S leaves the decoder as unmolested 16bit/44.1KHz (i.e., the decoder's oversampling function is not enabled).<br /><br />Can the OPUS board (or COD) be configured to accept this directly? i.e., without being upsampled to 96k/192k and having another 8 bits bolted on? Just plain old I2S/16bit/44.1kHz.<br /><br />The I2S flavour output by the decoder is standard Philips, and I tried tapping it off the resistors going into the existing DAC into a TDA1543 and it worked absolutely fine (didn't even sound too bad).<br /><br />Presumably a Metronome would take my 16/44.1 and make it 24/192 if I wanted.<br /><br />I could feed the DAC output to the Zapfilter, which can be configured for voltage out or current out DACs (i.e., OPUS or COD). My present configuration has a third-order passive filter on each DAC output, due the unfiltered nature of the present DACs raw PWM output. These filters wouldn't be necessary with modern DACs.<br /><br />Also, if I was to wait for a Buffalo, would there be any benefit to using the I2S lines over the S/PDIF line from the decoder? It sounds as though maybe not.<br /><br />Thanks,<br />Glenn</td></tr></table>2008-07-20T15:12:18-07:002008-07-20T15:12:18-07:00GSUM<table class="content postContainer" width="100%"><tr><td>Hi all,<br /><br />I have two old Marantz players; a Marantz CD67SE (with Tent XO2+XO supply + LC Audio Zapfilter2) and a stock CD63.<br /><br />Whilst the DAC in these players is well regarded, I fancy playing around with some modern alternatives.<br /><br />The DAC (NPC SM5872BS) has the 8x oversampling digital filter in it, so the I2S leaves the decoder as unmolested 16bit/44.1KHz (i.e., the decoder's oversampling function is not enabled).<br /><br />Can the OPUS board (or COD) be configured to accept this directly? i.e., without being upsampled to 96k/192k and having another 8 bits bolted on? Just plain old I2S/16bit/44.1kHz.<br /><br />The I2S flavour output by the decoder is standard Philips, and I tried tapping it off the resistors going into the existing DAC into a TDA1543 and it worked absolutely fine (didn't even sound too bad).<br /><br />Presumably a Metronome would take my 16/44.1 and make it 24/192 if I wanted.<br /><br />I could feed the DAC output to the Zapfilter, which can be configured for voltage out or current out DACs (i.e., OPUS or COD). My present configuration has a third-order passive filter on each DAC output, due the unfiltered nature of the present DACs raw PWM output. These filters wouldn't be necessary with modern DACs.<br /><br />Also, if I was to wait for a Buffalo, would there be any benefit to using the I2S lines over the S/PDIF line from the decoder? It sounds as though maybe not.<br /><br />Thanks,<br />Glenn</td></tr></table>