DCX2496 with 96kHz?
Hello out there and a happy new year.
Some days ago I downloaded audacity, a nice little freeware recorder software which supports the 24/96 format. From my modded DV-575A (578a) ( http://www.diyaudio.com/forums/showthread.php?s=&threadid=48393 ) I recorded a DVA-A from the Eagles with 24/96.
Playing back the recorded tracks via DCX2496 I didn’t get the quality I expected. Even worse it got if I connected the modded DV-575A directly to the DCX2496. In this configuration I got dropouts here and there. I opened the DCX-2496 and measured the recovered clock at pin 10 of the SRC (CS8420). The signal was very very bad with a high amount of jitter. Playing back the recorded tracks from the soundcard it was a bit better but not so much. I tried it with CDs (44.1kHz) and DVDs (48kHz) and there the recovered clock was absolute stable with no jitter at all. Playing a SACD with 88.2kHz it was worse than playing CDs or DVDs but better than DVD-A with 96kHz.
Going to the datasheet of the CS8420 I found out that the DCX2496 uses other values for the PLL filter as recommended in the datasheet. It also uses a different PCB layout as recommended. They recommend the caps and the resistor of the filter to be mounted directly aside the pins of the CS8420. Behringer mounted then on the opposite side of the PCB with long connection wires and some vias. The datasheet tells to avoid this in all cases!
My question is now. Who had already input a 96kHz digital signal into the DCX2496 and got the expected output quality?
Hello out there and a happy new year.
Some days ago I downloaded audacity, a nice little freeware recorder software which supports the 24/96 format. From my modded DV-575A (578a) ( http://www.diyaudio.com/forums/showthread.php?s=&threadid=48393 ) I recorded a DVA-A from the Eagles with 24/96.
Playing back the recorded tracks via DCX2496 I didn’t get the quality I expected. Even worse it got if I connected the modded DV-575A directly to the DCX2496. In this configuration I got dropouts here and there. I opened the DCX-2496 and measured the recovered clock at pin 10 of the SRC (CS8420). The signal was very very bad with a high amount of jitter. Playing back the recorded tracks from the soundcard it was a bit better but not so much. I tried it with CDs (44.1kHz) and DVDs (48kHz) and there the recovered clock was absolute stable with no jitter at all. Playing a SACD with 88.2kHz it was worse than playing CDs or DVDs but better than DVD-A with 96kHz.
Going to the datasheet of the CS8420 I found out that the DCX2496 uses other values for the PLL filter as recommended in the datasheet. It also uses a different PCB layout as recommended. They recommend the caps and the resistor of the filter to be mounted directly aside the pins of the CS8420. Behringer mounted then on the opposite side of the PCB with long connection wires and some vias. The datasheet tells to avoid this in all cases!
My question is now. Who had already input a 96kHz digital signal into the DCX2496 and got the expected output quality?
Hi, I cannot comment on that setup, but I had connected a philips 963 and a micromega stage 2 digitally with 75 ohm cable to the deq 2496. Both sounded better this way - no extensive blind testing done, just one track of my test cd - than routing the analog output from both through a bryston bp 20 and then to the deq. The output from the philips was 24/96 upsampled.
Right now i route everything through an src 2496 and thence to deq/dcx 2496 - very satisfactory sound.
Right now i route everything through an src 2496 and thence to deq/dcx 2496 - very satisfactory sound.
oehlrich, as far as my experiments show the DCX2496 uses CS8420 always so that the output I2S is clocked out always by the built in 24,576 MHz clock in 24/96 format.
I have installed a Tent XO2 to DCX and thus the I2S is clocked out from CS8420 according to this "master clock".
The scheme is better than recovering the clock from spdif, but is not flawless either. I have heard the effect as my own setup uses a clock feedback to Sony CD player.
I can listen to my system in two ways
a) a spdif coming from cd player as usual to DCX
b) spdif coming from cd player, but the cd player clock is formed from master clock inside DCX
The difference is there and is in favor of clock feedback scheme.
I have installed a Tent XO2 to DCX and thus the I2S is clocked out from CS8420 according to this "master clock".
The scheme is better than recovering the clock from spdif, but is not flawless either. I have heard the effect as my own setup uses a clock feedback to Sony CD player.
I can listen to my system in two ways
a) a spdif coming from cd player as usual to DCX
b) spdif coming from cd player, but the cd player clock is formed from master clock inside DCX
The difference is there and is in favor of clock feedback scheme.
Good idea, this option is also on my todo list and I guess it will make a big difference. Currently my problem ist not the output but the input of the CS8420. There is a PLL inside this chip which regelerates a 256*Fs clock from out of the incoming S/PDIF signal. This PLL did not run well im my dcx2496 submitting a 96kHz s/PDIF signal.ergo said:
ergo said:
I am planning of doing the same thing - ordering clock from Tent (based on Ergo's pdf instructions
), but I have some Questions - is it possible to use that clock as a master clock. My missunderstanding is: SRC and DEQ 2496 both acccept world clock, yet DCX 2496 doesn't. Is it because clock has to be where the A/D/A conversion takes place?If that is the case would it be possible to install BNC plug in the back of DCX and distribute that to DEQ and CD player?
AR2
Here is the link to my "work in progess" web page
Ergo's DIY site
I have so little time to make it that the info might rotten before I get done So I publish what I've got.
Ergo
Ergo's DIY site
I have so little time to make it that the info might rotten before I get done
So I publish what I've got.Ergo
comprehensive DCX mod
Hi guys!
My first post here. Been lurking for quite some time though, and it's about time I made one
I also posted this in the Yahoo group DCX 2496, but I'll probably reach more people here.
I'm working on the designs for a comprehensive mod of the DCX. I have two DCX's in my HT system, hooked in at exactly the wrong place (between preamp and poweramp, and no volume control after the DCX).
I want to modify both DCX's to take i2s directly in an all input channels. I'll mod a cheap DD/DTS receiver to get L/R/LS/RS/C/Sub serial i2s data out.
I'm thinking of using one of the Panasonix XR-xx for this, and take the signal right before the PWM output stage.
For the output of the Behringer, I will take the analog signal right of the DACs after adding series resistor and a polypropolene kondensator at each channel (4,7uF should be sufficient according to my calculations). I will feed the (total of 12 channels) to six motorized and IR-controlled stereo ALPS 10k log pots for the meantime. Might replace them with attuntators (how do I spell that?!) later on.
I've only looked at the input of the Behringer so far. I thought I might share what have found out to "test" the design with you guys.
Input channel A/B. The 8420 Sample Rate Converter is setup to run in hardware mode 1. This needs to be changed to hardware mode 2 (to support I2S instead of AES/EBU). This is achieved by connecting pin 20 to + instead of ground. AES/EBU connection will be removed from pin 4 and 5. Pin 4 is wired to + and pin 5 to ground to enable Serial I/O Format “2” (which is I2S). Serial data from the panasonic receiver will be wired to pin 12,13,14. Please let me know if I’m missing something or if I’m completely wrong.
Channel C is harder. No sample rate converter to hook in to. Now the ADC handles the output sample rate. The Shark expects the signal to be 24/96, so I guess the only thing to do is to add an other 8420 and set it up as above? Should do the trick, right? (Expensive little suckers, though).
If only one channel is used in the Shark. How do I flip the L/R in the I2s data stream in able to route the sub channel to one DCX and the center channel to the other? Can I just invert ILRCK (eg use a 74HC04) to achieve that?
If both channels are used, I’ll need to create a “dual mono” signal. Anyone has any ideas for that?
In any case, I’ll have to do something in the DD/DTS receiver in able to get four wires out of it. The Center/Sub stream must be “copied” somehow, or is it possible to feed two sources with the same wires of an i2s datastream? Maybe I should implement the “dual-mono” logic there. Ideas for components appreciated.
Does anyone know how far one can wire i2s signals without having to use a tranciever?
// Lyckman
Hi guys!
My first post here. Been lurking for quite some time though, and it's about time I made one
I also posted this in the Yahoo group DCX 2496, but I'll probably reach more people here.
I'm working on the designs for a comprehensive mod of the DCX. I have two DCX's in my HT system, hooked in at exactly the wrong place (between preamp and poweramp, and no volume control after the DCX).
I want to modify both DCX's to take i2s directly in an all input channels. I'll mod a cheap DD/DTS receiver to get L/R/LS/RS/C/Sub serial i2s data out.
I'm thinking of using one of the Panasonix XR-xx for this, and take the signal right before the PWM output stage.
For the output of the Behringer, I will take the analog signal right of the DACs after adding series resistor and a polypropolene kondensator at each channel (4,7uF should be sufficient according to my calculations). I will feed the (total of 12 channels) to six motorized and IR-controlled stereo ALPS 10k log pots for the meantime. Might replace them with attuntators (how do I spell that?!) later on.
I've only looked at the input of the Behringer so far. I thought I might share what have found out to "test" the design with you guys.
Input channel A/B. The 8420 Sample Rate Converter is setup to run in hardware mode 1. This needs to be changed to hardware mode 2 (to support I2S instead of AES/EBU). This is achieved by connecting pin 20 to + instead of ground. AES/EBU connection will be removed from pin 4 and 5. Pin 4 is wired to + and pin 5 to ground to enable Serial I/O Format “2” (which is I2S). Serial data from the panasonic receiver will be wired to pin 12,13,14. Please let me know if I’m missing something or if I’m completely wrong.
Channel C is harder. No sample rate converter to hook in to. Now the ADC handles the output sample rate. The Shark expects the signal to be 24/96, so I guess the only thing to do is to add an other 8420 and set it up as above? Should do the trick, right? (Expensive little suckers, though).
Facts:- All datasheets of DD/DTS decoders that I have looked at, has one “shared” i2s stream for sub and center channels.
- The analog input on DCX channel C is connected to both input L and R of the 5383 ADC in parallel. In conclusion, the output data stream wich feeds the shark has identical (hopefully) left and right channels.
Questions:- Does the behringer DSP software for some kinky reason expect a “dual-mono” I2S signal from channel C?
- If No, Does it only use data from one of the channels? (Would be the best for me, but why oh why did they have to connect the ADC in the way they did.. got me puzzled.)
(I don’t really expect anyone to know the answers to above questions )
If only one channel is used in the Shark. How do I flip the L/R in the I2s data stream in able to route the sub channel to one DCX and the center channel to the other? Can I just invert ILRCK (eg use a 74HC04) to achieve that?
If both channels are used, I’ll need to create a “dual mono” signal. Anyone has any ideas for that?
In any case, I’ll have to do something in the DD/DTS receiver in able to get four wires out of it. The Center/Sub stream must be “copied” somehow, or is it possible to feed two sources with the same wires of an i2s datastream? Maybe I should implement the “dual-mono” logic there. Ideas for components appreciated.
Does anyone know how far one can wire i2s signals without having to use a tranciever?
// Lyckman
Welcome to the forum, Lyckman. I wish everyone's first post was as good as yours!
As to your I2S question, the further they go, the more skew there will be between signals due to cable capacitance, and that will be translated into distortion. I've recently been investigating RS485 tranceivers (for my Behringer) and have been deeply unimpressed by the waveform attainable at the far end of a 3m piece of cable. And my tests were intended to check the characteristic impedance of the cable - not the quality of the tranceiver! I'm intending to use a current feedback video op-amp from the EL2030 family with a gain of 2 to drive the cable from the cable's characteristic impedance, then terminate the far end of the cable with the same impedance. That should ensure correct edges at the far end.
Don't forget that you can configure a Behringer to pick up only the A (or B) channel from an AES3 signal in order to use it as a six-channel crossover. You then need two Behringers and an AES3 distribution amplifier to feed them.
As to your I2S question, the further they go, the more skew there will be between signals due to cable capacitance, and that will be translated into distortion. I've recently been investigating RS485 tranceivers (for my Behringer) and have been deeply unimpressed by the waveform attainable at the far end of a 3m piece of cable. And my tests were intended to check the characteristic impedance of the cable - not the quality of the tranceiver! I'm intending to use a current feedback video op-amp from the EL2030 family with a gain of 2 to drive the cable from the cable's characteristic impedance, then terminate the far end of the cable with the same impedance. That should ensure correct edges at the far end.
Don't forget that you can configure a Behringer to pick up only the A (or B) channel from an AES3 signal in order to use it as a six-channel crossover. You then need two Behringers and an AES3 distribution amplifier to feed them.
Ok. 3m is way longer than I was aiming for. More like 30cm or so It would be nice if I could "be without" transievers.
I've Never heard of an AES3 distrubution amplifier, but by the name of it, it should most likely contain the circuitry I'll need for the i2s distribution. I'll do a little research on this. At least it is somewhere to start. Thanks.
// Lyckman
It would be nice if I could "be without" transievers.I've Never heard of an AES3 distrubution amplifier, but by the name of it, it should most likely contain the circuitry I'll need for the i2s distribution. I'll do a little research on this. At least it is somewhere to start. Thanks.
// Lyckman
I think you will like it. I have been using one in a 4way configuration for over 6 months now, but a couple of words of advice (which probably other members have already made);
i) You need to get the input level up, not only to avoid quantising noise, but to ensure that you get the stop-band rejection you expect from the crossover.
ii) I found that the auto delay feature on mine does not perform correctly. I can measure the impulse response of each driver/DCX channel and calculate the time of flight from the driver to the microphone, but in each channel the time delay is different -despite running the auto delay feature.
Anyway, if you have an MLS program like Soundeasy you can adjust the delays manually, to time-align drivers. Also it seems like the time delay is dependant upon the processing power of the DSP. chage the filter slope, them the delay changes as well.
Enjoy the Behringer, amazing value for money...
i) You need to get the input level up, not only to avoid quantising noise, but to ensure that you get the stop-band rejection you expect from the crossover.
ii) I found that the auto delay feature on mine does not perform correctly. I can measure the impulse response of each driver/DCX channel and calculate the time of flight from the driver to the microphone, but in each channel the time delay is different -despite running the auto delay feature.
Anyway, if you have an MLS program like Soundeasy you can adjust the delays manually, to time-align drivers. Also it seems like the time delay is dependant upon the processing power of the DSP. chage the filter slope, them the delay changes as well.
Enjoy the Behringer, amazing value for money...
cheap trick
hello
when removing the analog i/o board for checking the under side, i found the sound improved a lot after reassembling without having done anything at all.
my guess on that is, that the cabel from x1 to x13, carrying the audio signals don't like to be glued down to the metal, but benefit from having some 1-2 cm air isolation witch might lower capacitence to ground.
greetings
michael
hello
when removing the analog i/o board for checking the under side, i found the sound improved a lot after reassembling without having done anything at all.
my guess on that is, that the cabel from x1 to x13, carrying the audio signals don't like to be glued down to the metal, but benefit from having some 1-2 cm air isolation witch might lower capacitence to ground.
greetings
michael