SPDIF
Hi John,
I don't consider your post tread "pollution"
I have thought about using balanced signal from the player transporting it by two coax cables driving both inputs of the comparator. But after all I totally stopped working with the interface as I am now using the "one box solution".
Hi John,
I don't consider your post tread "pollution"
I have thought about using balanced signal from the player transporting it by two coax cables driving both inputs of the comparator. But after all I totally stopped working with the interface as I am now using the "one box solution".
On the subject of pulse transformers and how to obtain them, I've been happy with Scientific Conversions. They are located in Petaluma, CA which is local to me, and they will ship small quantities, although I've never ordered fewer than five. Their prices are also quite low, and they have both surface mount and through-hole, although the surface mounts are better.
Aes Ebu
"Doesn't anyone use balanced AES /EBU format with XLR's (at 5v P-P)?"
I have designed with this interface also. The potential for reflections is even greater than with an SPDIF with RCA connectors. XLR connectors are an incredable mismatch at RF frequencies and ever more difficult to design impedance nulling circuits for. They have a longer propagation delay and look more inductive. There is even more variation of these parameters for connectors from different manufactures than for RCA connectors.
Pulse transformer design become even more difficult since the higher impedance requires a larger primary inductance. This requires more turns for a given core type which leads to increased leakage inductance and reduction in bandwidth. The signal level also make saturation and distortion more of an issue.
Balanced cables consist of a twisted pair inside of a coax shield. Optimal ternination requires consideration of both the differential characteristic impepedance of the twisted wire pair and the common mode inpedance of the twisted pair in relation to the shield.
The much higher drive level also has the potential for more common mode noise to get from the transport to the DACs analog circuits. The higher drive levels do allow for more attenuation though. My premium AES cable use two 75 ohm coax
(one for each phase). Impedance matching was achieved with a resistor pad. The attenuation from this pad also helped to minimize signal reflections from the XLR connectors. These cables used extensive mechanical damping and a 40 volt bias voltage between the AC coupled shields and the center conductor. The only real advantage to the AES EBU balanced digital interface was that the Crystal input receiver was drive diffentially and at higher levels than the usual SPDIF implemetation. The interface circuit described above achieves the same benefit for driving the Crystal differentially and with much less tradeoffs the the usual AES/EBU balanced digital cable interface. I hated building the balanced cables and a suprising number of people paid $650 dollars retail for them.I think this cable is the hardest cable of any cables to get to sound good. It took over a year and a half for me to become satisfied with the design. It sounded very good but required way too much effort to get to where I was happy with the design.
"On the subject of pulse transformers and how to obtain them, I've been happy with Scientific Conversions"
I have evaluted them and found them to have the same design trade offs and simular measurents the Schott units and were considerably more expensive.
"Doesn't anyone use balanced AES /EBU format with XLR's (at 5v P-P)?"
I have designed with this interface also. The potential for reflections is even greater than with an SPDIF with RCA connectors. XLR connectors are an incredable mismatch at RF frequencies and ever more difficult to design impedance nulling circuits for. They have a longer propagation delay and look more inductive. There is even more variation of these parameters for connectors from different manufactures than for RCA connectors.
Pulse transformer design become even more difficult since the higher impedance requires a larger primary inductance. This requires more turns for a given core type which leads to increased leakage inductance and reduction in bandwidth. The signal level also make saturation and distortion more of an issue.
Balanced cables consist of a twisted pair inside of a coax shield. Optimal ternination requires consideration of both the differential characteristic impepedance of the twisted wire pair and the common mode inpedance of the twisted pair in relation to the shield.
The much higher drive level also has the potential for more common mode noise to get from the transport to the DACs analog circuits. The higher drive levels do allow for more attenuation though. My premium AES cable use two 75 ohm coax
(one for each phase). Impedance matching was achieved with a resistor pad. The attenuation from this pad also helped to minimize signal reflections from the XLR connectors. These cables used extensive mechanical damping and a 40 volt bias voltage between the AC coupled shields and the center conductor. The only real advantage to the AES EBU balanced digital interface was that the Crystal input receiver was drive diffentially and at higher levels than the usual SPDIF implemetation. The interface circuit described above achieves the same benefit for driving the Crystal differentially and with much less tradeoffs the the usual AES/EBU balanced digital cable interface. I hated building the balanced cables and a suprising number of people paid $650 dollars retail for them.I think this cable is the hardest cable of any cables to get to sound good. It took over a year and a half for me to become satisfied with the design. It sounded very good but required way too much effort to get to where I was happy with the design.
"On the subject of pulse transformers and how to obtain them, I've been happy with Scientific Conversions"
I have evaluted them and found them to have the same design trade offs and simular measurents the Schott units and were considerably more expensive.
How did I miss this one???
Elso:
I never told you it would not work, or not to do it. I just gave you some hints to make it work better. And it did, as I recall.
I saw no reason to upset your apple cart just because you chose a scheme other than what I would do.
OTOH, Tyler did ask me for my specific recommendations. So he received them.
dhaen:
What exactly is it that you don't agree with me on? I'm sure the gang would like to know. Start a new thread, if you prefer.
Jocko
Elso:
I never told you it would not work, or not to do it. I just gave you some hints to make it work better. And it did, as I recall.
I saw no reason to upset your apple cart just because you chose a scheme other than what I would do.
OTOH, Tyler did ask me for my specific recommendations. So he received them.
dhaen:
What exactly is it that you don't agree with me on? I'm sure the gang would like to know. Start a new thread, if you prefer.
Jocko
Needed: Schematic of Input Stage for CS8412
I am also interested in an improved input stage for the CS8412.
Does anyone have an operating schematic they would like to share?
I like the idea of the 74HCU04 with 74HC86 wired as a phase inverter, but how is this all achieved in practice? No site or thread seems to have the detail required to build this particular input stage.
I am also interested in an improved input stage for the CS8412.
Does anyone have an operating schematic they would like to share?
I like the idea of the 74HCU04 with 74HC86 wired as a phase inverter, but how is this all achieved in practice? No site or thread seems to have the detail required to build this particular input stage.
Any improvements to CS8412 Input stage?
Thanks rfbrw!
I gather you are referring to the diagram posted by Elso Kwak at the beginning of page two.
Modifications mentioned elsewhere in the thread include:
1) Delete C2 and C3.
2) Connect a 2k resistor at the node between U1:A and U1:B and bring this to ground.
Are these valid and do they sound better?
If I connect a pulse transformer across R3, leaving the input floating, is this okay? If not, where else should I tie the input of the transformer?
Thanks rfbrw!
I gather you are referring to the diagram posted by Elso Kwak at the beginning of page two.
Modifications mentioned elsewhere in the thread include:
1) Delete C2 and C3.
2) Connect a 2k resistor at the node between U1:A and U1:B and bring this to ground.
Are these valid and do they sound better?
If I connect a pulse transformer across R3, leaving the input floating, is this okay? If not, where else should I tie the input of the transformer?
Reread this thread. Please use the Search function for information on questions. You were probably not the first that asked that particular question before
http://www.diyaudio.com/forums/showthread.php?s=&threadid=13666&highlight=digital+transformer
http://www.diyaudio.com/forums/showthread.php?s=&threadid=13666&highlight=digital+transformer
The circuit is meant to replace the transformer. I cannot really comment on how it sounds as at the moment there is a lot junk after the CS8412 on the two boards that have the circuit but it does work with all the various sources I connect to it unlike my Mission DAC5.
When I built the circuit I did leave out the two caps at
RXN and RXP along with the 2K resistor and have had no problems.
When I built the circuit I did leave out the two caps at
RXN and RXP along with the 2K resistor and have had no problems.
tiroth said:
Hi Tiroth,
As I am using the circuit again:
You must ground the inverting input of the comparator.
Might be possible you have damaged the AD8561, I have blown up two. Hence the protection diodes in the new circuit.
Best directly connect the ouputs of the comparator to the CS8412.
This makes a huge difference sonically.
I am thrilled by the sound now!
http://www.diyaudio.com/forums/showthread.php?postid=227664#post227664
Ironically I did not get working Freds/Scott Nixons circuit completely succesfully
Elso,
I confess I gave up on the comparator a long time ago, but the noninverting input was connected to ground via 100 ohm just like in your schematic (the version posted at the very beginning of this thread...I know it has been updated several times now) I did attach to pot and vary from ground to Vcc...then gave up. Should it matter? I think the level was mV at most with 100 ohm.
I think the diodes are best omitted but then there is some danger to the comparator. It all depends on the goals I guess.
I have this thing Jocko drew on a napkin but it is far too complex for me to figure out. Maybe a few more years in this hobby...
Tyler
I confess I gave up on the comparator a long time ago, but the noninverting input was connected to ground via 100 ohm just like in your schematic (the version posted at the very beginning of this thread...I know it has been updated several times now) I did attach to pot and vary from ground to Vcc...then gave up. Should it matter? I think the level was mV at most with 100 ohm.
I think the diodes are best omitted but then there is some danger to the comparator. It all depends on the goals I guess.
I have this thing Jocko drew on a napkin but it is far too complex for me to figure out. Maybe a few more years in this hobby...
Tyler
Elso,
I confess I gave up on the comparator a long time ago, but the noninverting input was connected to ground via 100 ohm just like in your schematic (the version posted at the very beginning of this thread...I know it has been updated several times now) I did attach to pot and vary from ground to Vcc...then gave up. Should it matter? I think the level was mV at most with 100 ohm.
I think the diodes are best omitted but then there is some danger to the comparator. It all depends on the goals I guess.
I have this thing Jocko drew on a napkin but it is far too complex for me to figure out. Maybe a few more years in this hobby...
Did you try damping the input? RC or LC series resonant circuit? I needed one to get it to lock with "75 ohm" RCA patches and/or rat's nest wiring. Probably works fine if everything is true 75-ohm?
Tyler
I confess I gave up on the comparator a long time ago, but the noninverting input was connected to ground via 100 ohm just like in your schematic (the version posted at the very beginning of this thread...I know it has been updated several times now) I did attach to pot and vary from ground to Vcc...then gave up. Should it matter? I think the level was mV at most with 100 ohm.
I think the diodes are best omitted but then there is some danger to the comparator. It all depends on the goals I guess.
I have this thing Jocko drew on a napkin but it is far too complex for me to figure out. Maybe a few more years in this hobby...
Did you try damping the input? RC or LC series resonant circuit? I needed one to get it to lock with "75 ohm" RCA patches and/or rat's nest wiring. Probably works fine if everything is true 75-ohm?
Tyler
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.