I have a DAC that uses two 10nf caps for SPDIF coupling. What should I use for those - I tried so far WIMA FKP2 - too dark sounding and WIMA MKS02, which seems to be too harsh and 10nf ceramics which are 'dirty sounding'. Guess MKP or MKTs (low dielectric loss caps/good HF & pulse behaviour) would fit the bill better?
Well, part of it depends where they are in the circuit.
If they connect directly to the input, a lot of the sonic characteristics will be because of reflections.
I would expect ceramics to not sound good, although their lower inductance would be a benefit.
Jocko
If they connect directly to the input, a lot of the sonic characteristics will be because of reflections.
I would expect ceramics to not sound good, although their lower inductance would be a benefit.
Jocko
I had about the same and I just removed them. Since the source is a good device, I don't need them anymore.
Do you really need them? As far as I know the perfect coupling caps are no caps 😉
Do you really need them? As far as I know the perfect coupling caps are no caps 😉
Here's the schematic right out of the datasheet.
@roibm: are you sure they are safe to remove? The transport device is transformer coupled.
@roibm: are you sure they are safe to remove? The transport device is transformer coupled.
Size matters and smaller is often better!
Use a good metalized polypropelyne cap (MKP) with the lowest voltage you can get to keep the cap physical small. You can go to higher values and 0.047uF would be fine. I use some of the ERO 100V MKP caps to AC couple 75 load like the input of pulse transformers. Always put a cap in front of a pulse transform. I have heard a few 100 uA screw the sound of the pulse tranformer up big time when driving it with a video amp with a DC offset in the low millivolt range. NO DC CURRENT THROUGH THE PULSE TRANSFORMER! One of the Black Gate 0.1 uF nonpolar electrolytics also sounded extremly good in the digital interface.
Use a good metalized polypropelyne cap (MKP) with the lowest voltage you can get to keep the cap physical small. You can go to higher values and 0.047uF would be fine. I use some of the ERO 100V MKP caps to AC couple 75 load like the input of pulse transformers. Always put a cap in front of a pulse transform. I have heard a few 100 uA screw the sound of the pulse tranformer up big time when driving it with a video amp with a DC offset in the low millivolt range. NO DC CURRENT THROUGH THE PULSE TRANSFORMER! One of the Black Gate 0.1 uF nonpolar electrolytics also sounded extremly good in the digital interface.
Actually I meant that I have a transformer on the transport side and these grr... problematic caps on the DAC side.
So these ought to be ok? http://uk.farnell.com/jsp/endecaSearch/partDetail.jsp?SKU=3038543&N=401
So these ought to be ok? http://uk.farnell.com/jsp/endecaSearch/partDetail.jsp?SKU=3038543&N=401
If you are only going to use a transformer on one end, make it the RX end.
Kepping the outer conductor on th TX will reduce radiation, although there are ways around that. Besides, the RX side will perform better if you use a differential input, and the best way to do that is with a transformer.
Jocko
Kepping the outer conductor on th TX will reduce radiation, although there are ways around that. Besides, the RX side will perform better if you use a differential input, and the best way to do that is with a transformer.
Jocko
There nothing wrong with using a CKO5 style NPO, COG or even X7R ceramic material capicator in this application. There are many more areas of TX and RX interface to cause problems that those caps. However, make you use low inductance caps.
Jocko, he can elaborate on all the pit falls in the SPDIF interface. You might even get him to sell you a an impedance matched design the has been spectral tested. Then you won't have to get concerned about the caps and poor designs that plague 99.9% of all SPDIF interfaces.
My sermon is over for at least week, till some one asks the same questions again.
😉
Jocko, he can elaborate on all the pit falls in the SPDIF interface. You might even get him to sell you a an impedance matched design the has been spectral tested. Then you won't have to get concerned about the caps and poor designs that plague 99.9% of all SPDIF interfaces.
My sermon is over for at least week, till some one asks the same questions again.
😉
Well, in my case, I got no transport, it comes from the computer and I trust it. If I'll get problems, then I'll learn a lesson too 😉lucpes said:
Thanks, but I'm sure to mess up things if I do that, as I need to do more reading on this subject 🙄Jocko Homo said:
jewilson said:
Ceramic caps caused wreck havoc on the SPDIF signal (sonicly) at least in my case so I won't try it again. I know there are very knowledgeable people on this forum, that's why I posted here, and sorry for opening the same can of worms again
What soundcard are you using? I've seen some cards (M-Audio) which are transformer coupled & have a 10nf ceramic SMD in series with the SPDIF out which has to go.roibm said:
M-Audio Audiophile 2496.lucpes said:
Didn't check the spdif out path. I'll do it next time I open the machine.
The fact are that the ceramic caps that I mentioned or made for high frequency signals, it obvious by you statement that you don’t a have clue. The majority of high frequency oscillator phase lock loops use either Ceramic or Mica capacitor. There is a big difference in running a low frequency analog signal through a ceramic vrs a good PP PS film cap, but Bud we are not talking about analog signal are we.
Sorry I waisted your time, I did not realize by your statements that your were an ex-pert! 🙄
_IF_ 10nF is indeed "the" value, I'd use 10nF C0G SMT caps.
You can get 10nF/50V/C0G in a 1210 SMT package, which gives
you about as ideal a cap (good DF, DA, minimal parasitics) as
you can get for this sort of application.
You can get 10nF/50V/C0G in a 1210 SMT package, which gives
you about as ideal a cap (good DF, DA, minimal parasitics) as
you can get for this sort of application.
jewilson said:
Correct, I have almost no scientific knowledge of what I'm talking about and having to rely solely on ears to discern if a certain component is good for a specific application can be sometimes deceiving 🙂 But I have no chance to get CKO5 ceramic caps in my neck of the woods and silvered Mica would be a tad bit larger than the 5mm pitch I need for 10nf so it seems that I'm stuck to low inductance/low dielectric loss/good hf behaviour film caps instead or no cap at all.
Maybe this would be good for my application if I bend the leads a bit?
http://uk.farnell.com/jsp/endecaSea...ica&Nty=1&showImages=true&N=401&Ntk=gensearch
lucpes
Sure, two of these in parallel would work fine. I don't believe that being 600pf off is a big deal however you could get a 500pf to go with it.
This Farnell in One this company has the same look to their web site as US Newark Electronics http://www.newark.com they must be related. I see that Farnell sales NPO ceramic's. but the Mica is a little better.
Sure, two of these in parallel would work fine. I don't believe that being 600pf off is a big deal however you could get a 500pf to go with it.
This Farnell in One this company has the same look to their web site as US Newark Electronics http://www.newark.com they must be related. I see that Farnell sales NPO ceramic's. but the Mica is a little better.
roibm,
You must have hearing more sensitive than a dog or cat. I highly doubt simple tiny capacitors will make any difference in audio quality of an S/PDIF signal. I have an i2s to TTL to optical to RS485 to coax and into jwb's great DAC. It has a SC982-04 pulse transformer and a 0.1uF SMT cap on the input to block DC and couple the signal into the pulse transformer. The sound is excellent. It will sound the same if the cap was or wasn't there, if the pulse transformer was or wasn't there, it will make no difference. I advise you keep the capacitors and use a pulse transformer. That will keep everything safe. Use 10Mbps RS422/485 tranceiver IC's like the MAX3443 from Maxim, SP1485 from Sipex, SN75176 from TI, or the DS3695 from National Semi for your S/PDIF transmission and input the TTL signal right into the S/PDIF receiver. You will be satisfied. My RS485 to coax circuit uses a CMOS inverter, some resistors and a 1:1 pulse transformer. The output of the circuit gets connected to my DAC. I have transformer coupling on both sides, just for the hell of it. Even though the signal is going through all these conversions, it still sounds perfect.
You must have hearing more sensitive than a dog or cat. I highly doubt simple tiny capacitors will make any difference in audio quality of an S/PDIF signal. I have an i2s to TTL to optical to RS485 to coax and into jwb's great DAC. It has a SC982-04 pulse transformer and a 0.1uF SMT cap on the input to block DC and couple the signal into the pulse transformer. The sound is excellent. It will sound the same if the cap was or wasn't there, if the pulse transformer was or wasn't there, it will make no difference. I advise you keep the capacitors and use a pulse transformer. That will keep everything safe. Use 10Mbps RS422/485 tranceiver IC's like the MAX3443 from Maxim, SP1485 from Sipex, SN75176 from TI, or the DS3695 from National Semi for your S/PDIF transmission and input the TTL signal right into the S/PDIF receiver. You will be satisfied. My RS485 to coax circuit uses a CMOS inverter, some resistors and a 1:1 pulse transformer. The output of the circuit gets connected to my DAC. I have transformer coupling on both sides, just for the hell of it. Even though the signal is going through all these conversions, it still sounds perfect.
I never said I heard a difference when I removed the cap. But then again, there was a pause of several days between removing the cap and serious listening again.emuman100 said:
The cap was a really bad 0.1uF electrolytic, so I guess the change couldn't have been for worst.
As about your setup over there... Why? How long is the cable? Any numbers to show the difference between what you got and the standard spdif?
I figured as much.......
" The fact are that the ceramic caps that I mentioned or made for high frequency signals, it obvious by you statement that you don't a have clue."
I guess I don't have a clue either.......... BTW, find me a decent 0.1uF
(or even 0.01uF) ceramic or Mica cap.
All I can do is recommend people listen. It is not that hard to find a decent MKP cap from DigiKey of Mouser. The signal is digital statement is not going to fly very far since there is a whole engineering discipline devoted to optimizing the digital interface between ICs. Its called Signal Integrity. Tell them Howard Johnson said so.
http://www.sigcon.com/
" The fact are that the ceramic caps that I mentioned or made for high frequency signals, it obvious by you statement that you don't a have clue."
I guess I don't have a clue either.......... BTW, find me a decent 0.1uF
(or even 0.01uF) ceramic or Mica cap.
All I can do is recommend people listen. It is not that hard to find a decent MKP cap from DigiKey of Mouser. The signal is digital statement is not going to fly very far since there is a whole engineering discipline devoted to optimizing the digital interface between ICs. Its called Signal Integrity. Tell them Howard Johnson said so.http://www.sigcon.com/
roibm,
Only reason why is because I got the idea from a web site and it sounded pretty cool. With my tests, I've gotten 100ft with no noise or audio degridation. I don't have any numbers though. With these RS485 transceivers, you can use cable lengths of thousands of feet.
Only reason why is because I got the idea from a web site and it sounded pretty cool. With my tests, I've gotten 100ft with no noise or audio degridation. I don't have any numbers though. With these RS485 transceivers, you can use cable lengths of thousands of feet.
- Status
- Not open for further replies.