Hi, everyone!
A few days ago I made a small discovery regarding blocking HF noise coming from PC.
I have a professional sound card Prodif-88 that I use as a source. Next in line is SPDIF to I2S converter with SN75LBC176D transceiver and a DIY DAC with synchronous reclocking.
I know that PC radiates a lot of HF noise. The converter has a transformer at the input. It means galvanic isolation with PC. Until recently I have been of an opinion that SPDIF transformer is quite sufficient for blocking HF noise but recently I found out that it’s not.
Let me tell you how I came to that conclusion.
My soundcard has AES/EBU digital outputs in balanced mode. Recently I found out that SN75LBC176D in the converter is connected in balanced mode whereby it receives SPDIF signal in balanced mode and then it converts it into unbalanced mode and feeds DIR9001 with the signal which transforms SPDIF into 3 signals but masterclock is not used because the DAC has a synchronous reclocking trigger and the soundcard is slaved to oscillator in the DAC.
I did not understand that my converter can understand balanced mode and because of that I have been using a resistor divider in order to convert balanced AES/EBU into unbalanced SPDIF.
A year ago I experimented with a ferrite clamp that I clamped on the SPDIF cable in order to reduce HF noise but I found out that after putting on a ferrite clamp the quality noticeably deteriorated, I heard some kind of distortions which sound as a very slow sound. The sound started irritating me a lot and I disconnected the ferrite clamp.
Please, take a note that I connected sound card with converter via coaxial cable and I put a ferrite clamp on this coaxial cable. This is very important.
Then a few months ago as soon as I found out that my converter understands balanced AES/EBU I tried to switch to a specially designed balanced AES/EBU cable and the quality improved significantly.
This was my first discovery but I want to discuss the second discovery!
The second discovery that I made was that in spite of hearing deterioration of sound with a ferrite clamp on the coaxial cable I experimented with putting the same ferrite clamp on a balanced AES/EBU cable and to my amusement the sound improved drastically! Now I could not hear any digital artifacts after using ferrite clamp on the balanced cable. It became so natural that I was shocked by this experiment.
It means that the ferrite clamp cut out HF noise coming from the sound card and this noise no longer interfered with my DAC.
The question that I want to ask you is what is the reason for this phenomenon? A ferrite clamp deteriorates quality on unbalanced coaxial cable but it improves sound on balanced AES/EBU cable. Why is this?
And I have a second question. In my experiments I removed the stock SPDIF transformer in the converter and changed it to an AES/EBU transformer by Lundahl and this change also improved quality of my DAC.
How can a transformer change sound in the presence of the following circumstances?
1. the converter does not use master clock because the DAC has a synchronous reclocking trigger.
2. As far as I know sound cards of this type of quality are bit perfect.
These are 2 my questions. I’d appreciate if any experienced guys take a look at this case.
A few days ago I made a small discovery regarding blocking HF noise coming from PC.
I have a professional sound card Prodif-88 that I use as a source. Next in line is SPDIF to I2S converter with SN75LBC176D transceiver and a DIY DAC with synchronous reclocking.
I know that PC radiates a lot of HF noise. The converter has a transformer at the input. It means galvanic isolation with PC. Until recently I have been of an opinion that SPDIF transformer is quite sufficient for blocking HF noise but recently I found out that it’s not.
Let me tell you how I came to that conclusion.
My soundcard has AES/EBU digital outputs in balanced mode. Recently I found out that SN75LBC176D in the converter is connected in balanced mode whereby it receives SPDIF signal in balanced mode and then it converts it into unbalanced mode and feeds DIR9001 with the signal which transforms SPDIF into 3 signals but masterclock is not used because the DAC has a synchronous reclocking trigger and the soundcard is slaved to oscillator in the DAC.
I did not understand that my converter can understand balanced mode and because of that I have been using a resistor divider in order to convert balanced AES/EBU into unbalanced SPDIF.
A year ago I experimented with a ferrite clamp that I clamped on the SPDIF cable in order to reduce HF noise but I found out that after putting on a ferrite clamp the quality noticeably deteriorated, I heard some kind of distortions which sound as a very slow sound. The sound started irritating me a lot and I disconnected the ferrite clamp.
Please, take a note that I connected sound card with converter via coaxial cable and I put a ferrite clamp on this coaxial cable. This is very important.
Then a few months ago as soon as I found out that my converter understands balanced AES/EBU I tried to switch to a specially designed balanced AES/EBU cable and the quality improved significantly.
This was my first discovery but I want to discuss the second discovery!
The second discovery that I made was that in spite of hearing deterioration of sound with a ferrite clamp on the coaxial cable I experimented with putting the same ferrite clamp on a balanced AES/EBU cable and to my amusement the sound improved drastically! Now I could not hear any digital artifacts after using ferrite clamp on the balanced cable. It became so natural that I was shocked by this experiment.
It means that the ferrite clamp cut out HF noise coming from the sound card and this noise no longer interfered with my DAC.
The question that I want to ask you is what is the reason for this phenomenon? A ferrite clamp deteriorates quality on unbalanced coaxial cable but it improves sound on balanced AES/EBU cable. Why is this?
And I have a second question. In my experiments I removed the stock SPDIF transformer in the converter and changed it to an AES/EBU transformer by Lundahl and this change also improved quality of my DAC.
How can a transformer change sound in the presence of the following circumstances?
1. the converter does not use master clock because the DAC has a synchronous reclocking trigger.
2. As far as I know sound cards of this type of quality are bit perfect.
These are 2 my questions. I’d appreciate if any experienced guys take a look at this case.
Too many unknowns in your setup to know for sure, but there seems to be some kind of common-mode RF crap that might be getting into the DAC's analog stages or something, probably via the shield. I'm very much assuming that your converter and DAC setup is not AES48-2005 compliant and has an inherent Pin 1 Problem.
Putting the clip-on ferrite on forms a common-mode choke at high frequencies (often easily tens of MHz). The balanced signal basically does not see its inductance at all since it's (ideally) all differential, but common-mode (shield) currents do. Your SPDIF signal seems to be partly common mode even at high frequencies and marginal enough for that bit of filtering (or the choice of transformer) to make a difference in error rates.
I would aim to verify this by way of measurement somehow, ears are not too dependable a tool.
Putting the clip-on ferrite on forms a common-mode choke at high frequencies (often easily tens of MHz). The balanced signal basically does not see its inductance at all since it's (ideally) all differential, but common-mode (shield) currents do. Your SPDIF signal seems to be partly common mode even at high frequencies and marginal enough for that bit of filtering (or the choice of transformer) to make a difference in error rates.
I would aim to verify this by way of measurement somehow, ears are not too dependable a tool.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.