How CS8412 analog filter can affect sound?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hello, I have done a little modification to the analog PLL loop filter of CS8412 inserting a 4.7K resistor and making a total 427ohm, instead of the original 470ohm. I haven't done jitter measurements yet, but the sound is slightly better, a little bit open and clean on the mid-high. Difference is stuble but I will maintain the mod.

How this filter may affect analog sound? Is there a better scientific way to optimise it instead of copy other designs or do "try and error" ?
 

Attachments

  • filter.jpg
    filter.jpg
    7.5 KB · Views: 180
Jocko would have the answer if he was active here.

This is not exactly related to your question, but I thought I should write it. After a couple of months listening (suffering?) to a delta sigma dac (AD1852) I went back to my own diy design with the TDA1543 and it is clearly superior. There is no comparison in terms of emotional response. I stop being an "audiophile" nitpicking on what is wrong and start really enjoying the music. And keep playing albums and live shows one after another.

Then I went on to tweak some things. Look at the picture I attached. I have line filters in series (with some resistance between them) then a R-core trafo follows (note that the primaries are in series, for 220V line).

As you can see there are two separate secondaries and separate rectifier bridges. One secondary powers the reclocker board (with flip flops and a tent labs 11.2MHz) and the other secondary powers the TDA1543 dac. The receiver (CS8414), transmitter (CS8402) and the Xilinx CPLD are battery powered.

The tweak I tried was rewiring the R-core for 110V (primaries in parallel) and insert an additional isolation trafo, step down, 220 to 110V. IT COMPLETELY RUINED THE SOUND. Brian Adams MTV unplugged went from enjoyable to unlistenable. The bass seemed to increase, but everything else was worse. Voices, violins and the flute sounded metalic, dirty and annoying, which they didn´t before. Separation and soundstage suffered.

I think it happened because putting those primaries in parallel will increase the coupling between the supplies I was trying to keep separate (TDA1543 and reclocker). This is despite of the serious filtering and serious regulation on each leg (not shown). So it looks like it is some sort of jitter issue.

Reverted my failed tweak, everything was back to normal. Pretty good.
 

Attachments

  • 20190322_000025.jpg
    20190322_000025.jpg
    669.8 KB · Views: 129
Last edited:
I would say this...

The datasheet will probably suggest safe values that will make locking easy with a variety of sources, including jittery ones.

Whereas the tweak will probably make the pll perform a bit better (lower jitter) but might not lock on a very jittery source.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.