Author Topic: Digital audio frequency signal ghosting  (Read 1397 times)

0 Members and 1 Guest are viewing this topic.

Offline CkRtech

  • Regular Contributor
  • *
  • Posts: 63
  • Country: us
Digital audio frequency signal ghosting
« on: March 07, 2017, 01:09:37 am »
Hey all -

I built a Cirrus Logic CS8406 circuit for an old game console (Sega Saturn) using this, and I seem to be getting an odd signal on the output. The frequency should remain constant (I assume), but the output is rather noisy - looks like a ghost of the signal can be seen, and the frequency seems like it is 170 kHz ...but it will spike to 250 or 350 on my scope. At the moment i snapped the attached scope photo, the freq registered 340 kHz.

Can any of you guys recommend some common issues (Ground specifics, wire length/gauge, noisy clock, etc.) that tend to cause an issue like this?

Thanks!
 

Offline Audioguru

  • Super Contributor
  • ***
  • Posts: 1508
  • Country: ca
Re: Digital audio frequency signal ghosting
« Reply #1 on: March 07, 2017, 02:26:08 am »
Your 'scope is not sync'd to the signal.
The signal appears to be a squarewave that has lots of harmonics. The 340khz is the second harmonic of the 170kHz fundamental so your frequency counter is counting the harmonic instead of the fundamental frequency.
 
The following users thanked this post: CkRtech

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 10131
  • Country: gb
    • Having fun doing more, with less
Re: Digital audio frequency signal ghosting
« Reply #2 on: March 07, 2017, 09:31:33 am »
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 
The following users thanked this post: CkRtech

Offline radiogeek381

  • Regular Contributor
  • *
  • Posts: 57
  • Country: us
    • SoDaRadio
Re: Digital audio frequency signal ghosting
« Reply #3 on: March 07, 2017, 01:07:33 pm »
Your scope is sync'd just fine.

This is not ghosting.  It is called differential manchester coding or a bi-phase mark code.   https://en.wikipedia.org/wiki/Differential_Manchester_encoding

This is the way bits move in the S/PDIF spec.  To send a stream of bits, the signal is encoded so that there are TWO bit transitions in a cycle to signal a "0" and one transition to signal a 1.  This means that there is at least one transition per bit-time.  Manchester coding allows the receiver to recover the clock from a bit stream by looking at transitions.  Higher level protocols ensure that, regardless of the contents of the messages, a "1" bit appears often enough in the bit stream to allow us to distinguish a two-edge interval from a one edge interval. 

If you put your scope in single trigger mode and hit the run button a few times, you'll see different patterns appear.  That's why you see "ghost" images...

 
The following users thanked this post: CkRtech, tooki

Offline JRosario

  • Contributor
  • Posts: 9
  • Country: us
Re: Digital audio frequency signal ghosting
« Reply #4 on: March 16, 2017, 11:33:37 pm »
Why is there a dispute on whether or not the scope is synched?

What about the images or situation described leads you to that conclusion (either one)?

Just looking to find out how you are troubleshooting this scenario.
 

Offline Audioguru

  • Super Contributor
  • ***
  • Posts: 1508
  • Country: ca
Re: Digital audio frequency signal ghosting
« Reply #5 on: March 16, 2017, 11:55:35 pm »
Why is there a dispute on whether or not the scope is synched?

What about the images or situation described leads you to that conclusion (either one)?

Just looking to find out how you are troubleshooting this scenario.
The bright parts have good sync because they are repeated over and over. The dim parts have poor sync because they are not repetitive enough and overlap the bright parts.
Simply slow down the 'scope timebase to see the entire data stream and have better sync.
 

Online Andy Watson

  • Super Contributor
  • ***
  • Posts: 1715
Re: Digital audio frequency signal ghosting
« Reply #6 on: March 16, 2017, 11:57:28 pm »
Why is there a dispute on whether or not the scope is synched?

What about the images or situation described leads you to that conclusion (either one)?
It's not a dispute. The nature of the encoding is such that there is no fixed part of the waveform that is repetitive in the manner that is usually expected. The scope is triggered, as good as it can be, on the edge of a data bit whilst the surrounding data bits are continuously changing - hence the apparent ghosting.
 
The following users thanked this post: tooki, radiogeek381

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 10131
  • Country: gb
    • Having fun doing more, with less
Re: Digital audio frequency signal ghosting
« Reply #7 on: March 17, 2017, 12:26:10 am »
Why is there a dispute on whether or not the scope is synched?

What about the images or situation described leads you to that conclusion (either one)?

Just looking to find out how you are troubleshooting this scenario.
The bright parts have good sync because they are repeated over and over. The dim parts have poor sync because they are not repetitive enough and overlap the bright parts.
Simply slow down the 'scope timebase to see the entire data stream and have better sync.

No.

The scope is triggering correctly, and all bits are correctly synched all the time. The reason some sections are less bright is because they occur less frequently.

FFI, make sure you understand "eye diagrams".
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf