24-09-2020, 09:14 PM
Mike, you're not alone there. I spend a lot of time getting people to think about the analogue reality of digital signals. In practical scenarios, it's usually analogue problems that cause digital signals to fail - anything from the wrong cable type through to interference. In effect, a digital "message" is carried on the analogue medium, and therefore the analogue medium has to be sufficiently good to allow the message to be decoded.
For example, one way to carry digital audio is a standard called AES3. This has much in common with the domestic "S/P-DIF" connections that you might have seen on the rear of some CD and DVD players. This particular standard is widely used, and is very reliable, but when it fails, it's almost always because of analogue problems like HF loss in very long wires, or a mis-connection that introduces unterminated stubs, which in term cause reflections and hence distortion of the analogue waveform that can prevent the receiver from being able to decide if the analogue voltage it gets is supposed to represent a logic 0 or a logic 1.
In practice, these "physical layer" problems dominate; it's very rare to have logic problem with the data carried on the interface. There is dedicated test equipment that looks at the data carried on the interface, but most of the time, an oscilloscope is what you need.
ADSL or VDSL were mentioned earlier. Previously I've detailed how I tripled my ADSL speed by attending to the analogue problems with the phone wiring in my house - eliminating unterminated stubs and ensuring the correct characteristic impedance of the wiring did that. A digital signal needs to sit on a solid analogue foundation
Even a CD is analogue. Ever looked at an eye waveform? It comes about because there are 2 amounts of reflectivity on a disc. The exact amounts vary from disc to disc, and of course, recordable CDs have rather less than a commercial pressing, but it's an analogue signal that comes back from the laser pickup, and the front end of the processing circuitry has to decide whether the analogue voltage is representing a logic 0 or logic 1.
The business about taking care of the analogue side of digital systems is usually called "signal integrity": https://en.wikipedia.org/wiki/Signal_integrity
For example, one way to carry digital audio is a standard called AES3. This has much in common with the domestic "S/P-DIF" connections that you might have seen on the rear of some CD and DVD players. This particular standard is widely used, and is very reliable, but when it fails, it's almost always because of analogue problems like HF loss in very long wires, or a mis-connection that introduces unterminated stubs, which in term cause reflections and hence distortion of the analogue waveform that can prevent the receiver from being able to decide if the analogue voltage it gets is supposed to represent a logic 0 or a logic 1.
In practice, these "physical layer" problems dominate; it's very rare to have logic problem with the data carried on the interface. There is dedicated test equipment that looks at the data carried on the interface, but most of the time, an oscilloscope is what you need.
ADSL or VDSL were mentioned earlier. Previously I've detailed how I tripled my ADSL speed by attending to the analogue problems with the phone wiring in my house - eliminating unterminated stubs and ensuring the correct characteristic impedance of the wiring did that. A digital signal needs to sit on a solid analogue foundation

Even a CD is analogue. Ever looked at an eye waveform? It comes about because there are 2 amounts of reflectivity on a disc. The exact amounts vary from disc to disc, and of course, recordable CDs have rather less than a commercial pressing, but it's an analogue signal that comes back from the laser pickup, and the front end of the processing circuitry has to decide whether the analogue voltage is representing a logic 0 or logic 1.
The business about taking care of the analogue side of digital systems is usually called "signal integrity": https://en.wikipedia.org/wiki/Signal_integrity







