@tvrgeek Nice, the Vidar amplifiers are very nice. Thank you helps to put things in perspective, with your post.
- ...
- 84 posts total
- 84 posts total
@tvrgeek Nice, the Vidar amplifiers are very nice. Thank you helps to put things in perspective, with your post. |
Did a deeper dive. ( getting used to my new DAC, so lots of time) First, related is PCM. Source and end 75 Ohms, so a 75 Ohm coax is the "correct" cable. Length should not matter until it is so long as signal loss comes to play. No issue with jitter caused by reflections but you are dependent on the source clock. "S" and "P" knew what they were doing. OK, USB. First of all, it is NRZ encoding. Shape of the rise and fall is irrelevant. It's a 90 Ohm specified cable, twisted pair. Shielding was not part of the spec for USB 1 and 2 as it was designed for keyboards or printers. So only needed to have reasonable CMR. Now we use this for audio between devices with different power and ground paths. That's our world. The end point ( DAC) design needs to recognize this. Both for loss and for reflections, there is a distance limitation of 5M. There is no specification for minimum length. ( Thumb drives work just fine @ 0) A "0" is +/- .01V and a "1" is 3.6 to 4.5 V. As you can see almost the full 4V hysteresis so it would take a HUGE amount of noise or reflection to be detected and with matched impedance, that is not happening. I again stand by my "decent cable and well designed input" as all that is required for USB audio. USB mode 2. Going past that requires "extra ordinary claims". Not saying they are not valid, but the require proof as the technology says it just ain't happening. I noticed a You-Tube video where our friend at PS was talking about USB jitter in relation to the source clock. That was ONLY USB mode 1. We don't do that any more! The DAC buffers and re-clocks. PS: USB-3 is even more robust including transport layer CRC, but we are not there yet. |
Forgot to add. USB 1 and 2 are not "guaranteed" transports like TCP/IP is. More like UDP/IP. It is up to the application layer which I do not believe is either for audio. But as I have shown, it is quite reliable. That does mean a poorly implemented receiver can in fact make a mistake not caught and if horribly enough of them, maybe audible. It would likely need to be by coincidence MSB or close to it. I know this won't close the USB debate by "true believers" Fine, there are plenty of sources that will take your Money for Nothing :) |