Toslink/ADAT Cables - Suggestions

There is no reason to worry about the optical cables.

Just use the cheap ones.

spdif/AES protocol is very slow. You also know that digital is on or off. The transceivers trigger off of edges of the transitions, so there is no worry about amplitude issues from lesser polishing on ends. The amplitude is still high enough that inclusions or cracks in the plastic won't attenuate much.

A lot of fiber, even high speed, high cost stuff, is still plastic cored.

Mainly with fiber it's the end polish that makes most of the difference unless you're using multiple wavelengths on single fibers.

Also, there is no performance difference between coax and fiber in spdif/AES except at higher bitrates, which the cheap optical transceivers can't handle anyway, so don't worry about trying to go to coax from an optical port.

The only downside to any of the lightpipe protocols (spdif/aes/adat) is that the common LED based transceivers are super slow, and the better ones used for things like madi are much more expensive which is why they're not used in commodity products.

Another interesting tidbit is that most toslink transceivers don't have polished interfaces, so your expensive super-polished cable isn't doing much at all



BTW, there was never a coax standard for adat like there was for spdif/AES, which is why you never see coax adat.

Even though spdif/AES and adat are biphase clock type signals, they are nothing alike at the protocol layer and are 100% incompatible.

BTW 2, I designed a jitter test unit for 2.5Ghz optical multimode fiber. It measured jitter in the femptosecond range over miles of test fiber..

A couple feet of plastic fiber absolutely won't matter for a few 96khz signals.

I designed a converter using spdif/AES and it was quite the learning experience, and one thing I can tell you from that is that the receiver and transmitter chips are the same for optical and coax, spdif and AES. Usually the designer just chooses one or the other and the coax is run directly into or out of the chip while the optical goes into or out of the chip to the driving LED or receiving phototransistor.

Not really much between them to create a difference in sound, unless the designers did something outlandish in their buffering, which would mean something is wrong with the design, not that coax is better than optical.

Also, by the nature of the biphase clocking in the streaming signal, which means the bitclock is part of the digital audio streams, if there was significant enough change in the stream to elicit a change in fidelity, the transceiver PLLs would lose streamlock and you'd get clicks on the audio.

These chips don't recognize grey areas in the digital streams. They use something called Schmitt triggers on their input pins. These are essentially hard decision points for the chip, and the signal is either seen as a High/1 or Low/0. The chip ignores everything in between that doesn't meet the threshold.

If a bit doesn't meet the threshold then you'll get a bit error and your PLLs will start to slip out of lock and you'll possibly get a click or pop or lose the signal altogether if enough of the bits are lost.

I think the biggest problem with the super cheap toslink cables is how well they fit into their receptacles. The cheap ones rarely fit, either too tight or too loose.

I had some that I got rid of because they'd vibrate out of the transceiver sockets over time from the fan vibration of the computer in the same rack. I think I also had one that just wouldn't go into a socket.

I bought some 5$ ones instead of the 1$ ones simply because they fit well, with a good solid click into the socket.