SPDIF to TTL

That's a simple AC signal to DC logic conversion,

. | \ . AC ---||---+---| >O--+---- . | | / | . | | . '---/\/\---'

I'd say, yes, in the absence of any signal, an HCU04 gate would do a better job of settling at 1/2 Vcc, and perhaps do so with less class-A current, but in the absence of a signal, who cares what happens? In practice, S/PDIF lines have a continuous AC signal, so this simplified circuit's feedback RC node should find the average of the incoming data stream. Also, the continuous huge signal means the gates won't have a class-A rail-rail current problem.

s/pdif data doesn't have a 50% duty cycle, which gives me pause, but Andrew Kilpatrick, Randy McAnally and ESP = Rod Elliott swear by it, so it must be adequate. :-)

OTOH, you can get official s/pdif receiver connectors and ICs for a few bucks, and that seems a bit more professional.

Check that, s/pdif's biphase mark code is more than close enough to a 50% average for good RC coupling.

Averages 50% in the longer term, that's why it's an AC coupled signal.

Yes, but there's more to life than doing everything efficiently ;)

Grant.

"Dirk Bruere at NeoPax" schrieb im Newsbeitrag news: snipped-for-privacy@mid.individual.net...

In the following schematics page 17 upper left corner, you see a working solution (I have the eval board, its working nicely in coax SPDIF mode).

MIKE

It's interesting that TI picks an 'U04. An 74LVC2GU04 to be exact. They do use this simple ac-coupled-circuit in a PurePath Digital reference design for their rather sophisticated custom silicon for high-performance audio.

'U04' Variants have been used like this "forever". Some loonies have even been known to use them as amplifiers :-) ...Jim Thompson

Right, I'm well aware of that. Old stuff, dating back to the 60s when I started using cmos, or COS/MOS as RCA called it back then. But the question was about the viability of non-U types for this, do you have an opinion about that? Not what's naturally best, but whether it'll work OK.

It'll function. The problem I have with it is: an 'HC04 is effectively THREE 'HCU04's in series, with size scaling for the gate-source capacitance build-up. So a low-pass feed-back biases at considerable rail-to-rail current.

It'll function, but I don't recommend it.

For my custom stuff, where such an animal is absolutely needed, I add current sources in the respective (device) sources to limit bias build-up.

In a recent "monster" actuator driver I resorted to inverters where one side was a mirror until "swifty" speeds could be assured. ...Jim Thompson