Relays in digital signal path

[Prune]

I'm using relays for I/O switching in the DAC. I'm concerned of whether that may introduce problems. My internal wiring for the digital signal is also coax, to try to maintan the characteristic impedance as much as possible, and even the jacks and cable plugs are 75 Ohm. However, this is clearly not the case for the relays. My question is how much of a problem that is, when for a couple of inches the signal path is not with 75 Ohm impedance. Also, the relays are latching, so no current flows through the coil except momentarily when switching, and they are sealed, so contacs should last. But I'm worried about reflections and so on.

[tiroth]

It is going to cause problems. You should use 75 ohm relays intended for this purpose, or terminate the line and use logic.

[Prune]

Well I searched for coaxial relays, and it comes up for $125. That's not reasonable for one relay.

[andrei]

Go to DigiKey and search for COTO 9002 series SIP reed relays. They are coaxially shielded (50 ohm) and not expensive. Alternatively, you could use solid state switches (like SN74CBT series).

[rmorris]

digital audio signals (spdif / aes-3) are not actually very sensitive to impedance matching. You shouldn't go too far out of course, but relatively minor impedance mismatch will result in 'ripples' at the top and bottom of the waveform but with minimal influence on the signal timing as it passes through the logic transition level.
Video signals are sensitive to this hence the emphasis on Z-matching there.
Nothing is ever perfect of course and i don't know what signal path length you are looking at but if I were you I'd try it then evaluate for real. I reckon cable capacitance / program induced jitter is much more of a problem.
regards

[OliFilth]

SPDIF is digital, so why use relays (mechanical or solid-state) that are designed for analogue signals?
Why not use a digital multiplexer, like the MC14512, which is designed for applications like this? (and make sure you terminate the digital line appropriately)

[emuman100]

Prune,

I would use solid state switching. Convert the coax to TTL and use a multiplexor like OliFilth said. It would be much better, with less loss, noise and transient replay "pops".

I'm jealous of you're DAC working. Well, now that I have everything but the soldering station (which comes in a few days) I'll be able to do a much better soldering job. After all, I'm certified to operate Pace equipment, so it'll be easy.

[rmorris]

Like the last two posts say it's probably better to use logic gates/switching. I don't know your particular application but if you can do the switching at logic level rather than the spdif signal itself then it becomes much simpler and you can terminate the lines correctly. If you need to switch the spdif/iec950 itself then there are probably devices designed for video signal switching that will do the job - maybe look at Elantec semiconductors ?
Having said that there can be cases where relays might be preferable depending on power supply voltages available etc etc...

[Jocko Homo]

I once used a relay to switch SPDIF siganls......and a cheap one from NAIS worked better than the expensive one designed for RF.

You can do it, but it does take some work to keep reflections down. CMOS solutions can also contribute to reflections, so it is not a foolproof solution.

Jocko

[rmorris]

what i'm suggesting is a properly terminated 75ohm receiver interface which produces a logic level which can then be switched/distributed locally without influence on the 75ohm line.
See Crystal Semiconductor / AKM datasheets / application notes for examples also the 'Digital Audio Interface' (or something like that) book by John Watkins (or similar - this from memory) or if you want a big read on the whole digital audio field - 'The Art of Digital Audio'.
Not that it has to be complicated.

What sort of problem symptoms did you get with the rf relay ?
You can get problems if there is not enough current for the contact design.