Digital Coax to XLR adapters?

The input on your DAC is AES/EBU which is a balanced XLR connection, but intended for a digital signal at 110 ohms. Your coax input on your DAC and output on your CD player is a coax S/PDIF input, which is a 75 ohm digital signal. So no, you can't use a connector to adapt one to the other. They are very different connections and use different standards and protocols. You need to go coax to coax using a digital RCA cable (NOT an analog RCA cable). Perhaps there is a digital coax splitter you could find, that would allow you to switch between inputs, but I don't know much about that. Google is your friend. Or maybe someone else can chime in.

Actually, what you need is 75 ohm to 110 ohm impedance transformer. Get one of these: Canare BCJ-XP-TRB 75 Ohm to 110 Ohm Digital Audio Impedance Transformer. For the money, it's worth a try. Note that it wants a BNC connection on the coax in side, so you may need an adapter for that.

I used this transformer and it works (with Benchmark DAC1). Be careful with the 6MHz Canare transformer bandwidth since protocol requires always 64 bits per sample (32bit per each channel)

44.1kHz x 64bit=2.82MHz OK
96kHz x 64bit=6.1MHz OK
192kHz x 64bit=12.3MHz Not working

Protocol is pretty much the same but characteristic impedance and levels are different. Since you need to get from S/Pdif to AES/EBU you need to change characteristic impedance from 75ohm to 110ohm and increase voltage from 0.5V to 5V (active circuit or transformer required)

Bandwidth limit of 6MHz will also slow down transitions to about 27ns. It won't affect typical player that has about 25ns transition but will affect more expensive transports' faster transitions - for better or for worse. I wouldn't worry about it now since it is system dependent and effect are unpredictable. Transformer is cheap and also sold on Ebay where I bought mine. Be sure it is BNC to XLR and not the other way around (XLR on transformer has to be male). Use number Drubin provided to avoid mistake.

Excellent points by Kijanki, as usual. How did you calculate the 27 ns, though? If it were similar to a first order RC-type rolloff, for instance, the risetime and falltime corresponding to a 6 MHz bandwidth would be 0.35/6 MHz = 58 ns. Also, putting a bandwidth limitation corresponding to a 27 ns transition time in series with 25 ns I believe would result in risetimes and falltimes of about 37 ns, corresponding to the square root of the sum of the squares (I think :-)), which could very conceivably result in an audibly significant increase in jitter (timing fluctuations) with some DACs in some setups.

Another possibility to consider is shown here, a (two-way) S/PDIF to AES/EBU converter that utilizes active circuitry, rather than (or perhaps in addition to) a transformer. It would not require an RCA-to-BNC adapter, and is described as being suitable for applications up to 24-bits/96 kHz.

I have no experience with any of these devices, though, and no feel for the extent to which inserting the Hosa converter into the signal path might result in an increase in jitter (as a consequence of either its effects on signal risetimes and falltimes, or noise it may add to the signal, or impedance inaccuracy). Although the user comments that you will see posted at the B&H link by various pro audio people are generally very favorable, FWIW.

Another alternative, at considerably higher price points, would be certain "reclockers," whose purpose is to improve sound quality by reducing jitter in the digital signal. Some of them can accept S/PDIF inputs while providing AES/EBU outputs.

Apologies if my question is naive.

At this point in the thread it would seem to be well established that it is not :-)

Best regards,
-- Al

Al, you're right 58ns it is. I took fastest hypothetical slew rate corresponding to 6MHz being 1/6e6/2pi=27ns. It would be more realistic to take 1/3 of the cycle for -3dB attenuated signal. I'm not sure if 27ns in series with 25ns should be calculated as square root of the sums of squares but if it is we get SQRT(25^2+58^2)=63ns - very slow. It might be a very bad thing in a noisy system. Reclocker is always wonderful solution opening possibility of PC Audio server. My Benchamark DAC1 has reclocker built in reducing jitter effects making possible problems that transformer brought (jitter) inaudible.