Clock cable question

[Ian B]

This might be slightly left field for this forum, but I've been trying a Grimm cc2 clock with my professional Burl Audio Mothership analog to digital converters for my recording studio.

I'm liking many things the Grimm clock does in terms of resolution, speed, vividness, and 3D depth. But it is altering the timbre of things I record in a not good way, suppressing bass and midrange, and adding brightness and bite. One colleague said that it sounds "less musical" despite the increase in audio quality.

I used the cheapest 75 ohm BNC cable I could find for the demo. My question for those with experience with external clocks and cables: Can a better clock cable correct the treble-heavy timbre I'm getting, or is this more likely just a clock and converter mismatch? Thanks.

 

[DonH50]

For about $20 you can buy a 1-m Amphenol RF cable, or $100 gets you an H+S (Huber+Suhner) cable, either of which is a qualified cable from a reputable vendor. Available from Digikey or other electronics places. I do not like Pasternack but have used Fairview Microwave for RF cables in the past so you might want to look at their selection. However, it's hard to conceive of a cable mechanism that would affect the clock to the extent of altering the sound as you describe. I'd guess something else is going on. Do you know the signal levels and interface impedance requirements of the devices? Perhaps you need to use a shorter cable, or a 50-ohm cable instead of 75-ohm cable for the clock? BNC's are normally 50 ohms though "skeletonized" 75-ohm versions exist.

 

[Ian B]

I'm just using a one meter Belden that cost $23. The cable and clock output says 75 ohm, so shouldn't that be correct?

It could just be what the clock does to my converters is a mixed bag, which is something people tend to say about external clocks.

I might try a lower end Shunyata because I've had good results with their digital cables, and the factory right around the corner (so to speak). But I am also skeptical that while I probably can increase performance, I might not be able to do anything about the timbre that is happening.

 

[tony22]

I'm just using a one meter Belden that cost $23. The cable and clock output says 75 ohm, so shouldn't that be correct?

It could just be what the clock does to my converters is a mixed bag, which is something people tend to say about external clocks.

I might try a lower end Shunyata because I've had good results with their digital cables, and the factory right around the corner (so to speak). But I am also skeptical that while I probably can increase performance, I might not be able to do anything about the timbre that is happening.

Ian, if I may, listen to Don and try the Amphenol or H+S before spending significantly more money on something with the sort of profit margins typical in high end audio. Not that you couldn’t possibly find a good cable, but these two companies have been making precision RF cables for decades. I suspect they’ve forgotten more about building RF cables than the audio companies who didn’t know what a 10MHz cable is up until a couple of years ago. In my working days we used both brands in Radar prototyping and manufacturing, the kind of stuff that using a 10MHz clock for audio looks like the child’s play that it is.

 

[Paul1960]

I don't know about your specific configuration but I used to use a Gustard R26 with a LHY OCK2 external clock. I was flabbergasted by the difference the clock cables made as well as the fact that they needed 100+ hours before they really sounded right. I have no explanation for that and perhaps it was a psycho-acoustic delusion but to my ears it was easily and repeatably noticeable. The R26 took a 50 ohm cable and the LMR400 I bought off ebay for $25 was nearly as good as the Harmonic-Technology DC III, 50 ohm Single Crystal Copper Cable I bought from Hong Kong for $110. Other 50 Ohm cables I tried were less good. I did buy a couple of 75 ohm cables for use with my Etherregen but I heard no significant difference with the sound there. Bottom line is the cable can make a difference, I don't think you need to spend a ton of money, but do try a few cables to see what you like and just leave them connected and powered on for a week before you evaluate. Sorry not to have a specific recommendation for a 75 ohm cable.

 

[Al M.]

I don't know about your specific configuration but I used to use a Gustard R26 with a LHY OCK2 external clock. I was flabbergasted by the difference the clock cables made as well as the fact that they needed 100+ hours before they really sounded right. I have no explanation for that and perhaps it was a psycho-acoustic delusion but to my ears it was easily and repeatably noticeable. The R26 took a 50 ohm cable and the LMR400 I bought off ebay for $25 was nearly as good as the Harmonic-Technology DC III, 50 ohm Single Crystal Copper Cable I bought from Hong Kong for $110. Other 50 Ohm cables I tried were less good. I did buy a couple of 75 ohm cables for use with my Etherregen but I heard no significant difference with the sound there. Bottom line is the cable can make a difference, I don't think you need to spend a ton of money, but do try a few cables to see what you like and just leave them connected and powered on for a week before you evaluate. Sorry not to have a specific recommendation for a 75 ohm cable.

Yes, cables make a difference.

Also, cable length can make a difference, as explained here:

Why longer is generally better for an S/PDIF Digital Cable

Both my digital cables are 1.5 meter in length.

 

[tony22]

Yes, cables make a difference.

Also, cable length can make a difference, as explained here:

Why longer is generally better for an S/PDIF Digital Cable

Both my digital cables are 1.5 meter in length.

A typical 10Mhz square wave clock signal has about a 10ns rise time.

 

[DonH50]

Best I can tell, the Grimm and Burl units are both specifying AES3 which is a 75-ohm interface. It is pretty typical to use 50-ohm BNC connectors for 75-ohm video (etc. ) cables. You could try setting the Grimm's output impedance to "low" and see if that makes a difference. It's also possible you simply got a bad cable. Or something else is going on. Affecting just a portion of the signal, particularly the bass, due to a clock cable is quite unexpected and a mechanism for that is not clear to me. Unless the clocks end up causing signals to be out of phase somehow.

The S/PDIF article is basically saying triple transit times matter when impedances are not matched. True, but in most applications (including S/PDIF and AES3) performance is best when impedances are properly matched and cables are as short as possible. Longer cables have higher loss and lead to higher jitter, big problems for serial data (Ethernet, PCIe, SAS/SATA, USB, etc.) or RF systems, not so much for digital audio where a relatively large amount of jitter is tolerable and frequencies are relatively low.

It's also possible the external clock connection is adding ground noise or a ground loop that creates low-level "hash" perceived as a shift in timbre.

 

[Al M.]

A typical 10Mhz square wave clock signal has about a 10ns rise time.

Thanks. I don't have a 10 MHz clock yet, but that would explain why, for example, the clock cables that come with Afterdark 10 MHz clocks are shorter.

 

[DonH50]

A typical 10Mhz square wave clock signal has about a 10ns rise time.

Is that for oscillators targeting audio clocks? I know the majority of 10 MHz oscillators I used in my day job (Epson, Seiko, Toyo, etc. packaged oscillators) had risetimes in the 2-4 ns range, but those targeted much higher-frequency systems (and we often used higher-frequency oscillators anyway). I don't know much about specs for clocks targeting audio stuff but assumed (probably wrongly) they were using standard off-the-shelf oscillators.

 

[tony22]

Is that for oscillators targeting audio clocks? I know the majority of 10 MHz oscillators I used in my day job (Epson, Seiko, Toyo, etc. packaged oscillators) had risetimes in the 2-4 ns range, but those targeted much higher-frequency systems (and we often used higher-frequency oscillators anyway). I don't know much about specs for clocks targeting audio stuff but assumed (probably wrongly) they were using standard off-the-shelf oscillators.

My comment was based on stuff from 20 or more years ago for stuff at lower frequency systems. It wouldn’t surprise me a bit that the newer oscillators and those targeted for higher frequencies operate with speedier rise times as you note.

 

[antigrunge]

I just went from a Sommer 75 Ohm video cable to a Huber +Suhner RG214 for less than $50er cable and the sifference is vast. Go for it and you won’t regret it! And btw: the 1.5m reflection recommendation applies to SP/Dif and not BNC. The shorter the better to avoid Rfi/EMI incursion

 

[tony22]

Hmm. The RG214 is a 50 Ohm cable. Is it possible you had the 75 Ohm cable on a 50 Ohm feed?

 

[antigrunge]

On a sine rather than square wave clock the cable impedance doesn’t matter

 

[tony22]

True on paper (the 10x guideline), but personally I wouldn’t leave anything to chance. Just me.

 

[DonH50]

Reflections are less an issue with a pure sine wave, but mismatched impedances will reduce amplitude depending upon cable length.

The oft-cited xx-m requirement for S/PDIF, or anything else, assumes mismatches and that they are significant enough to cause errors. In practice shorter is almost always better (unless you overload the receiver).

 

[tony22]

Agreed, Don. In my mind there’s no need to introduce any more potential variables if there’s no need to.

 

[antigrunge]

The cables are 2 feet long and the audible results confirm the theory
I also use Mini Circuits low pass filters on both with unknown impact on impedance

 

[orange55]

Yes clock cables can change the sound. Here is a good review: https://audiobacon.net/2018/08/11/the-audiophiles-short-list-the-best-digital-coaxial-cables/

I have the SOtM cable and it does exactly as the review suggests.

 

[tony22]

I’ll state the obvious, but clock cables are not audio cables. If there is a change to the sound using different clock cables (and I know there is, having heard it myself), it likely has at least something to do with how the cable and its connectors may be degrading the transmission of the clock signal. My guess is most if not all of the clock cable manufacturers mentioned in Bacon’s review have not used anything like a VNA or a TDR to determine how well their cables meet RF performance metrics.