Audio Clocks and Aging

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

john12ax7

Well-known member
Joined
Oct 15, 2010
Messages
2,463
Location
California, US
My studio is getting to the point where I need a master clock for several devices.

Are there any DIY word  clock projects?

For commercial units is there any downside to an older unit,  i. e. for the same make / model will a brand new unit be any better than one 5-10 years old.?  Crystals age,  at what time frame will there be a negative performance impact on jitter?
 
First ask yourself if you really need a clock.  And if you really do, buy a Grimm.  Many people using clocks could get by without and some would even have better audio quality without.

My understanding is that as crystal oscillators age they drift in absolute value but actually improve in short term stability.  Absolute value does not matter at all for our purposes, short term stability and phase noise is everything.

Find everything you can written by Bruno Putzeys on the topic, there are whitepapers on the Grimm site.

 
Every device you have that needs an external word clock is going to use it's own internal PLL to up-convert to it's required internal clocks. For instance, you may have ADC's/DAC's inside that require 64x SampleRate (48kHz etc).

If you consider the sources for error (or phase noise to use the correct term in clock-land), there's much to be said for the PLL inside your product being much more prone to error than your distributed word clock.

Consider using S/PDIF from your ADC as a clock source to the rest of the system. The S/PDIF Clock regenerators inside the S/PDIF reciever chips by TI, Cirrus, Wolfson are all pretty darn good, and with the faster clock rate of the S/PDIF source, have to do less multiplication up to your converter clock rate. I'd argue that helps (having your error corrected more often).

That, would be a free option to you with the tech you have in your studio.

If money is still burning a hole in your pocket, please go ahead and buy the grimm. Pro Audio needs every dollar possible these days.

/R
 
john12ax7 said:
My studio is getting to the point where I need a master clock for several devices.

Are there any DIY word  clock projects?

Make your own?

- a clean 3.3V supply.
- two brand new low ppm 3.3V crystal oscillators, 22.5792MHz and 24.576MHz.
- a 2 position rotary switch to select which oscillator as the audio master clock (1st = 48K/96K/192K, 2nd = 44.1K/88.2K/176.4K)
- two 4-bit logic counters, cascaded to divide the selected audio master clock by 128, 256 and 512.
- a 3 position rotary switch to select one of those three word clocks (1st = fs/512, 2nd = fs/256, 3rd = fs/128)
- a single buffer logic to amplify the selected word clock and send it out via a BNC connector.
 
metalb00b00 said:
Make your own?

- a clean 3.3V supply.
- two brand new low ppm 3.3V crystal oscillators, 22.5792MHz and 24.576MHz.
- a 2 position rotary switch to select which oscillator as the audio master clock (1st = 48K/96K/192K, 2nd = 44.1K/88.2K/176.4K)
- two 4-bit logic counters, cascaded to divide the selected audio master clock by 128, 256 and 512.
- a 3 position rotary switch to select one of those three word clocks (1st = fs/512, 2nd = fs/256, 3rd = fs/128)
- a single buffer logic to amplify the selected word clock and send it out via a BNC connector.

This. All of the logic can be rolled up into a small CPLD. Use a Silicon Labs dual oscillator (Si513) factory programmed for 24.576 and 22.5792, one pin selects the frequency. I'd use 74LVC buffers on the outputs to drive the BNCs.

It's really trivial.
 
Andy Peters said:
This. All of the logic can be rolled up into a small CPLD. Use a Silicon Labs dual oscillator (Si513) factory programmed for 24.576 and 22.5792, one pin selects the frequency. I'd use 74LVC buffers on the outputs to drive the BNCs.

It's really trivial.

Yep, I'd use a CPLD and a programmable oscillator as well.
It's just not DIY friendly because you have to know VHDL/Verilog and also need a programmer to burn in the firmware.
 
Actually as a simpler option, what if I just use one of my existing converters as the master word clock? And then build some parallel distribution amplifiers to drive all the other gear?

Then it would seem it's just a case of making 75 ohm unity gain DC coupled amplifiers?
 
john12ax7 said:
Actually as a simpler option, what if I just use one of my existing converters as the master word clock? And then build some parallel distribution amplifiers to drive all the other gear?

That would work, too.

Then it would seem it's just a case of making 75 ohm unity gain DC coupled amplifiers?

74LVC family buffers will do the job, easy, just put a 50-ohm resistor in series with the output and you're done. Decouple well,

-a

 
external clocking with Word Clock is not done to improve quality, but to tie large complex setups together that would otherwise be impossible/unpractical to realise.
Its a broadcast-thing, never intended for the audiophile abuses it sees today. A 48kHz rectangular signal is just way to loose to tie a 5mHz digital signal tightly together :)

If you want precise clocking, use DARS (if your equipment supports), or try to figure out a layout where you can clock through the digital audio signal.
 
If something has and aes-3 inputs does that mean it'll be compatible with DARs clocking?

Is AES clock sync the same as DARS?

 
For many years i thought that a master clock was only needed for multiple devices  digital setups . Since i have only have a DAW i thought i didn't need a clock. I was given a Apogee Bigben and i'm very surprised to say that i notice a change in sound from internal and external clock. May be i'm getting old :)
 
To me it's better, tighter stereo image , more precise high freq and reverb seems deeper. I don't know how to describe it very well but i can pick it 100% in blind test. I had people at my studio and they can hear it as well.
 
ubxf said:
To me it's better, tighter stereo image , more precise high freq and reverb seems deeper. I don't know how to describe it very well but i can pick it 100% in blind test. I had people at my studio and they can hear it as well.

Which converter were you using?
 
Yes, the Digidesign hardware was the mother of all necessary clocking.

That said, in recent hardware I haven't noted any benefit from external clocking - other than the obvious where you want multiple converters to sync/cooperate in multitrack mode

Jakob E.
 
metalb00b00 said:
Yep, I'd use a CPLD and a programmable oscillator as well.
It's just not DIY friendly because you have to know VHDL/Verilog and also need a programmer to burn in the firmware.

Define burn in the firmware
 
It means to "program" the firmware, to put the 1s and 0s in permanent memory, much like programming (putting the code into non-volatile memory) a FLASH-memory-based microcontroller (they're all FLASH memory based now).

Not to be confused with "burning in" an audio cable.
 
Back
Top