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.
benb said:
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.

I want to stick to the topic here so I'll be brief but there is no need to burn in an audio cable since it does not have any active components. A flex and continuity test is enough. You generally burn in active units such as the CPLD it's self overnight because there could be a faulty component surrounding it or a prolonged failure. There may be dualism to the the term "burn in." You may be referring to the compiler and the uploading of something like a .jed file to the CPLD. In my experience this has always been called "uploading the firmware."
 
buildafriend said:
There may be dualism to the the term "burn in." You may be referring to the compiler and the uploading of something like a .jed file to the CPLD. In my experience this has always been called "uploading the firmware."

I've been doing this a long time, and going back to the days of one-time-programmable PROMs and PALs, through EPROMs and erasable PALs (UV-erasable with ceramic windows) through EEPROMs (electrically-erasable) and now flash cells in CPLDs and some FPGAs with same, "burning" has always meant "programming the thing."
 
abbey road d enfer said:
Nevertheless, burning-in equipment before shipping it to the market is a more ancient procedure than the existence of ROM's.
+1...  "Burning" (proms) is different than "burning in gear," which is different than burning in audio cables (that is total BS).  ::)

I suspect the original term may even come from a completely different industry. Since many electronic product failures are infant failures that occur early in the product's life, some manufacturers would burn in gear for hours (days?) to weed out these early infant failures.

Back in the 80s I would run rack gear overnight in a burn-in rack. A power timer would cycle these products on and off, putting them through multiple thermal cycles to help accelerate any marginal solder joints or flaky components.  Later at Peavey many power amp burn-in racks only took hours (or less) and included short circuit stress tests, but the same principle applies to weed out infant failures.

Later the audiophool community adopted burning in as some voodoo to overcome buyer remorse if some product did not immediately wow the customer... "it needs to burn in for several days"  (cough). I researched this for a reality check (back last century) and the closest mechanism I could find was "self-annealing".  Low noise semiconductor junctions can get noisy if allowed to reverse zener, which is why many low noise designs add clamp diodes in reverse across  low noise base-emitter junctions. Forced annealing to quiet down a noisy junction involves running elevated current forward through the junction. In theory self-annealing can occur at reduced current over longer time (I never confirmed this).

Caveat,, annealing refers to improving damaged junctions (and never completely fixes them), so burning in an undamaged junction should make no difference in that context.

JR
 
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.

I had a similar situation regarding the clock on my Mytek DA, the superclock (internal clock of the Mytek) sounded better than the internal clock of the audio interface, several people noticed it too in blind tests, however the difference is really subtle and I would not spend a dime on a clock just to get a "better sound"
 
Nowdays I don't use an external clock,
My system is based on the Focusrite RED4 Pre interface, and I'm happier than ever with the conversion and clock quality of this unit.

Some years ago in the studio I was working  we had 192 and 96 Digidesign interfaces, they were used together with Apogee and Lavry converters, we had a Master Clock unit at the time and the studio owner said it made a difference in the sound quality.
We were quite busy all the time and I never made any Blind or A/B tests. I was happy with the sound that was enough for me.
So I can't confirm the external clock improved the audio quality.
I can confirm that like a lot of people I was also never impressed with the sound of the Digi 192 and 96 units, although I loved the Protools HD system

The only time I really noticed an improvement in sound quality using an external clock was actually in live sound.
It was with Yamaha live digital consoles. I used to get a lot of M7CL digital mixers at gigs,  the sound of this console was always week, being it the Mic Pres, the convertion or everything it just sounds cheap and hissy.
One time one engineer asked me to do a test with internal clocking or external clocking in this console,
I listened to it, with music and my voice. Some other Engineers did the same, we were asked to write our thoughts on a piece of paper.
And  we all wrote basically the same things.
The sound improved with external clocking on the Yamaha M7CL console, and it was easily noticeable.
How much improvement you ask me?
Well the console still sounds crap, and I reject it in all my Riders, but it improved enough to make me always take an External clock source to the gigs.

Just to point out that this console was released in 2005, it's old digital technology for nowadays standards.

At the present even Behringer was able to make a pretty good sounding digital mixer the X32



 
Whoops said:
The only time I really noticed an improvement in sound quality using an external clock was actually in live sound.
It was with Yamaha live digital consoles. ...
...The sound improved with external clocking on the Yamaha M7CL console, and it was easily noticeable.

Just to point out that this console was released in 2005, it's old digital technology for nowadays standards.
That's quite surprizing, since clock technology was pretty mature then, and clock reconstruction was not as advanced as today.
Venturing a dubious explanation: good clocking would highlight the other flaws when less stable clock would add jitter that may conceal them.  ???
 
abbey road d enfer said:
Venturing a dubious explanation: good clocking would highlight the other flaws when less stable clock would add jitter that may conceal them.  ???

Maybe you're right, might be a possible explanation or maybe the Yamaha M7CL clock was just plain Bad

 
abbey road d enfer said:
One has to work really hard to make a fixed clock that has so much jitter it degrades the performance.

Possibly, I can only ascertain and give opinions about my experience with the equipment and that was in fact the only time I've tried different clocks on a piece of gear and A/B them.
To be honest even in this situation the external clock used was not anything pristine or expensive, first time it was a Motu 828 (or similar) interface as the master clock, after it I've tried with what I had it was a Swissonic AD24 converter as the master clock, with one unit or the other as the master clock the sound of the M7CL was improved in my opinion.

Just to let you know it was not the case we bought some expensive clock and want to believe it was better.
In reality I prefer to travel light when Im on the road so my preference would always be to not have to travel with an extra unit just to be a master clock of a particular console only. But yes, the difference in sound was enough personally to justify the extra weight.

With the Yamaha PM5D an older model than the M7CL I did the same tests and couldn't discern a difference with internal or external clock. The PM5D was not in the same product range as the M7CL console, it was if I remember correctly the top of the Yamaha range.

I wouldn't completely set aside that the clock of the M7CL was badly designed, even if it's something hard to do,  they failed in a lot of departments.
This console has really hissy preamps, thats a well know problem, and the other digital consoles at the time from other brands were not hissy. Soundwise being the Mic Pres, the converters, the signal path, the internal digital summing ,or the clock or everything this was a badly designed product but had great software, user interface, and price for the time.
 
My 2c, I am using a Lynx Aurora and years ago I added a big ben to tie it and the AES16e together and to be honest I could never tell a difference between using it and letting the Lynx synchrolock run things.  Big Ben's go for pretty cheap nowadays too, so might that be a quick option?
 
i have a Drawmer M-Clock, one of the 1st made by the factory, very low serial = old.

has dual TCXO,  like all decent analog hybrid clocks, like Grimm Audio CC1, etc...
single quartz oscillator clocks are usually DDS or PLL "inferior", unless it has 64-Bit DDS controlled by a OCXO with Rubidium 10MHz option, like Antelope Trinity.

Aging usually is done at the Quartz cristal oscillator factory "burn-in", because it seems that drifts the 1st days very fast.
Urban Legend style...

the TCXO i have, can be calibrated with a small plastic screw driver, each, moves/tunes the capacitance of the circuit feeding the quartz oscillator.
increasing or decreasing the oscillating frequency in very small amounts...

i have an Oscilloscope with Real High Quality Frequency counter, most Digital Oscilloscopes are Software counters "crap/useless", like Keysight, Pico scopes, etc...

i can calibrate the TCXO to be exactly 44.1 x512 & 48khz x512 =
22,579200 MHz
24,576000 MHz..

some clocks are x256 other x512, other x1024, internally,
but the wordclock output signal is 192KHz max,
AES/EBU or S/PDIF black / empty signal is 6MHz aprox.
https://groupdiy.com/index.php?topic=61952.msg791132#msg791132
because the internal clock division, the real wordclock output frequency is:
44.1000 KHz,
48.0000 KHz,
88.2000 KHz,
96.0000 KHZ +/-1~2  ej: 96.0002 ~ 95.9998
176.4000 KHz +/-1~2
192.0000 KHZ +/-1~2

thats called 1ppm "1 part per million drift."

Drift is different than jitter, i recommend you read the Grimm Audio white papers in .pdf to learn more about that.

for example:
a +-/1ppm clock drift will have perfect pitch over 200Hz,

but jitter is different, related to Phase Noise, but Not the same...
Grimm Audio has the lowest Phase Noise design ive seen.

Apogee Big Ben has lower RMS than Mutec iClock, but iClock has less/smaller peaks.

there are 3rd harmonic Resonant clocks design like Black Lion Audio mk2,
they create the frequency needed, indirectly, as a harmonic of other, and eliminating the fundamental frequency.

Analog Digital Hybrid, "dual Quartz, HCMOS circuit or Dual TX with Digital circuit like Rosendahl Nanoclocks.", better designs can be Temperature Controlled / compensated.
http://studiorepair.com/gallery/misc/Rosendahl_Nanoclocks/slides/Rosendahl_Nanoclocks_Circuitry__STUDIOREPAIR_10101801_1210302530.html

DDS "Direct Digital Synthesis" limited by Negative feedback and Bits of resolution... the best is Keysight TrueForm DDS generator, Similar Apogee Big Ben, mutec iclock, Antelope, and others...

3rd Harmonic overtone like BLA Mk2

1x Quartz, silicon PLL type. "usually cheap, included in low cost equipment like Focusrite 18i20 mk2"

VCXO, tuning frequency is controlled by voltage ..

are some of the types of oscillators "XO",...

 
Back
Top