here you go:
Originally Posted by Dan Lavry
I'm wondering why you say internal clock is always best for an AD converter ? Why are there so many people that say an external clock improved their AD conversion ? You hear people say … tighter bass.. better stereo field… etc… I've heard this a lot from people who say they clock their Lynx Aurora to an external clock such as Big ben.. etc.. IF people CAN tell a difference …. it must sound better…. no ? Does it just depend on the converter ? If the internal clock isn't that good, then maybe it can be improved from external clocking ? What's the scoop on this ?
There are a lot of people saying all sorts of things. Some of them that happened to be great ears that I respect agree with me. Others do not agree.
But what I say is not about "how many people say whatever". With enough advertising money, with the help of some false gurus, you can convince a lot of people of all sorts of things. What I say is solid and iron clad. I am talking about some basic principles of technology, not about opinions. For example, 1+1 =2, leaving no room for opinions. The same is true for the fact that the best way to clock a converter is to use internal clock. And BTW, as I mentioned earlier, Digidesign paper on clocking is backing me up on that, 2 years or so after I stated that internal is the way to go!
An ideal clock needs to have each cycle time be identical to any other cycle time. Slow and gradual variations (slow drift) will slow down or speed up the performance. That is not the issue regarding internal or external clocking. It is the fast variations that you want to avoid. You do not want to have a longer cycle, then a shorter one and another shorter… You want all the cycles to be the same!
The short term time variations between cycles is called jitter, so we want low jitter activity.
So say I have a great clock circuit. Where should I put it? The best place is to locate it INSIDE the converter chassis. Lets take the same circuit and put it in an external chassis. Now we have more causes that increase jitter. To begin with, we have the clock in a different chassis with a different power supply and ground. We have a cable which can pick up electric interference, and we are dealing with cable impedance and termination tolerance issues. We have a cable driver on one side, and a receiver on the other. By the time we get to the inside of the AD we have accumulated a lot of jitter, and that could be avoided by placing the same clock inside, next to the converter circuits (no driver, receiver, cable, interference, termination issues, ground loops….)
In addition, the clock we sync to has to be at the sample rate, but inside the converter we need to generate some higher frequency clocks at multiples of the sample rate. When we use internal clock, we start off with the highest frequency, and generate the lower "versions" by simple division. But for external sync, we receive that already more jittery clock and we now need to generate the higher frequencies, and that also ends up as more jitter. But even that aside, the arriving clock needs some cleanup which is typically done with a PLL circuit, and that too adds jitter…
So some "pro external clock" forces tried to sell the idea that their clock is better because there is some "magic" stuff inside which is proprietary. That is a real bad crock, though a lot of people bought into it. The whole concept is flawed, because music is "anything that you can fit within the hearing range".
The external clock box sends a clock to the AD. The clock box receives ZERO INFORMATION from the AD. So lets take a few AD's
AD #1 is perfect.
AD #2 has a lot of 1KHz jitter due to some radio signal interference
AD #3 has 60Hz jitter due to AC power getting in
AD # 4 is turned off
What kind of a signal will make the 4 converters better? It has to be the same signal for all converters, because the clock box has no idea what is at the end of the cable.
A doctor needs to cure a number of patients, but it does not ever know how they are, what their problems are (some may be dead). No phone, no fax.. The doctor needs to send a medication to all of them (the same medication). This is the perfect analogy.
The whole concept of conversion is based on equal clock time intervals, and the best clock is one that provides it. Getting that circuit away from the converter and piling up causes for more jitter is the wrong way to go. Claims of "special signals" in the clock are a crock, and they also contradict the low jitter specifications. If you alter the clock from a low jitter device to a carrier of some signal, you increase jitter, and no manufacture will tell you that they offer higher jitter!
I am sorry that so many are being talked into subjective garbage. I will stay with what is correct, and most of the great ears seem to agree with me, as expected. In fact, I tend to trust the ears of those that prefer internal clocks. They prefer transparency to reduced dynamic range and distortions.
When someone buys into a crock, it is difficult to admit being suckered. But it is best to try and revisit the issue and realize that you are being talked into accepting more noise, or some low level very inharmonic distortions as a good thing.
There is time when external clock is needed. So use internal when you must. But given the choice, use internal clock.
The only way an external clock will be better is if the specific internal clock design is very poor. That is unlikely, because the same designer also designed the receiver an PLL for the external clock case, and that circuitry is much more difficult and critical. If a unit has a good PLL and a good pull able clock, it is most likely will have a good internal stand alone clock. Yes, it is possible to find a specific gear with poor internal clock and good external clock circuits, but that would be a rare exception.