Why is low DC-content desirable for data stored on a CD?

[c.j[dot]w]

Hello,

On a CD, information is stored in 14-bit words. Between those words are
three merging bits. These contain no information, but are chosen,
partially, "to keep signal DC content low", according to a text I'm
reading. This is further explaned by "Low digital sum value (the number
of one bits minus the number of zero bits) minimizes the low frequency
and DC content of the data signal which permits optimal servo system
operation."

Why does low DC content permit "optimal servo system operation", and
what is at all meant by "servo system operation"?

Thanks in advance,
Regards Carl

 

[Bob Masta]

Hello,

On a CD, information is stored in 14-bit words. Between those words are
three merging bits. These contain no information, but are chosen,
partially, "to keep signal DC content low", according to a text I'm
reading. This is further explaned by "Low digital sum value (the number
of one bits minus the number of zero bits) minimizes the low frequency
and DC content of the data signal which permits optimal servo system
operation."

Why does low DC content permit "optimal servo system operation", and
what is at all meant by "servo system operation"?

Thanks in advance,
Regards Carl

I'll take a guess at this. I assume the servo system they are
talking about is the one responsible for tracking the signal
on the disk. It tries to maximize signal strength by moving the
sensor head radially as needed. Assuming it looks at the
raw signal and is AC coupled, long stretches of 1s or 0s
would give an erroneously low response that might interefere
with tracking. There might also be issues if it is trying to
derive a data clock based upon the frequency of edge
transitions.

Again, just a guess. Maybe somebody who actually knows
what they are talking about will step forward... <g>

Best regards,



Bob Masta
dqatechATdaqartaDOTcom

D A Q A R T A
Data AcQuisition And Real-Time Analysis
www.daqarta.com

 

[c.j[dot]w]

Thanks for the answer.

I'll take a guess at this. I assume the servo system they are
talking about is the one responsible for tracking the signal
on the disk. It tries to maximize signal strength by moving the
sensor head radially as needed. Assuming it looks at the
raw signal and is AC coupled, long stretches of 1s or 0s
would give an erroneously low response that might interefere
with tracking.

I don't really get it - could you explain further?

There might also be issues if it is trying to
derive a data clock based upon the frequency of edge
transitions.

This is true, and is one reason for wanting to avoid long streams of 1's
and 0's. But as I understand the explanation I read, the other reason is
more important.

Anyone else is also welcome with other ideas, or support to the above
answer.

 

[Active8]

Hello,

On a CD, information is stored in 14-bit words. Between those words are
three merging bits. These contain no information, but are chosen,
partially, "to keep signal DC content low", according to a text I'm
reading. This is further explaned by "Low digital sum value (the number
of one bits minus the number of zero bits) minimizes the low frequency
and DC content of the data signal which permits optimal servo system
operation."

Why does low DC content permit "optimal servo system operation", and
what is at all meant by "servo system operation"?

Thanks in advance,
Regards Carl

I think Bob's second answer is right. Normal information content is
random and by keeping it balanced (low DC content) it appears more
white noise like and is sure to have a frequency component at the
data clock frequency or a multiple thereof say, bit rate * 14 bits,
maybe. You reconstruct this clock frequency and compare it to a
reference clock to get an error signal which "controls" the servo.

That assumes the author considers the spindle speed as part of the
servo system.

As for radial tracking, I thought that was part of the focus system
where the photo detectors keep the laser over the data track. At any
rate, if it capacitively coupled, you don't want DC in the data, so
3 zeros between all ones words might fix that.

 

[BobGardner]

CDs use CIRC and EFM.... Cross Interleave Reed-Solomon Coding and
Eight-to-Fourteen Modulation. CIRC is an error correction scheme that adds
about 50% redundancy to the other channel one frame behind, so hopefully, an
error can be corrected with this addtional info (somehow?). But the EFM is a
lookup table of 256 of the avilable 16 thou 14 bit codes carefully chosen so
there is now run of 0s greater than 3. This allows the pll to stay synched even
on quiet passeges with lots of 0s being transmitted. The analog data is 2 8 bit
bytes... they feed each byte into the efm looker-upper. This is explained
pretty well in Pohlman Digital Audio.