Forums

Beating Nyquist?

Started by Andor July 25, 2007

Clay wrote:
>> >>>Andor wrote: >> >>I know a CD player that dithers the read-out clock of the CD.
?????? The clock jitter of the DAC is the thing which is carefully avoided in the audio.
> In >>audio, it is generally accepted that correlated noise (showing up as >>lines in the spectrum) is more disturbing (audio wise) than >>uncorrelated wideband noise - this can be translated to transforming >>spectral lines in the audio jitter to wideband noise.
This is what dithering is used for. The clock jitter has a very detrimental effect.
>> >>Why is clock jitter dithering used for processors?
For the formal compliance to the EMC standards which specify the measurements by a narrow band analyser.
> Sometimes clock dithering is used for processors to keep their emitted > radiation down to a lower level.
The noise power is the same. It is just spread around the spectrum.
> A processor I like to use definitely > has this feature (Rabbit 4000). Computer clocks tend to be a horrible > source of noise, so anything that can reduce it is welcome.
Quite the contrary. The typical solution for the sensitive electronics is to lock all of the clocks to the same reference. So the interference frequencies are known and kept away from the bands of interest. Vladimir Vassilevsky DSP and Mixed Signal Design Consultant http://www.abvolt.com
Clay wrote:
(snip)

> Sometimes clock dithering is used for processors to keep their emitted > radiation down to a lower level. A processor I like to use definitely > has this feature (Rabbit 4000). Computer clocks tend to be a horrible > source of noise, so anything that can reduce it is welcome.
It isn't that it is a lower level, it is that the peak is lower. I spreads out in frequency space, which is mostly good. -- glen
On Jul 27, 3:43 pm, glen herrmannsfeldt <g...@ugcs.caltech.edu> wrote:
> Clay wrote: > > (snip) > > > Sometimes clock dithering is used for processors to keep their emitted > > radiation down to a lower level. A processor I like to use definitely > > has this feature (Rabbit 4000). Computer clocks tend to be a horrible > > source of noise, so anything that can reduce it is welcome. > > It isn't that it is a lower level, it is that the peak is lower. > I spreads out in frequency space, which is mostly good. > > -- glen
It depends on how you look at it. If the detector's bandwidth is less than the bandwidth of the emission, it will be at a lower level in the detector. while the total emitted power won't be likely reduced by jitter, the power density is. Clay
On Jul 27, 11:21 am, Vladimir Vassilevsky <antispam_bo...@hotmail.com>
wrote:
> Clay wrote: > > >>>Andor wrote: > > >>I know a CD player that dithers the read-out clock of the CD. > > ?????? > The clock jitter of the DAC is the thing which is carefully avoided in > the audio. > > > In > >>audio, it is generally accepted that correlated noise (showing up as > >>lines in the spectrum) is more disturbing (audio wise) than > >>uncorrelated wideband noise - this can be translated to transforming > >>spectral lines in the audio jitter to wideband noise. > > This is what dithering is used for. The clock jitter has a very > detrimental effect. > > > > >>Why is clock jitter dithering used for processors? > > For the formal compliance to the EMC standards which specify the > measurements by a narrow band analyser. > > > Sometimes clock dithering is used for processors to keep their emitted > > radiation down to a lower level. > > The noise power is the same. It is just spread around the spectrum. > > > A processor I like to use definitely > > has this feature (Rabbit 4000). Computer clocks tend to be a horrible > > source of noise, so anything that can reduce it is welcome. > > Quite the contrary. The typical solution for the sensitive electronics > is to lock all of the clocks to the same reference. So the interference > frequencies are known and kept away from the bands of interest. >
Sometimes one ends up with several separate pieces of equipment and locking of clocks is not pracical. Also the internal workins of some of the equipment may not be made known. However if shielding and dithering reduce the noise power density enough, then the other pieces of equipment may not be interfered with. If on the other hand one is designing a receiver, then having all of the clocks locked and if a proper frequency plan is implemented where all of the interference can be moved to don't care regions of the spectrum, then dithering is not needed. It all depends! Clay
On Jul 27, 11:21 am, Vladimir Vassilevsky <antispam_bo...@hotmail.com> 
wrote:

(snip)

> Sometimes one ends up with several separate pieces of equipment and > locking of clocks is not pracical. Also the internal workins of some > of the equipment may not be made known. However if shielding and > dithering reduce the noise power density enough, then the other pieces > of equipment may not be interfered with. If on the other hand one is > designing a receiver, then having all of the clocks locked and if a > proper frequency plan is implemented where all of the interference can > be moved to don't care regions of the spectrum, then dithering is not > needed. It all depends!
I remember seeing the schematic to the Commodore-64, where it is very obvious what happened. The processor part is on one corner, (triangle in matrix terms), the video logic in the other, and a PLL in between. It seems that the first design had the two running of separate clocks, resulting in a "swimming" effect on the video display. (Especially when the two clocks are close to a simple ratio.) The human (among others) visual system is very sensitive to some types of motion. -- glen
glen herrmannsfeldt wrote:

> > On Jul 27, 11:21 am, Vladimir Vassilevsky <antispam_bo...@hotmail.com> > wrote:
Sorry, I did the snip wrong in the last post. It was supposed to follow up Vladamir's: > Quite the contrary. The typical solution for the sensitive > electronics is to lock all of the clocks to the same reference. > So the interference frequencies are known and kept away from > the bands of interest. -- glen
Vladimir Vassilevsky wrote:
> > > Clay wrote: >>> Why is clock jitter dithering used for processors? > > For the formal compliance to the EMC standards which specify the > measurements by a narrow band analyser. > >> Sometimes clock dithering is used for processors to keep their emitted >> radiation down to a lower level. > > The noise power is the same. It is just spread around the spectrum.
It is more often a matter of simple getting through those approvals tests, than achieving anything useful. A PC BIOS typically has a clock blurring option, to help with approvals. After approval it gets turned off, to improve the clock margins.
>> A processor I like to use definitely >> has this feature (Rabbit 4000). Computer clocks tend to be a horrible >> source of noise, so anything that can reduce it is welcome. > > Quite the contrary. The typical solution for the sensitive electronics > is to lock all of the clocks to the same reference. So the interference > frequencies are known and kept away from the bands of interest.
More typically, it is the edges, rather than the frequencies which are kept away from the band of interest. For example, in a sigma delta converter, an analogue delay is usually used to push the modulator sampling as far as possible from the edge of the clock for the filters, whilst still being synced to them. Steve
"glen herrmannsfeldt" <gah@ugcs.caltech.edu> wrote in message 
news:PMGdnWA4q9urYjXbnZ2dnUVZ_ramnZ2d@comcast.com...
> Fred Marshall wrote: > (snip) > >> I don't see why not. Unless you're talking about periodic epochs and >> Dirichlets. >> If you pass the non-uniform samples through a lowpass filter / convolve >> with a particularly structured sinc / then that would seem to work fine. > > Each basis function should be one at its sample point, and zero at all > the other sample points. I do agree that it isn't obvious that is > consistent with a Nyquist frequency. > >> Now I can imagine having a different sinc for each sample but then what >> would be it's related lowpass? You'd have to relate at least one sample >> to another, etc. etc. And, in the end, the sincs are of infinite extent >> and so forth. So, I think they all have to be of the same periodicity. >> Now, there *are* other basis sets but I believe they can always be >> decomposed into a basis of sincs as long as the underlying signal space >> is bandlimited. And, that's a condition we accept perforce isn't it? >> That is, once sampled you can't tell the difference. > > I am trying to remember the way it is done in crystallography. > (Crystallography is pretty much sampling theory in three dimensions.) > In that case, more complicated structures are described using > a simpler lattice with a basis, where the basis is the combination > of atoms in the unit cell. See: > > http://en.wikipedia.org/wiki/Structure_factor > > To me, the basis that you want is the basis that is one at the > corresponding sample point and zero at others sample points. > I think, though, that it will not be orthogonal in continuous space. > It is the right one, though, if you consider reconstruction as a > sum over basis functions multiplied by the sample values. > > -- glen
Glen, It can be shown that a sinc is just a sum of sinusoids and that sines and cosines form an equally valid temporal basis set for a finite and regularly discrete spectrum. Just being finite and discrete makes it pretty obvious, eh? Fred
Fred Marshall wrote:
(snip)

> It can be shown that a sinc is just a sum of sinusoids and that sines and > cosines form an equally valid temporal basis set for a finite and regularly > discrete spectrum. Just being finite and discrete makes it pretty obvious, > eh?
Any linear, and linearly independent, combination of basis functions can be used as a new basis. The sinc basis is convenient for uniform spaced samples, as each sample is represented by one basis function in the reconstruction. Similarly, there are basis functions that represent the reconstructed samples of a non-linearly sampled signal. Those are not sinc. -- glen
On 31 Jul., 02:29, glen herrmannsfeldt <g...@ugcs.caltech.edu> wrote:
> Fred Marshall wrote: > > (snip) > > > It can be shown that a sinc is just a sum of sinusoids and that sines and > > cosines form an equally valid temporal basis set for a finite and regularly > > discrete spectrum. Just being finite and discrete makes it pretty obvious, > > eh? > > Any linear, and linearly independent, combination of basis functions > can be used as a new basis. The sinc basis is convenient for uniform > spaced samples, as each sample is represented by one basis function > in the reconstruction. > > Similarly, there are basis functions that represent the reconstructed > samples of a non-linearly sampled signal. Those are not sinc.
Glen, can you tell me more about those basis functions for non- linearly sampled signals? Regards, Andor