| Author |
Message |
Raul Fajardo
Guest
|
 Posted:
Tue Dec 21, 2004 9:10 pm Post subject:
Converting a sampled signal |
|
|
Hello,
I thought this question would touch somehow this community since this
task belongs to signal processing...
I always learned, that when I have a signal which has a known
information frequency band, I must sample this signal with at least
double this frequency, and take care to use a Low Pass Filter.
What if I have a already sampled signal but not quantized nor
digitalized. Should I use any kind of filter before trying to convert
this signal?
(Sampled -> Digital)
Would you add any filter between the AD conversor and this signal
output?
Thanks in advance
Raul Fajardo |
|
| Back to top |
|
 |
Tim Wescott
Guest
|
| Posted:
Tue Dec 21, 2004 9:16 pm Post subject:
Re: Converting a sampled signal |
|
|
Raul Fajardo wrote:
| Quote: | Hello,
I thought this question would touch somehow this community since this
task belongs to signal processing...
I always learned, that when I have a signal which has a known
information frequency band, I must sample this signal with at least
double this frequency, and take care to use a Low Pass Filter.
What if I have a already sampled signal but not quantized nor
digitalized. Should I use any kind of filter before trying to convert
this signal?
(Sampled -> Digital)
Would you add any filter between the AD conversor and this signal
output?
Thanks in advance
Raul Fajardo
|
No. It is the sampling process itself that aliases the frequencies --
once it's sampled your best bet is to get it to the ADC as accurately as
possible.
--
Tim Wescott
Wescott Design Services
http://www.wescottdesign.com |
|
| Back to top |
|
|
Jerry Avins
Guest
|
| Posted:
Wed Dec 22, 2004 1:03 am Post subject:
Re: Converting a sampled signal |
|
|
Raul Fajardo wrote:
| Quote: | Hello,
I thought this question would touch somehow this community since this
task belongs to signal processing...
I always learned, that when I have a signal which has a known
information frequency band, I must sample this signal with at least
double this frequency, and take care to use a Low Pass Filter.
|
Strictly, more than double.
| Quote: | What if I have a already sampled signal but not quantized nor
digitalized.
|
That's a neat trick. One such sample at a time can reside on the S&H's
capacitor, but it's rather difficult to think about "processing" a
single sample.
| Quote: | Should I use any kind of filter before trying to convert this signal?
(Sampled -> Digital)
|
It's not a signal. It's just a sample. All you can do is quantify it by
giving it a number.
| Quote: | Would you add any filter between the AD conversor and this signal
output?
|
That depends on whether I want the signal to be modified.
| Quote: | Thanks in advance
Raul Fajardo
|
Jerry
--
Engineering is the art of making what you want from things you can get.
ŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻ |
|
| Back to top |
|
|
glen herrmannsfeldt
Guest
|
| Posted:
Wed Dec 22, 2004 2:08 am Post subject:
Re: Converting a sampled signal |
|
|
Jerry Avins wrote:
| Quote: | Raul Fajardo wrote:
|
(snip)
| Quote: | I always learned, that when I have a signal which has a known
information frequency band, I must sample this signal with at least
double this frequency, and take care to use a Low Pass Filter.
Strictly, more than double.
|
If you are close enough to notice, I would worry about other
things, too. It would require an infinitely sharp low
pass filter. Crystals have finite frequency tolerance,
so you would have to add the possible crystal frequency
deviation.
| Quote: | What if I have a already sampled signal but not quantized nor
digitalized.
That's a neat trick. One such sample at a time can reside on the S&H's
capacitor, but it's rather difficult to think about "processing" a
single sample.
|
There are analog shift registers that can store sampled analog
signals. I believe there is also something more like DRAM than
a shift register that can hold analog voltages stored on capacitors.
They are used for simple voice record/playback toys, for example.
I go agree that is an unusual case, though. If you are going
to digitize it you might as well do it as soon as possible.
-- glen |
|
| Back to top |
|
|
Jerry Avins
Guest
|
| Posted:
Wed Dec 22, 2004 5:33 am Post subject:
Re: Converting a sampled signal |
|
|
glen herrmannsfeldt wrote:
| Quote: |
Jerry Avins wrote:
Raul Fajardo wrote:
(snip)
I always learned, that when I have a signal which has a known
information frequency band, I must sample this signal with at least
double this frequency, and take care to use a Low Pass Filter.
Strictly, more than double.
If you are close enough to notice, I would worry about other
things, too. It would require an infinitely sharp low
pass filter. Crystals have finite frequency tolerance,
so you would have to add the possible crystal frequency
deviation.
What if I have a already sampled signal but not quantized nor
digitalized.
That's a neat trick. One such sample at a time can reside on the S&H's
capacitor, but it's rather difficult to think about "processing" a
single sample.
There are analog shift registers that can store sampled analog
signals. I believe there is also something more like DRAM than
a shift register that can hold analog voltages stored on capacitors.
They are used for simple voice record/playback toys, for example.
I go agree that is an unusual case, though. If you are going
to digitize it you might as well do it as soon as possible.
|
While I have to agree than one can put a charge bucket brigade between a
sampling circuit and a digital quantizer, the time when it might have
made sense (except in video arrays) is long past. Do you imagine that
it's the sort of setup that the OP had in mind?
Jerry
--
Engineering is the art of making what you want from things you can get.
ŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻ |
|
| Back to top |
|
|
glen herrmannsfeldt
Guest
|
| Posted:
Wed Dec 22, 2004 6:28 am Post subject:
Re: Converting a sampled signal |
|
|
Jerry Avins wrote:
| Quote: | glen herrmannsfeldt wrote:
Jerry Avins wrote:
Raul Fajardo wrote:
|
(snip)
| Quote: | What if I have a already sampled signal but not quantized nor
digitalized.
That's a neat trick. One such sample at a time can reside on the S&H's
capacitor, but it's rather difficult to think about "processing" a
single sample.
There are analog shift registers that can store sampled analog
signals. I believe there is also something more like DRAM than
a shift register that can hold analog voltages stored on capacitors.
They are used for simple voice record/playback toys, for example.
I go agree that is an unusual case, though. If you are going
to digitize it you might as well do it as soon as possible.
While I have to agree than one can put a charge bucket brigade between a
sampling circuit and a digital quantizer, the time when it might have
made sense (except in video arrays) is long past. Do you imagine that
it's the sort of setup that the OP had in mind?
|
Or he already has the sample & hold, and is continuing down
the circuit board to the next step? The modern ADCs that
I know of include the S&H also, but I suppose there could be
reasons for adding one.
I agree it seems unlikely.
-- glen |
|
| Back to top |
|
|
Jerry Avins
Guest
|
| Posted:
Wed Dec 22, 2004 8:03 am Post subject:
Re: Converting a sampled signal |
|
|
glen herrmannsfeldt wrote:
| Quote: |
Jerry Avins wrote:
glen herrmannsfeldt wrote:
Jerry Avins wrote:
Raul Fajardo wrote:
(snip)
What if I have a already sampled signal but not quantized nor
digitalized.
That's a neat trick. One such sample at a time can reside on the S&H's
capacitor, but it's rather difficult to think about "processing" a
single sample.
There are analog shift registers that can store sampled analog
signals. I believe there is also something more like DRAM than
a shift register that can hold analog voltages stored on capacitors.
They are used for simple voice record/playback toys, for example.
I go agree that is an unusual case, though. If you are going
to digitize it you might as well do it as soon as possible.
While I have to agree than one can put a charge bucket brigade between a
sampling circuit and a digital quantizer, the time when it might have
made sense (except in video arrays) is long past. Do you imagine that
it's the sort of setup that the OP had in mind?
Or he already has the sample & hold, and is continuing down
the circuit board to the next step? The modern ADCs that
I know of include the S&H also, but I suppose there could be
reasons for adding one.
I agree it seems unlikely.
|
I suppose that the question arises from a theoretical framework
completely separated from a practical embodiment. It brings to mind the
Brazilian graduate student in optics interviewed by Richard Feinman who
knew all the math about and qualities of "plane parallel plate" but
could think of no example where it might actually apply. "Window pane"
was not lofty enough for scientific analysis.
Jerry
--
Engineering is the art of making what you want from things you can get.
ŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻ |
|
| Back to top |
|
|
|
|
|
|
FAQ
Memberlist
Register
Profile
Log in to check your private messages
Log in