Hi All, I am going to be selecting a specific sigma-delta converter for an application and am getting a bit confused on the terminology. I have read up on the theory and understand it. Looking at this datasheet: http://focus.ti.com/lit/ds/symlink/ads1251.pdf On the first page, it says 20kHz data rate. Further down, on page 8, it talks about the oversampling and says that the output rate will scale so that the oversampling ratio stays at 64. So my understanding is that if i use 8Mhz clock, the maximum signal input bandwidth can be 10kHz right?, and for example if i scale it down to 3.072Mhz, since the datasheet says that the output rate is 8000Hz, meaning that my maximum input signal bandwidth can be 4kHz for proper conversion w/out aliasing?

# Question related to Sigma-Delta Datasheets

Started by ●June 12, 2007

Reply by ●June 12, 20072007-06-12

Kiran wrote:> Hi All, > I am going to be selecting a specific sigma-delta converter for an > application and am getting a bit confused on the terminology. I have > read up on the theory and understand it. > > Looking at this datasheet: http://focus.ti.com/lit/ds/symlink/ads1251.pdf > > On the first page, it says 20kHz data rate. Further down, on page 8, > it talks about the oversampling and says that the output rate will > scale so that the oversampling ratio stays at 64. So my understanding > is that if i use 8Mhz clock, the maximum signal input bandwidth can be > 10kHz right?, and for example if i scale it down to 3.072Mhz, since > the datasheet says that the output rate is 8000Hz, meaning that my > maximum input signal bandwidth can be 4kHz for proper conversion w/out > aliasing? >This, of course, assumes that the Nyquist rate is a prescription, not a limit that you can never reach. But Nyquist didn't say that. See http://www.wescottdesign.com/articles/Sampling/sampling.html for details. The data sheet should give you graphs of the performance of the thing's internal filters, and these filters' frequency responses will, being digital, scale with your clock rate. The data sheets _should_ detail what happens to the filters' responses as the clock rate changes, and should also give concrete suggestions about what you need to provide for anti-aliasing ahead of the converter, if it isn't built in. -- Tim Wescott Wescott Design Services http://www.wescottdesign.com Do you need to implement control loops in software? "Applied Control Theory for Embedded Systems" gives you just what it says. See details at http://www.wescottdesign.com/actfes/actfes.html

Reply by ●June 13, 20072007-06-13

Hi Tim, First of all, thanks for your response. I guess what I'm trying to ask is: lets say we had the following situation: A signal was completely bandlimited to 5kHz. Then, according to Nyquist's theorem, w/out any noise or any other extraneous effects, we'd buy a SAR adc or flash adc that would sample at >= 10kHz. Now on ADC spec sheets, they call this the sample rate, and they would have denoted 10ksps (kilo samples/sec). Now my question is, if we were to buy a sigma-delta converter to do the same thing, then would you look at the spec sheet and buy one that said 10ksps as its sample rate?, or are you supposed to take the oversampling into account yourself and buy 5*2*64 = 64ksps sigma-delta adc to do it? i know its probably a very simple and beginner question, but I appreciate everybody's help!, thanks On Jun 12, 5:38 pm, Tim Wescott <t...@seemywebsite.com> wrote:> Kiran wrote: > > Hi All, > > I am going to be selecting a specific sigma-delta converter for an > > application and am getting a bit confused on the terminology. I have > > read up on the theory and understand it. > > > Looking at this datasheet:http://focus.ti.com/lit/ds/symlink/ads1251.pdf > > > On the first page, it says 20kHz data rate. Further down, on page 8, > > it talks about the oversampling and says that the output rate will > > scale so that the oversampling ratio stays at 64. So my understanding > > is that if i use 8Mhz clock, the maximum signal input bandwidth can be > > 10kHz right?, and for example if i scale it down to 3.072Mhz, since > > the datasheet says that the output rate is 8000Hz, meaning that my > > maximum input signal bandwidth can be 4kHz for proper conversion w/out > > aliasing? > > This, of course, assumes that the Nyquist rate is a prescription, not a > limit that you can never reach. But Nyquist didn't say that. Seehttp://www.wescottdesign.com/articles/Sampling/sampling.htmlfor details. > > The data sheet should give you graphs of the performance of the thing's > internal filters, and these filters' frequency responses will, being > digital, scale with your clock rate. The data sheets _should_ detail > what happens to the filters' responses as the clock rate changes, and > should also give concrete suggestions about what you need to provide for > anti-aliasing ahead of the converter, if it isn't built in. > > -- > > Tim Wescott > Wescott Design Serviceshttp://www.wescottdesign.com > > Do you need to implement control loops in software? > "Applied Control Theory for Embedded Systems" gives you just what it says. > See details athttp://www.wescottdesign.com/actfes/actfes.html

Reply by ●June 13, 20072007-06-13

On Jun 13, 11:58 am, Kiran <Kiran.Ka...@gmail.com> wrote:> Hi Tim, > First of all, thanks for your response. > I guess what I'm trying to ask is: > lets say we had the following situation: > A signal was completely bandlimited to 5kHz. Then, according to > Nyquist's theorem, w/out any noise or any other extraneous effects, > we'd buy a SAR adc or flash adc that would sample at >= 10kHz. Now on > ADC spec sheets, they call this the sample rate, and they would have > denoted 10ksps (kilo samples/sec). Now my question is, if we were to > buy a sigma-delta converter to do the same thing, then would you look > at the spec sheet and buy one that said 10ksps as its sample rate?, or > are you supposed to take the oversampling into account yourself and > buy 5*2*64 = 64ksps sigma-delta adc to do it? > > i know its probably a very simple and beginner question, but I > appreciate everybody's help!, > thanks > > On Jun 12, 5:38 pm, Tim Wescott <t...@seemywebsite.com> wrote: > > > Kiran wrote: > > > Hi All, > > > I am going to be selecting a specific sigma-delta converter for an > > > application and am getting a bit confused on the terminology. I have > > > read up on the theory and understand it. > > > > Looking at this datasheet:http://focus.ti.com/lit/ds/symlink/ads1251.pdf > > > > On the first page, it says 20kHz data rate. Further down, on page 8, > > > it talks about the oversampling and says that the output rate will > > > scale so that the oversampling ratio stays at 64. So my understanding > > > is that if i use 8Mhz clock, the maximum signal input bandwidth can be > > > 10kHz right?, and for example if i scale it down to 3.072Mhz, since > > > the datasheet says that the output rate is 8000Hz, meaning that my > > > maximum input signal bandwidth can be 4kHz for proper conversion w/out > > > aliasing? > > > This, of course, assumes that the Nyquist rate is a prescription, not a > > limit that you can never reach. But Nyquist didn't say that. Seehttp://www.wescottdesign.com/articles/Sampling/sampling.htmlfordetails. > > > The data sheet should give you graphs of the performance of the thing's > > internal filters, and these filters' frequency responses will, being > > digital, scale with your clock rate. The data sheets _should_ detail > > what happens to the filters' responses as the clock rate changes, and > > should also give concrete suggestions about what you need to provide for > > anti-aliasing ahead of the converter, if it isn't built in. > > > -- > > > Tim Wescott > > Wescott Design Serviceshttp://www.wescottdesign.com > > > Do you need to implement control loops in software? > > "Applied Control Theory for Embedded Systems" gives you just what it says. > > See details athttp://www.wescottdesign.com/actfes/actfes.htmlYay nicknames!

Reply by ●June 13, 20072007-06-13

On Jun 13, 11:58 am, Kiran <Kiran.Ka...@gmail.com> wrote:> Hi Tim, > First of all, thanks for your response. > I guess what I'm trying to ask is: > lets say we had the following situation: > A signal was completely bandlimited to 5kHz. Then, according to > Nyquist's theorem, w/out any noise or any other extraneous effects, > we'd buy a SAR adc or flash adc that would sample at >= 10kHz. Now on > ADC spec sheets, they call this the sample rate, and they would have > denoted 10ksps (kilo samples/sec). Now my question is, if we were to > buy a sigma-delta converter to do the same thing, then would you look > at the spec sheet and buy one that said 10ksps as its sample rate?, or > are you supposed to take the oversampling into account yourself and > buy 5*2*64 = 64ksps sigma-delta adc to do it? > > i know its probably a very simple and beginner question, but I > appreciate everybody's help!, > thanks > > On Jun 12, 5:38 pm, Tim Wescott <t...@seemywebsite.com> wrote: > > > Kiran wrote: > > > Hi All, > > > I am going to be selecting a specific sigma-delta converter for an > > > application and am getting a bit confused on the terminology. I have > > > read up on the theory and understand it. > > > > Looking at this datasheet:http://focus.ti.com/lit/ds/symlink/ads1251.pdf > > > > On the first page, it says 20kHz data rate. Further down, on page 8, > > > it talks about the oversampling and says that the output rate will > > > scale so that the oversampling ratio stays at 64. So my understanding > > > is that if i use 8Mhz clock, the maximum signal input bandwidth can be > > > 10kHz right?, and for example if i scale it down to 3.072Mhz, since > > > the datasheet says that the output rate is 8000Hz, meaning that my > > > maximum input signal bandwidth can be 4kHz for proper conversion w/out > > > aliasing? > > > This, of course, assumes that the Nyquist rate is a prescription, not a > > limit that you can never reach. But Nyquist didn't say that. Seehttp://www.wescottdesign.com/articles/Sampling/sampling.htmlfordetails. > > > The data sheet should give you graphs of the performance of the thing's > > internal filters, and these filters' frequency responses will, being > > digital, scale with your clock rate. The data sheets _should_ detail > > what happens to the filters' responses as the clock rate changes, and > > should also give concrete suggestions about what you need to provide for > > anti-aliasing ahead of the converter, if it isn't built in. > > > -- > > > Tim Wescott > > Wescott Design Serviceshttp://www.wescottdesign.com > > > Do you need to implement control loops in software? > > "Applied Control Theory for Embedded Systems" gives you just what it says. > > See details athttp://www.wescottdesign.com/actfes/actfes.htmlchr15 is a n00b.

Reply by ●June 13, 20072007-06-13

On Jun 13, 2:58 pm, Kiran <Kiran.Ka...@gmail.com> wrote:> Hi Tim, > First of all, thanks for your response. > I guess what I'm trying to ask is: > lets say we had the following situation: > A signal was completely bandlimited to 5kHz. Then, according to > Nyquist's theorem, w/out any noise or any other extraneous effects, > we'd buy a SAR adc or flash adc that would sample at >= 10kHz. Now on > ADC spec sheets, they call this the sample rate, and they would have > denoted 10ksps (kilo samples/sec). Now my question is, if we were to > buy a sigma-delta converter to do the same thing, then would you look > at the spec sheet and buy one that said 10ksps as its sample rate?, or > are you supposed to take the oversampling into account yourself and > buy 5*2*64 = 64ksps sigma-delta adc to do it?- you just buy a 10ksps sigma-delta A/D, ignore the internal sample rate and look at the datasheet min bandwidth spec

Reply by ●June 13, 20072007-06-13

On Wed, 13 Jun 2007 15:29:41 -0700, steve wrote:> On Jun 13, 2:58 pm, Kiran <Kiran.Ka...@gmail.com> wrote: >> Hi Tim, >> First of all, thanks for your response. >> I guess what I'm trying to ask is: >> lets say we had the following situation: >> A signal was completely bandlimited to 5kHz. Then, according to >> Nyquist's theorem, w/out any noise or any other extraneous effects, >> we'd buy a SAR adc or flash adc that would sample at >= 10kHz. Now on >> ADC spec sheets, they call this the sample rate, and they would have >> denoted 10ksps (kilo samples/sec). Now my question is, if we were to >> buy a sigma-delta converter to do the same thing, then would you look >> at the spec sheet and buy one that said 10ksps as its sample rate?, or >> are you supposed to take the oversampling into account yourself and >> buy 5*2*64 = 64ksps sigma-delta adc to do it? > - > > you just buy a 10ksps sigma-delta A/D, ignore the internal sample rate > and look at the datasheet min bandwidth specTo expand: A S-D ADC works by sampling really fast in a way that generates a lot of noise, but at frequencies higher than the one you care about. Then it filters the snot out of the signal and decimates. So you'd be safe with your 10kHz sample rate out, but you may find that your signal has been filtered heavily up toward 5kHz, depending on the converter. All of the S-D ADC data sheets that I've seen specify the filter behavior in detail; yours should too. -- Tim Wescott Control systems and communications consulting http://www.wescottdesign.com Need to learn how to apply control theory in your embedded system? "Applied Control Theory for Embedded Systems" by Tim Wescott Elsevier/Newnes, http://www.wescottdesign.com/actfes/actfes.html

Reply by ●June 14, 20072007-06-14

Tim Wescott <tim@seemywebsite.com> writes:> A S-D ADCIt's "D-S," or "delta-sigma!" -- % Randy Yates % "I met someone who looks alot like you, %% Fuquay-Varina, NC % she does the things you do, %%% 919-577-9882 % but she is an IBM." %%%% <yates@ieee.org> % 'Yours Truly, 2095', *Time*, ELO http://home.earthlink.net/~yatescr