R/E/P Community

Please login or register.

Login with username, password and session length
Advanced search  

Pages: [1]   Go Down

Author Topic: What about multibit ADC's ??  (Read 4105 times)

Chuck

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 89
What about multibit ADC's ??
« on: May 10, 2004, 01:38:31 pm »

Hi there,

I was wondering if any of you guys still use multibit ADCs for recording ?

Today, those sigma delta ADCs seem to be very popular, but maybe only because they are cheaper to make...

I personally consider the sigma-delta conversion process more a interpretation or calculation result, than a real conversion.

A multibit ADC is a real ADC, in that it converts the amplitude of a signal to a x-bit number at a fixed time point. It is not averaging over time but a true high-speed measurement.

Multibit ADCs are precision devices and naturally more expensive. Their resolution is limited to 16..18 bits, as a higher resolution is not possible.

I am considering to construct a true multibit ADC with 18-bits precision up to 500kHz sampling, and would like to know if there would be any interest from your side in trying it out.

Thanks for input.

Charles Smile

Logged
ALTMANN MICRO MACHINES - GERMANY

davidc

  • Full Member
  • ***
  • Offline Offline
  • Posts: 168
Re: What about multibit ADC's ??
« Reply #1 on: May 10, 2004, 04:03:35 pm »

Hi Chuck,

I personally would agree with you that in principle that a multi-bit ADC would be far better than a delta sigma. I use a delta sigma because I don't know of anyone making a multi-bit.

I use a multi-bit DAC (an old modified hi-fi DAC), and it is far superior to any delta sigmas that I have heard. I would be very interested in hearing your ADC. What chipset are you thinking of using?

Best Regards

David C
Logged

Nika Aldrich

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 832
Re: What about multibit ADC's ??
« Reply #2 on: May 10, 2004, 05:08:50 pm »

Chuck wrote on Mon, 10 May 2004 18:38

Hi there,

I was wondering if any of you guys still use multibit ADCs for recording ?

Today, those sigma delta ADCs seem to be very popular, but maybe only because they are cheaper to make...

I personally consider the sigma-delta conversion process more a interpretation or calculation result, than a real conversion.

A multibit ADC is a real ADC, in that it converts the amplitude of a signal to a x-bit number at a fixed time point. It is not averaging over time but a true high-speed measurement.

Multibit ADCs are precision devices and naturally more expensive. Their resolution is limited to 16..18 bits, as a higher resolution is not possible.

I am considering to construct a true multibit ADC with 18-bits precision up to 500kHz sampling, and would like to know if there would be any interest from your side in trying it out.



I think several things need clarifying, here.  First, most of the converters currently used in the industry are multibit designs.  Yes, they use delta sigma modulators, but they use multibit variations of them.  I've seen 2, 4, and 5 bit delta sigma modulators in use.  I think it will be very difficult to find a 1 bit DSM in use right now from any of the big designers (AKM, Burr Brown, Crystal, Analog Devices, Sony, dCS, Lavry, Meitner, etc.)

I believe what you might be talking about is not a multibit design, in general terms, but rather a Successive Approximation Register, or SAR design.  This type of design has been largely abandoned in the industry as having too much distortion (or noise), negating any benefits it might bring.  This has been overcome by using heavily oversampled 1 bit designs, that have now become heavily oversampled multibit designs.  The weaknesses in the 1 bit designs were particular types of distortion that would present itself in certain, very specific conditions, but the benefits were that it provided nearly perfect integral and differential linearity.  Further, the noiseshaping (an inherent part of a DSM) was capable of pushing the noisefloor into a range that allowed far improved performance in the audible range.  The new multibit DSM designs couple the benefits of SAR converters (not having the specific distortion components yielded by 1 bit designs) with the benefits of 1 bit designs (low noisefloor in the audible range and low differential and integral non-linearity).  I think you would have an extremely difficult time designing an SAR based converter that exceeded the performance of even the mediocre designs used on the market right now.

Nika.
Logged
"Digital Audio Explained" now available on sale.

Click above for sample chapter, table of contents, and more.

Chuck

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 89
Re: What about multibit ADC's ??
« Reply #3 on: May 11, 2004, 07:25:45 am »

Hi David,,

I'm glad to hear of someone who knows about the sonic difference between a sigma-delta DAC and a good R2R DAC.

Up to now - according to my level of listening awareness - nothing comes close to a well implemented PCM1704 colinear DAC-chip, but also the old TDA1543 is very good... although not as accurate ...

As for the multibit ADC, I have not decided which chip I am going to use. I have found interesting multi-bit ADC-chips from Maxim, Burr-Brown and Analog Devices. They are available in 16 and 18-bit designs. I think you may want the 18-bit version ?

Charles Smile

Logged
ALTMANN MICRO MACHINES - GERMANY

Chuck

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 89
Re: What about multibit ADC's ??
« Reply #4 on: May 11, 2004, 08:16:07 am »

Nika Aldrich wrote on Mon, 10 May 2004 23:08



I believe what you might be talking about is not a multibit design, in general terms, but rather a Successive Approximation Register, or SAR design.  This type of design has been largely abandoned in the industry as having too much distortion (or noise), negating any benefits it might bring.  This has been overcome by using heavily oversampled 1 bit designs, that have now become heavily oversampled multibit designs.  The weaknesses in the 1 bit designs were particular types of distortion that would present itself in certain, very specific conditions, but the benefits were that it provided nearly perfect integral and differential linearity.  Further, the noiseshaping (an inherent part of a DSM) was capable of pushing the noisefloor into a range that allowed far improved performance in the audible range.  The new multibit DSM designs couple the benefits of SAR converters (not having the specific distortion components yielded by 1 bit designs) with the benefits of 1 bit designs (low noisefloor in the audible range and low differential and integral non-linearity).  I think you would have an extremely difficult time designing an SAR based converter that exceeded the performance of even the mediocre designs used on the market right now.

Nika.


Hi Nika,,

yes, by saying multibit, I mean a SAR design. I think they are very good for audio, because they can sample and convert a complete 18-bit precision value at a single and exact point in time, whereas a sigma-delta ADC operates on many very low precisions samples that are averaged.

I agree that a 4 or 5-level sigma delta converter may be much better than a 1-bit. Unfortunately I did not find the number of bits in the respective data-sheets.

However, I consider a high-precision 18-bit conversion much better, in that it is able to do a real conversion or sample at time point X without averaging over time and computation. It is compatible with the theory of sampling, whereas I do not see the validity of the sigma delta process.

As I read the data sheet of the AKM5394 (which I think is one of the very good sigma-delta ADCs), there is a very interesting THD plot on page 21. It shows that THD+N is in fact very low (-104dB) for a full scale input. But at a -30dB input level THD linearly rises to -88dB and the diagram suggests that for a -60dB input the THD would rise to really unacceptable levels.
Exactly the same combination is a fact for sigma-delta DACs. They have very low distortion for a full-scale output, but for a low level, where low distortion is needed, they completely fail in presence of a good R2R multi-bit DAC.

As for a modern SAR-ADC the distortion is specified flat at about -100dB for input signals down to -60dB. That would suggest to me, that a multibit ADC is much better for music, as it is able to precisely convert low level inputs, and it is low-level signals that would need a low distortion figure.

As I understand, sigma delta ADC need the noise-shaping process, because of the incredible high quantization noise that comes out of the 1- or very low-bit modulator.

A high precision multibit ADC has extremely low quantization errors, and thus the complete average, decimate and calculate to insignificant 24-bit numbers process of sigma-delta ADCs is not necessary.

It is just straightforward precision sampling.

I actually don't know what is better.

I think I have to try it out.

Charles Smile
Logged
ALTMANN MICRO MACHINES - GERMANY

Nika Aldrich

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 832
Re: What about multibit ADC's ??
« Reply #5 on: May 11, 2004, 09:38:19 am »

Chuck,

Before you design a new SAR based ADC I would do some research into DSM designs.  Your post indicates to me that a little bit of research may be helpful.  I will do my best to point you in a few directions through responding to various points in your post.

Chuck wrote on Tue, 11 May 2004 13:16


whereas a sigma-delta ADC operates on many very low precisions samples that are averaged.


No, a delta sigma converter works in essentially the same way as an SAR converter in that it takes discreet samples at given moments in time.  The sample taken really is the simple quantization of the waveform at that moment of time.  The only exception is that there is a feedback loop built in which serves to filter the quantization error.  Therefore, the sample at a given moment in time is just as valid as the sample at a given moment in time in an SAR converter with the exception of the feedback loop of the error, which serves to filter the noisefloor to get it out of the audible range.

Quote:

I agree that a 4 or 5-level sigma delta converter may be much better than a 1-bit. Unfortunately I did not find the number of bits in the respective data-sheets.


The AKM5394 that you mention is clearly a multibit design - as is just about every chip released since around 2000.

Quote:

However, I consider a high-precision 18-bit conversion much better, in that it is able to do a real conversion or sample at time point X without averaging over time and computation.


This means of comparison is relatively irrelevant.  No sample system is accurate with a single sample.  A single sample in an SAR design is useless and does not conform to Nyquist/Shannon because there isn't enough data with which to filter and reconstruct.  Disregarding that, however, a single sample in delta sigma designs is just as much a simple, quantized sample of the amplitude at a given moment of time.  It is when more samples are taken in an SAR design that it becomes slightly more complex, but only in that there is a feedback loop, lowering the amplitude.  

Quote:

It is compatible with the theory of sampling, whereas I do not see the validity of the sigma delta process.


The fact that you do not see it yet does not mean that it does not conform.  *S*   It does actually conform very simply to Nyquist and Shannon, though popular press about it confuses how it works by talking about it in terms that don't relate.  I often see talk about delta sigma modulators "averaging" or "measuring only the difference between amplitudes" or any number of other things.  I can just as easily explain them to you as a pure Nyquist-based quantizer, and for the sake of discussion in the audio industry this is the best way to explain it.

Quote:

As I read the data sheet of the AKM5394 ... there is a very interesting THD plot on page 21. It shows that THD+N is in fact very low (-104dB) for a full scale input. But at a -30dB input level THD linearly rises to -88dB and the diagram suggests that for a -60dB input the THD would rise to really unacceptable levels.


Can you show me this chart?  You can email it to me at Nika_Aldrich@sweetwater.com and I will help interpret it.  I have two .pdfs on the AKM5394A and I see no chart like this?  

Quote:

Exactly the same combination is a fact for sigma-delta DACs. They have very low distortion for a full-scale output, but for a low level, where low distortion is needed, they completely fail in presence of a good R2R multi-bit DAC.


This is incorrect, but I will need more information about your perspective on these converters in order to most efficiently explain how they work and why this statement is erroneous.

Cheers,
Nika
Logged
"Digital Audio Explained" now available on sale.

Click above for sample chapter, table of contents, and more.

davidc

  • Full Member
  • ***
  • Offline Offline
  • Posts: 168
Re: What about multibit ADC's ??
« Reply #6 on: May 11, 2004, 02:44:51 pm »

Chuck wrote on Tue, 11 May 2004 12:25

Hi David,,

I'm glad to hear of someone who knows about the sonic difference between a sigma-delta DAC and a good R2R DAC.

Up to now - according to my level of listening awareness - nothing comes close to a well implemented PCM1704 colinear DAC-chip, but also the old TDA1543 is very good... although not as accurate ...

As for the multibit ADC, I have not decided which chip I am going to use. I have found interesting multi-bit ADC-chips from Maxim, Burr-Brown and Analog Devices. They are available in 16 and 18-bit designs. I think you may want the 18-bit version ?

Charles Smile




Hi Charles,

I have to say I really like the TDA1543. It is a very involving DAC and give vocals tremendous expression.

I have alway liked Analog Devices chips, and my reference DAC is an 18 bit Analog Devices chip (AD1863 if I remember rightly).

What sampling rates are you going to have this thing handle - 192, 384, 768  Laughing . Seriously 88.2 would be really useful.

Best Regards

David
Logged

davidc

  • Full Member
  • ***
  • Offline Offline
  • Posts: 168
Re: What about multibit ADC's ??
« Reply #7 on: May 11, 2004, 02:54:09 pm »

Nika,

Does what you are saying apply to DACs as well?

Best Regards

David

Logged

Chuck

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 89
Re: What about multibit ADC's ??
« Reply #8 on: May 12, 2004, 12:13:55 am »

davidc wrote on Tue, 11 May 2004 20:44


Hi Charles,

I have to say I really like the TDA1543. It is a very involving DAC and give vocals tremendous expression.

I have alway liked Analog Devices chips, and my reference DAC is an 18 bit Analog Devices chip (AD1863 if I remember rightly).

What sampling rates are you going to have this thing handle - 192, 384, 768  Laughing . Seriously 88.2 would be really useful.

Best Regards

David



Hi David,,

My current project is a low-cost DAC with the TDA1543, that runs with a car-battery.

It has a custom receiver-chip with two precision VCXO clocks, that makes it work for multiples of 44.1 and 48kHz, up to 192kHz, non-oversampling.

The AD1862 you are mentioning is a premium DAC chip. I find it a pity, that almost any R2R is out of production today, with exemption of the excellent, yet expensive PCM1704.

"marketing bits vs. real bits".

As for the multibit ADC, it would however do multiples of 44.1 and 48kHz, up to... what ever...

Charles Smile
Logged
ALTMANN MICRO MACHINES - GERMANY

Nika Aldrich

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 832
Re: What about multibit ADC's ??
« Reply #9 on: May 12, 2004, 01:44:59 am »

davidc wrote on Tue, 11 May 2004 19:54

Nika,

Does what you are saying apply to DACs as well?



Certainly.  The advantages are more obvious on the ADC side because traditional, resistor based ADCs were always a step behind their respective DAC counterparts.  Fortunately, with multi-bit delta sigma designs they're pretty equal.

Nika.
Logged
"Digital Audio Explained" now available on sale.

Click above for sample chapter, table of contents, and more.

davidc

  • Full Member
  • ***
  • Offline Offline
  • Posts: 168
Re: What about multibit ADC's ??
« Reply #10 on: May 12, 2004, 05:46:40 am »

Nika Aldrich wrote on Wed, 12 May 2004 06:44

davidc wrote on Tue, 11 May 2004 19:54

Nika,

Does what you are saying apply to DACs as well?



Certainly.  The advantages are more obvious on the ADC side because traditional, resistor based ADCs were always a step behind their respective DAC counterparts.  Fortunately, with multi-bit delta sigma designs they're pretty equal.

Nika.



Well that is interesting Nika, because IMHO multi bit DACs subjectively stomp all over delta sigma DACs. So maybe it is the same for ADCs. I have not been able to tell, because I have not been able to get my hands on one.

Best Regards

David C
Logged

Nika Aldrich

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 832
Re: What about multibit ADC's ??
« Reply #11 on: May 12, 2004, 09:11:33 am »

davidc wrote on Wed, 12 May 2004 10:46

Well that is interesting Nika, because IMHO multi bit DACs subjectively stomp all over delta sigma DACs. So maybe it is the same for ADCs. I have not been able to tell, because I have not been able to get my hands on one.

Best Regards

David C



I am afraid I am not familiar with model numbers, but it might be worth clarifying that the delta sigma designs really are multi-level variations.  I It is also possible that the additional non-linearities in R-2R designs are such that you find them to be "euphonic" even though they are less accurate.

Nika.
Logged
"Digital Audio Explained" now available on sale.

Click above for sample chapter, table of contents, and more.
Pages: [1]   Go Up