[drtechno]

Hi guys. I wanted to show you guys how signal level is directly associated with signal to noise when capturing.
 
below is a data sheet page of a common adc chip used in audio converters:
 
you'll also notice the s/n changes with sampling rate. But this might be a product of the filters too.

 

[MindsMirror]

Of course signal is one variable of signal/noise, changing the signal level affects the S/N.
 
Higher sample rates contain higher frequencies, and these higher frequencies contain additional noise, so increasing the sample rate will increase the noise. You'll notice that the A-weighted S/N is not specified at different sample rates. That is because the A-weighting only considers the audible frequencies, the additional noise from the higher sample rates is not present in the audible range so it would have no effect.

 

[gregorio]

  I wanted to show you guys how signal level is directly associated with signal to noise when capturing.

 
What a strange thing to post. The ratio of signal level to noise level (SNR) must obviously be affected by signal level (and also obviously by the noise level!). Pointing this out is a bit like pointing out that the result of the equation 2 + 3 is "directly associated" with the value of the number "2". If we were to change the "2" to a different integer then the result would no longer be 5.
 
Of more potential interest is your choice of wording; "directly associated" and "when capturing" because that depends on what we're capturing (recording), where we're capturing it and with what. For example, are we talking about capturing a test signal directly injected into an ADC chip or are we talking about the real world recording of say live musicians, in a studio/concert venue with microphones and mic-preamps? If it's the latter, then the noise introduced by the ADC chip cannot really be said to be "directly associated" with SNR. Adding noise to a signal which is say 1,000 times lower in level than the noise already in that signal makes no meaningful/audible difference to the SNR. Using the simple math equation analogy above, changing the "2" to 2.001 doesn't significantly change the result, which is still effectively 5.
 
G

 

[castleofargh]

I'm also not sure about the point of this post. signal to noise ratio will change if signal or noise or both change. there's a captain obvious feel to this.

 

[drtechno]

  Of course signal is one variable of signal/noise, changing the signal level affects the S/N.
 
Higher sample rates contain higher frequencies, and these higher frequencies contain additional noise, so increasing the sample rate will increase the noise. You'll notice that the A-weighted S/N is not specified at different sample rates. That is because the A-weighting only considers the audible frequencies, the additional noise from the higher sample rates is not present in the audible range so it would have no effect.

 
but A weight is very non linear and really not that suitable to use as a baseline to compare/anyalize ADCs.
 
A little background of A weight:
"While most audio engineers are familiar with the A-weighting curve, which was based on the 40 phon equal-loudness contour derived initially by Fletcher and Munson (1933) the later CCIR-468 weighting curve, now supported as an ITU standard is less well known outside of the UK and Europe. Originally incorporated into an ANSI standard for sound level meters, A-weighting was intended for measurement of the audibility of sounds by themselves. It was never specifically intended for the measurement of the more random (near-white or pink) noise in electronic equipment, though has been used for this purpose by most microphone manufacturers since the 1970s. The human ear responds quite differently to clicks and bursts of random noise, and it is this difference that gave rise to the 468-weighting, which together with quasi-peak measurement (rather than the rms measurement used with A-weighting) became widely used by broadcasters throughout Britain, Europe, and former British Commonwealth countries, where engineers were heavily influenced by BBC test methods. Telephone companies worldwide have also used methods similar to 468 weighting with quasi-peak measurement to describe objectionable interference induced in one telephone circuit by switching transients in another." ....source https://en.wikipedia.org/wiki/ITU-R_468_noise_weighting
 
 
That is why I look at C weight (which is the S/(n+d) ) which is the total product of what is going on ( s/n of the stage plus the quantization noise after spirous noise compensation circuits are applied ).
 
 
attached is the A weight curve for people who don't know or too lazy to go to wikipedia.
 
Also attached is the THD+N(ratio) vs. Input Level for the two recommended input circuit. and figure 12 which is the difference in input coupling capacitor size ( btw 90% of black lion's mods on adcs exploits this)
 

 

 

[drtechno]

oops forgot attachments. fixed,

 

[drtechno]

Normally an EE will never use weighting curves. These are exceptions because an ADC is a type of modulation circuit. That is why when someone builds an ADC for market, they have to submit it to RF regulating agencies (like the FCC). Because there is only so much modulation leakage on the inputs that is acceptable by the regulating authority.
 
That is why the diy-er "moc boutique companies brought to you by e-bay" only can really put out a DAC and not a ADC because of the ramifications that would incur if they didn't get FCC tested (for any modulating equipment sold in the US)
 
DACs are consider "receiving equipment" by the FCC so there isn't a certification they have to really go under.

 

[pinnahertz]

  Normally an EE will never use weighting curves. These are exceptions because an ADC is a type of modulation circuit. That is why when someone builds an ADC for market, they have to submit it to RF regulating agencies (like the FCC). Because there is only so much modulation leakage on the inputs that is acceptable by the regulating authority.
 
That is why the diy-er "moc boutique companies brought to you by e-bay" only can really put out a DAC and not a ADC because of the ramifications that would incur if they didn't get FCC tested (for any modulating equipment sold in the US)
 
DACs are consider "receiving equipment" by the FCC so there isn't a certification they have to really go under.

A designer will ( and should) use whatever weighting curve is appropriate to the project, regardless if he has earned his EE or not.  Weighting curves help relate measured noise to actual audibility, and do make sense when you understand which curves to use and when.
 
An ADC is not a modulation circuit.  
 
ADCs are subject to the same FCC regulations as DACs, and all other computing devices (see CFR 47, Part 15...suggest you read it thoroughly before making any further assumptions). 
 
The FCC isn't concerned with "modulation leakage" (there's no such thing), but is concerned with unintentional and incidental (as well as all intentional) RF radiation.  That can come from an input, output, the case, the power cord, anything, but with proper design will easily meet FCC Part 15 qualifications.  That's true of an ADC, DAC, or any other computing device. 
 
The FCC doesn't regularly test devices for Part 15 compliance, it's done only occasionally, and is not contingent to certification.  The section that pertains is 15.29:
 
(a) Any equipment or device subject to the provisions of this part, together with any certificate, notice of registration or any technical data required to be kept on file by the operator, supplier or party responsible for compliance of the device shall be made available for inspection by a Commission representative upon reasonable request.
(b) The owner or operator of a radio frequency device subject to this part shall promptly furnish to the Commission or its representative such information as may be requested concerning the operation of the radio frequency device.
(c) The party responsible for the compliance of any device subject to this part shall promptly furnish to the Commission or its representatives such information as may be requested concerning the operation of the device, including a copy of any measurements made for obtaining an equipment authorization or demonstrating compliance with the regulations.
(d) The Commission, from time to time, may request the party responsible for compliance, including an importer, to submit to the FCC Laboratory in Columbia, Maryland, various equipment to determine that the equipment continues to comply with the applicable standards. Shipping costs to the Commission's Laboratory and return shall be borne by the responsible party. Testing by the Commission will be performed using the measurement procedure(s) that was in effect at the time the equipment was authorized or verified.
 
"DACs are consider "receiving equipment" by the FCC so there isn't a certification they have to really go under."
 
No, a DAC fits under Part 15.3 (k) (excerpted)
 
(k) Digital device. (Previously defined as a computing device). An unintentional radiator (device or system) that generates and uses timing signals or pulses at a rate in excess of 9,000 pulses (cycles) per second and uses digital techniques; inclusive of telephone equipment that uses digital techniques or any device or system that generates and uses radio frequency energy for the purpose of performing data processing functions, such as electronic computations, operations, transformations, recording, filing, sorting, storage, retrieval, or transfer.
 
Part 15.3 includes many definitions, and details many receiving devices which also under Part 15 compliance rules, as they include internal (local) oscillators above 9kHz, and even though designed to receive radio signals, are capable of incidental radiation as well.  A DAC doesn't receive radio signals, doesn't match any FCC definition of "receiver", but has internal clock signals well in excess of 9kHz, and is therefore subject to Part 15 compliance. 
 
Please...Please...try to refrain from posting statements about which you are unfamiliar, and are drawing unfounded conclusions.

 

[drtechno]

its part 15 but adc's are still considered a as a type of modulation device.
If it wasn't, it would not have rf spirous emissions that they have  to test.
Part 15 was revised two years ago. To go along with the new classifications that is in a transitional change for converters with the European reclassification:
 
 
 
Emissions: EN55032 (CISPR32)
For emissions, EN55032 replaces EN55022 (IT equipment), EN55013 (broadcast receivers and associated equipment) and EN55103 (audio & studio equipment).  EN55032 was approved in December 2012 and consists of CISPR32+A1+A2. It has officially been added to the OJ – meaning it can now be used by manufacturers – with a transition period until March 2017. 
Apart from the power disturbance measurement, which is no longer permitted, all the other limit values and measurement procedures remain broadly unchanged.
 
source: Metlab regulatory certification lab:  http://www.metlabs.com/emc/en55032-replacing-en55022-and-others-for-ce-marking-of-multimedia-equipment/
 
 
I still doubt those DACs on ebay are certified.

 

[pinnahertz]

  its part 15 but adc's are still considered a as a type of modulation device.
If it wasn't, it would not have rf spirous emissions that they have  to test.
Part 15 was revised two years ago. To go along with the new classifications that is in a transitional change for converters with the European reclassification:
 
 
 
Emissions: EN55032 (CISPR32)
For emissions, EN55032 replaces EN55022 (IT equipment), EN55013 (broadcast receivers and associated equipment) and EN55103 (audio & studio equipment).  EN55032 was approved in December 2012 and consists of CISPR32+A1+A2. It has officially been added to the OJ – meaning it can now be used by manufacturers – with a transition period until March 2017. 
Apart from the power disturbance measurement, which is no longer permitted, all the other limit values and measurement procedures remain broadly unchanged.
 
source: Metlab regulatory certification lab:  http://www.metlabs.com/emc/en55032-replacing-en55022-and-others-for-ce-marking-of-multimedia-equipment/
 
 
I still doubt those DACs on ebay are certified.

"its part 15 but adc's are still considered a as a type of modulation device."

Please cite a reference. 
 
"If it wasn't, it would not have rf spirous emissions that they have  to test."
 
You can have modulation without RF, and RF without modulation.  If you have modulation without an RF carrier, nobody cares because it won't radiate.   The issue with Part 15 is the RF radiation, not the modulation, that is of concern.   Equipment is certified based on radiation.  And the word you are searching for is "spurious".
 
"Part 15 was revised two years ago. To go along with the new classifications that is in a transitional change for converters with the European reclassification.."

I excerpted the current version.  Point?
 
 
"I still doubt those DACs on ebay are certified."
 
Who cares if they are? Your information is still wrong.

 

[drtechno]

looks like they revised it in 2014. and moved the converters from #15 class A to class B.
 
but it still has to get tested as the exempt status § 15.103(h) won't cover it as it is plugged into the wall instead of running off of batteries.
 
Also I notice that if the production is less than five units at a time, its exempt. So this is a new provision was added for the diy sector.
 
 
http://www.metlabs.com/emc/some-electronic-devices-are-exempt-from-fcc-emc-testing/
 
here is the detail on the exemption:
http://www.gpo.gov/fdsys/pkg/CFR-2002-title47-vol1/pdf/CFR-2002-title47-vol1-sec15-103.pdf

 

[drtechno]

Spirous noise is there because its in all signal domain changes weather it is to an RF, IF or digital.
This was covered for me in year 4 of EE school. about 22 years ago.
I do see data sheets mentioning it. So the laws of physics hasn't changed.