TI E2E Community
High Speed Data Converters
High Speed Data Converters Forum
A couple of general questions about the ADC12D1600 data sheet. I am looking at the (old) National data sheet dated Oct 17 2011 and with a TI cover page, Literature Number SNAS480K.
On page 41, in the bottom row, right side of page is a graph "NPRvs. RMS NOISE LOADING LEVEL (ADC12D1600)". The horizontal axis is labeled "Vrms LOADING LEVEL (dB)". What is "Vrms LOADING LEVEL" I do not find any definition of the term in the data sheet. What is the relationship to the "V in_FSR" specification on page 20?
Is there a list of high speed converter specification definitions?
The datasheet does contain a general section of definitions: look for the Specification Definitions. This section does contain a simple definition of Noise Power Ratio (NPR), but it looks like you are looking for more details. Please see the attached presentation for a more detailed overview of NPR on the GSPS ADCs.
5775.ADC1xD1x00 NPR for E2E.pdf
Let me know if you have further questions.
Thanks for the response and the information on NPR testing. My question concerns how to interpert an inportant technical specification: RMS LOADING LEVEL. I tried to attacha screenchot to this email , but looks like attachments are not allowed. I have re-stated my question. Any information would be appreciated
Here is a screen shot of the TI (former National) ADC12D1600 high speed A/D, page 1 and page 41.
At the bottom of page 41 is a graph of NPR vs “RMS LOADING LEVEL (dB). My question: the X axis is labeled “RMS LOADING LEVEL (dB)”. “dB” in this case is a relative term. What is the reference for the RMS LOADING LEVEL ? Is it relative to one of the Table 8 (Page 20) analog parameters? If so, which one and shat is the power of the input signal in dBm. If the input signal is stated in volts peak to peak, what is the impedance (50 ohm, 100 ohm??) so that the input signal power can be calculated?
Thanks for a quick response.
I think I understand your question this time... let me know this response answers your actual question:
First, let us consider the power in a full-scale sine wave. The power is:
P [Watts] = [ (Vpp/2)^2 ] / [2 * Rin, diff ]
P [dBm] = 10 * log10( P / 0.001 )
For this example, Vpp = 0.8V; Rin, diff = 100 ohm. Therefore,
P = -0.97 dBm
What does this equate to in dBFS? This is a bit of a debate... since this is the maximum amplitude sine wave which can be input to the ADC without clipping, some folks like to label this '0dBFS' for a sine wave. However, it is actually -3dBFS since the rms power which is in a sine wave is 3dB less than that in a full-scale DC signal.
Now, let's examine the RMS noise loading level. From p.9 of the pdf document:
2 * Vo = 0.8 Vpp
Vo = 0.4 Vp
Assuming the input is a FS sine wave, sigma = 0.8 / 2 / sqrt(2) = 0.28
k = Vo / sigma = 1.414
1/k = 0.707
RMS noise loading level = 20 * log10(k) = -3 dBFS
So, the RMS noise loading level is in units of dBFS, with the understanding that a full-scale sine wave is -3dBFS, not 0dBFS, as it is sometimes considered to be for an ADC.
Thanks for the additional information. I am beginning to see the relationships, just a couple of additional comments if you do not mind.
So, for the NPR Vs RMS NOISE LOADING LEVEL graph the X axis is labeled “Vrms LOADING LEVEL (dB)” represents a DC voltage applied to the converter input? The X axis “VRMS LOADING LEVEL” value of 0 dB would correspond to -3dBFS for a sine wave (according to your explanation, more on that later). From the graph, the peak NPR occurs at a X axis VRMS LOADING LEVEL of -12 dB and a corresponding NPR value of ~ 49 dB. So the equivalent power in the sine wave to produce max NPR is 15 dB below the sine wave value to produce FS.
In your example, a sine wave of 0.8 Vpp has a RMS power of -0.97 dBm, corresponding to -3dBFS. The sine wave input to produce max NPR would have a power of -0.97 dBm – 15dB = -15.97 dBm. Corresponding to 0.025 millwatt , Vrms = 0.503 millivolt rms or 1.4 millVolt pp.
You made the statement, “”However, it is actually -3dBFS since the rms power which is in a sine wave is 3dB less than that in a full-scale DC signal”. The definition of RMS voltage for a sine wave is “ a 1 volt RMS sine wave has the identical heating power as 1 volt DC.” 1 volt rms sine wave applied to a 1 ohm resistance produces 1 watt of heat (or power) and 1volt DC applied to 1 ohm resistance produces 1 watt of heat (or power). Why the 3dB difference in your statement? Is it in the definition of “… a full scale DC signal”?
Your comments are appreciated.
One non technical question about the forum. Is there a way to have the forum notify me by email when a reply is posted? I thought I set that up, but I do not seem to be receiving e-mail notices of your posts.
Thanks again for your time
I apologize for a tardy reply...
It looks like you are trying to understand what kind of signal is used to test the NPR. To answer that, I can refer to the document which I posted earlier, see especially slides 12 and 13. The input signal is neither a DC signal nor a sine wave. It is broad band noise from DC to the Nyquist frequency, which is notched at a smaller bandwidth. In the case of this measurement, the notched noise signal is generated by an ARB which produces a multi-tone output to simulate the noise; where the tones are absent becomes the notch. Does this answer your question?
I'm not sure about how to get notified when an answer is posted - I thought that would happen automatically. I have signed up to follow the high-speed data converter forum, so I get notifications about every new post in that one.
Thank you for the reply. A little bit of housekeeping first ---- you posted on May 31. I received notice of the post on June 5, apparently after "JimBrinkhurst84999 has marked your post RE: ADC12D1600 in the High Speed Data Converters Forum as answered" Do not know what that really means, but it appears that need to veryify all of the posts.
Your answer comes a little closer. Here are some additional details about my question. We do testing of ADC's using both a CW sign wave, continuous noise (AWGN) and notched noise for NPR tests. I an familiar with the notched noise NPR set up - Aglient has a signal generator that can produce true notched AWGN rather than using tones. We use swept sign wave input to measure frequency response and two or three sign wave to measure distortion. The question is " What RMS power level (as measured by a true RMS power meter) is to be used for the different input signals, so that the ADC is accurately characterized? Your example of notched noise mentioned DC to the Nyquest frequency. For illustration purposes, lets say the RSM power in your signal is -10 dBm, and this signal is applied to the ADC input. Could a -10 dBm sign wave at 1/2 the Nyquest frequency be used to measure frequency response, or two tones, each tone -13 dBm at 1/2 Nyquest +/- 5 MHz, be used to measure two tone distortion? Or would some correction factor, say the -3dB(FS) factor you mentioned in an earlier post on 27 April need to be applied? The object is to have the three different test signals (RMS power?) operate the ADC input at the same point so that the performance is accurately characterized.
Thanks for your time.
Thanks for the chat on Friday - I think it was good to speak in person as sometimes writing emails just doesn't get the message across! Let me know if you have further questions.
All content and materials on this site are provided "as is". TI and its respective suppliers and providers of content make no representations about the suitability of these materials for any purpose and disclaim all warranties and conditions with regard to these materials, including but not limited to all implied warranties and conditions of merchantability, fitness for a particular purpose, title and non-infringement of any third party intellectual property right. TI and its respective suppliers and providers of content make no representations about the suitability of these materials for any purpose and disclaim all warranties and conditions with respect to these materials. No license, either express or implied, by estoppel or otherwise, is granted by TI. Use of the information on this site may require a license from a third party, or a license from TI.
TI is a global semiconductor design and manufacturing company. Innovate with 100,000+ analog ICs andembedded processors, along with software, tools and the industry’s largest sales/support staff.