We are planning to use CDCM7005 in our design.
1. 5 Clock outputs (Fout = 102.4MHz) to drive 4 ADCs and External reference.
2. We are planning to use ADC AD9467 (Analog devices). The ADC supports 75dBc SNR (Input 70+/- 10MHz / 1.5MHz to 30MHz , Sampled at 102.4MHz) .
3. CDCM7005 should lock to an External OCXO (Reference Clock) with sine wave 10MHz (Phase noise -130dBc/Hz @10KHz and higher) which is yet to be selected.
1. Will the CDCM7005 will provide sufficient low jitter clock to meet the SNR spec of the ADC (RMS clcok jitter should be less than approximately 350fS. Correct me if I am wrong).
2. Since the OCXO is yet to be selected. How to derive the jitter requirement of the OCXO from the design requirement for the CDCM7005 (RMS Jitter < 350fS)
3. Is there any frequency planner available for CDCM7005 (Such as one available for CDCM61001 http://e2e.ti.com/support/clocks/m/videos__files/576535.aspx ). I want to derive the VCXO frequency for the 10MHz reference and 102.4MHz output with the possible M, N & P divider values.
4. We are planning to use the VCXO (http://www.silabs.com/Support%20Documents/TechnicalDocs/si570.pdf). Please let me know how can I estimate the phase noise of CDCM7005 output from VCXO and OCXO phase noise information. I refereed the http://e2e.ti.com/support/clocks/f/48/t/153019.aspx. But unable to find out where to give the phase noise informations of the VCXO and OCXO to get the resultant phase noise of the CDCM7005.
thank you for your interest in our device. Please find below the answer to your questions:
1) the jitter requirement for the ADC depends on the maximum analog input frequency which is sampled, the target SNR and the aperture jitter of the ADC itself. For example if you want to sample a 50MHz input signal with an SNR of 78dB with the AD9467 (60 ps typ aperture jitter) your clock jitter does not have to exceed 390fs rms. If you consider some margin due to the noise of the analog signal (which also degrades the overall SNR), 350fs should be your targeted rms jitter. At this point, I would give you another advice about the reference source. A sinewave clock signal is more sensitive to noise than a square wave one. If you like to stick with a sinewave clock source, you should have at least a high slew-rate signal. Otherwise a clipped sine wave or square wave source is even a more clean source.
2) the CDCM7005 will work as jitter cleaner which behaves like a very lowpass filter. Therefore you should take care of the input reference performance in the frequency range up to the device PLL- bandwidth which you defined. For example if you set the External Loop Filter in such a way that the PLL bandwidth is 100Hz, you should take care of the phase noise of the input reference up to a offset frequency of 100Hz. From that point on, the output clock perfromance will be dominated by the external VCXO you are going to use.
3) In the following screen shot, you can find the divider value you need to set into the device
4) Here some other link you may use for your project. They include same calculator to see the noise contribution from each PLL block and the GUI interface to program the device.
I hope this fulfill your requests.
Thanks for your detailed clarifications on query 1 & 2. Please clarify me further on the 3 & 4
>> 3) In the following screen shot, you can find the divider value you need to set into the device
I tried that "Texas Instruments CDCM7005 SPI Software". How the divider values for the M divider =100, N divider =768, P-divider=16 were acheived by the software (Where we have give the Ref_Clk frequency and the desired Output_Clock as input in the software to calculate the divider values. How do I find the required VCXO frequency for 102.4MHz output clock)
>> 4) Here some other link you may use for your project. They include same calculator to see the noise contribution from each PLL block and the GUI interface to program the >>>device.
I tried the excel version of the tool and I am unable to find where should I include the phase noise information of the VCXO and OCXO (Ref_Clk) to estimate the output phase noise of the CDCM7005. Please clarify.
the current available version of the GUI does not include a frequency planner tool. The VCXO possibilities for a defined output frequency (i.e. fout=102.4MHz) are given by multiplying fout x1, x2, x3, x4, x6, x8, x16. In your case I wanted to suggest you a VCXO frequency of 614.4MHz (which is 102.4M x 6).
The next step would be to calculate the M, P and N values. How to do this? You should define the PFD frequency, which is given by the greatest common divisor between the selected VCXO frequency and the input reference frequency (i.e. 10MHz), which you can get within the available M, P and N value range. In your case the greatest common divisor is 100kHz. In case you have multiple choices I would suggest you to pick up always the highest possible PFD frequency.
Based on the calculated PFD frequency, you can derive the values of M, P and N dividers.
With regards to TI PLL sim, the excel version provides you only an high-level approach to the device, while the Labview version has more features, like the possibilty to perform the noise contribution analysis. As you can see in the attachement, under "Noise Cons." menu you can insert the Inpur reference phase noise levels, the VCXO phase noise and you get an estimation of the output phase noise. In the Filter & PLL Cal. section, you can also tried out different Loop Filters in order to achieve the best performance out of the CDCM7005.
Let me please ask you for which application you want to use this device.
Feel free to ask for further support if needed.
8764.QUAD ADC GENERAL BLOCK - TI Support.pdf
Once again thanks for your detailed clarifications. Currently I dont have access to Labview I am checking for the possibilities to use the labview version of the tool.
Regarding the application we are planning to use the CDCM7005 to clock the Four ADCs. This is a XMC based quad channel ADC card. It may be used for Software defined radio. Here with I have attached the block diagram for your reference.
Since our OCXO (Ref_clock = 10MHz) is a sine wave, we are planning to use a sine to square wave converter (since CDCM7005 doesnt support sine wave input). Please refer http://e2e.ti.com/support/clocks/f/48/t/163080.aspx where I was recommended to use ILMV7219 for sine wave to square wave conversion. I was just wondering device like "CDC3RL02" might have been better suit my application however it just supports 1.8V but CDCM7005 needs 3.3V. Please let me know your suggestion if any to implement the clocking solution for the ADC.
sorry for the delayed answer. The CDC3RL02 could work as Sine-to-Square Wave buffer. Do you have the chance to place a 3.3V supply in your board as well?
I was also wondering about the clock solution you are considering in for your FPGAs (Virtex 5 and Virtex 6). The CDCE62005 could be used for those FPGAs.
Thanks and best regards,
I mean to say that since CDC3RL02 supports only 1.8V output I am unable to use it with CDCM7005 (Since the CDCM7005 requires 3.3V compatible LVCMOS). Any recommendation for the interface between these two chips highly appreciated. Yes, we have on board 3.3V. And as you said we are planning to clock the FPGA with the CDCM7005.
I mean to say that since CDC3RL02 supports only 1.8V output so I am unable to use it with CDCM7005 (Since the CDCM7005 requires 3.3V compatible LVCMOS). Any recommendation for the interface between these two chips highly appreciated. Yes, we have on board 3.3V. And as you said we are planning to clock the FPGA with the CDCM7005.
sorry for the late response. I would suggest you to use the CDCS503. This device has a wide input hysteresis that can make it work also with small slew rate input signal.
I am afraid are you recommending the CDCS503 as an interface between the CDC3RL02 (1.8V LVCMOS output) & CDCM7005 (3.3V LVCMOS input)? Since CDCS503 is operating at 3.3V and its VIH min is 0.7VCC (2.31V). Please clarify.
I meant to use the CDCS503 instead of CDC3RL02. But in this case you should better use a Reference Input of 30MHz and set the M divider in the CDCM7005 to 300. Would this be possible for you?
My reference clock is 10MHz sine wave. Mean while I found a chip that support sine wave input and LVCMOS output.
PL130-07SI (Sine to square wave converter) from phase link
Thanks a lot for your efforts and clarifications to resolve my queries.
Hi, We are using the Jitter cleaner CDCM7005 in our board.We are providing VCXO Ref.input from Si571AEC000107DG & reference clock from external OCXO 501669013 Wenzel associates part 501-09451. We are facing issue in CDCM7005 locking.We want to know whether the VCXO & reference signal are synchronous in their rising edge. What will happen if the VCXO & reference input are 180 degree out of phase.We have got the PDF as 400KHz & kept lock detect window to 18.5ns.We want to know how the rising edge synchronous lock is ensured using our set up,& how to achieve locking?
Thanks & Regards,
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.