I'm looking at using the ADCLK905 to provide a differential PECL clock for the AD9914- wondering if I can give it a sinewave as per the single-ended example Fig 29 and what minimum amplitude it would require at the input?
The minimum single ended input Vpp voltage swing is 200mV to guarantee that the part functions. The output phase noise performance and propagation delay is going to be limited by your input slew rate (the faster the better). The output of the ADCLK905 will then limit your performance on the AD9914.
What kind of performance do you require from the AD9914?
This is one of those cases where the sky's the limit. We are going for the best absolute phase noise we can get from the AD9914 in the frequency range from 0.5 to 42MHz, which I realize is low for the 9914, but we also need the same module to be able to do PSK at 900MHz.
Currently we are planning to clock the 9914 from a low noise OCXO like the Sprinter series, multiplied up from 100MHz to 3.5GHz to give us maximum flexibility from the DDS. I'm trying to figure out the final output stage of the multiplier, which will produce a clean fundamental at 3.5GHz. Any tips?
For the best performance I would suggest skipping the ADCLK905 all together and run the 100MHz OCXO into the AD9525 (Analog PLL) with an external 3.5GHz VCXO and skip the internal PLL of the AD9914.
You will also be best off if you can run the DDS at a high output frequency and then divide down the output externally to increase slew rate.
Feel free to email me at firstname.lastname@example.org. If you are able to provide a basic block diagram of your system I should be able to give you some more specific tradeoffs.
We tested AD9914 with single ended REF CLK. You can generate 3.5GHz by simply multiplying 100MHz using low phase noise multipliers (Schottky diode) or NLTL multipliers. We have noticed that for low power REF CLK, DDS will work for even -10dBm (0.07V) but spurious will be more and for proper operation REF CLK amplitude should be more than 10dBm (0.7V). There are some multipliers from Wenzel and also low phase noise amplifiers, may be you can have a look to there products.
I have a similar question:
In my design, I have a AD9915 which would be clocked by a ADF4351 whose output is a sine wave. Would i gain anything from buffering the signal using the ADCLK905 as in the evaluation board?
Retrieving data ...