Hi. I want to use ADF4159 to quickly sweep a frequency from 3GHz to 10 GHz, with 50MHz step. What is the quickest sweep time possible with AFD4159? Thanks for help.
Generally speaking, the speed at which the loop can ramp/sweep is limited by the loop filter bandwidth (LBW). Internally, the ADF4159 can change output almost instantly; the issue is the amount of time (lock-time) it takes for the loop (ADF4159 + filter + VCO) to follow the output. A wide LBW allows for fast lock-time, as the expense of phase noise. A narrow LBW slows down lock-time, but also reduces phase noise.
We provide a free simulation tool which allows you to simulate an ADF4159 loop and optimize lock-time versus phase noise. I have created a tutorial video using the tool which will give you a good introduction. The easiest thing is to watch the video and then come back to me if you have any questions.
Free download of ADIsimPLL and basic usage video: www.analog.com/adisimpll
3 GHz to 10 GHz is very wide. See Bob's question above.
sweep time = (number of steps) * (settling time of the loop filter + programming time/step)
if you program each step individually, though the settling time of the loop filter will dominate.
An even better question is: Do you have a VCO that will tune from 3 GHz to 10 GHz?
Many thanks for answer. I'm relatively new to the topic, so my questions might be a bit silly.
Anyway, I have more general question. I want to sweep VCO, say 3-4ghz, but it does not have to be linear like in FMCW. My main concern is the speed, and my frequency steps will be relatively sparse. Which solution would you advise to use? ADF4158, 4159 or maybe something else?
The ADF4158 is better for your system because it covers your operating frequencies and it is released (unlike the ADF4159). If you look at pages 26-30 of the ADF4158 data sheet, you will see the variety of waveforms it can produce.
Please watch the video above to see how to optimize ramp time. Let me know if you've any questions.
Great, thanks for advice. I'll definitely watch the video.
Just one more silly question. I've seen some guys using DDS and multiplication only for sweeping. E.g. get AD9914 to generate 500MHz-1GHz sweep, and multiply this directly up to 4-8GHz band. Is that something you would recommend? Can this be any quicker that using ADF4158+VCO ? Many thanks for help.
I'm not familiar with the DDS parts. You will have to ask on the DDS forum, or maybe one of the DDS guys will see this.
I believe either solution can work. As far as I know, PLL + VCO solution is generally lower power and cost.
as it is mentioned above, DDS solution will be more expensive and its spur performance will be worse. but it will be faster.
you can check those casts for comparison,
Fundamentals of Frequency Synthesis, Part 1: Phased Locked Loops
Fundamentals of Frequency Synthesis, Part 2: Direct Digital Synthesis (DDS)
if u want to seewp .5-1G with dds you will also need a ref clock higher than 2G practically 2.5G, but some ADI DDS`s may have internal REF multiplier inside them
another method is DAS solution. performance of DAS solution will be the best but it is expensive and complicated thus requires time and budget.
i would use adf4158 cause it has a good ramp feature as mentioned on the above post, if it can meet your seep time requirements.
Retrieving data ...