Sweep a large range of frequencies in one experiment with Phaser board

jqt3

Hello everyone,

I am trying to sweep through a range of frequencies from 1.475 GHz to 4.275 GHz in one experiment. At the moment I can go from +-200 MHz of 2.875 GHz (so I am able to get frequencies between 2.675 GHz to 3.075 GHz with the set_duc_frequency() function), and an additional +-10 MHz from the set_frequency() function in the oscillator class.

From a previous post I made some time ago, I learned that I can change the base frequency or f_trf, by changing the tx_div_sel parameter to either 0-3 in the device_db.py file. I have the upconverter variant and originally, my base is 2.875 GHz, so when I changed the tx_div_sel to either 1 to 3, the base frequency changes to 1437.5, 718.75, 359.375 MHz respectively.

However, when I measure the frequency coming out of the phaser board when I changed the tx_div_sel to either 1 to 3, the base frequency doesn't change from 2.875 GHz until I restart my artiq_master/dashboard. Additionally, when I change the output frequency with set_duc_frequency(), the highest frequency I can get is 2.875 GHz+210 MHz, which is not enough to get to the 4.275 GHz I need.

Is there a way to sweep through 1.475 GHz to 4.275 GHz in one experiment? The frequency step size I plan to use is probably 1 to 20 MHz.

I've been reading the trf372017 datasheet and I see that the register with tx_div_sel is Register 6, bit24, and was thinking of using the trf_write() and trf_read() functions under the Phaser channel class. Would this be the work around instead of modifying the device_db.py file? If so, the other issue still stands, where I am not sure how to get all the frequencies in between 1.475 GHz to 4.275 GHz.

Thanks

claytonho

jqt3
Yes, the only way to sweep such a large range within an experiment is to directly write to the TRF registers.
However, a lot more parameters need to be changed: pll_div_sel, rdiv, nint, prsc_sel, etc. Section 7.3.2 of the TRF datasheet provides some examples of how to set the TRF output range.
We use the upconverter at low frequency (₈₀ MHz) range, so my suggestions might be different for you, but my experience was that a lot of parameters need to be played with, e.g. tx_div_bias & lo_div_bias, to really make the output happy. Without those changes, you won't get the right amount of power where you want it and there'll be spurs everywhere.
I don't recall ever finding any helpful guides that issue step by step instructions, and the vibe I got was that figuring out the TRF can only be done by playing with it and using some black magic; the issues section on the sinara phaser github was the closest I got to a guide.
One important note: there's going to be a lot of spurs in the output spectrum, especially since you require such a large range (e.g. use at 1.475 GHz will cause spurs at 3.6 GHz, and vice versa). You may have to look into tunable filters or come up with a jank solution.

jqt3

claytonho When writing to the registers, are there any other parameters besides lo_div_sel, pll_div_sel, rdiv, nint, prsc_sel, and tx_div_sel that we have to change? I was able to successfully change the output frequency by setting those 6 values in the device_db.py file, but when I tried the same thing but using the trf_write() function, I don't get the same output.
For instance, when I wanted to change the output frequency to 1655 MHz, I had these settings in my device_db.py, which worked:

"trf0": {
    "lo_div_sel": 0b01,
    "pll_div_sel": 0b01,
    "rdiv": 25,
    "nint": 331,
    "prsc_sel": 1,
    "tx_div_sel": 0b01
}

But when using the trf_write() function, and using those same parameter values to output 1655 MHz, I get an output of 1.22 GHz instead. As of now I only write to trf registers 1, 2, and 6.

jqt3

Hi @claytonho thanks for your response - I was able to figure out how to calculate the parameters to output whatever frequency I like.