PLL board for HP sweep oscillators

[EggertEnjoyer123]

I created a phase lock board based on the LMX2491 to phase lock my sweep oscillator and turn it into a synthesizer. It seems to work with my HP 8620C and HP 86290A (some modification might be required for the 8350).

The design files for everything are attached to this post. The microcontroller is a standard STM32 bluepill, the display is a standard 20x4 with I2C adapter board, and the keypad is just the standard 4x4 one.

There are two modifications that need to be done to the 86290A's FM board. The mods ensure that the FM tuning sensitivity doesn't change when bands are switched. Apart from that, they shouldn't affect operation normally (when the switch is not set to PL). A CN-50 connector from Amazon was used to mate with the connector on the back of the 8620C. The board derives its power from the +20, -10, and +5 supplies on the 8620C.

The loop bandwidth of the PLL is 300kHz and the phase margin is around 70 degrees.

Some issues:
1. The noise floor increases when the sweep generator is not in CW mode. In CW mode, you have to manually set the frequency close enough for the PLL to lock in, while in Ext Input mode, the DAC voltage gets fed in and automatically sets the frequency. I believe this is due to improper shielding, but not exactly sure. In CW mode, I got comparable performance to a commercial PLL module (well as far as my 8566 could measure).
2. Spurs (especially at higher denominators). At 15.000001000 GHz, the fractional spurs are only -25 dBc. There probably is a way to get around this (for example: using two PLLs so the LMX2491 only needs to operate in integer N mode). There's also the possibility of adding dithering, but that just redistributes the noise and doesn't fix anything.
3. My code is half baked and many features aren't implemented yet (like ramp/modulation).
4. I am using a crappy DAC and the datasheet failed to mention it wasn't rail to rail (damn you LCSC). There's a bodge to get around that with some op amps, and you'll have to adjust the op amps if you use a different DAC.
5. I have not tested the external 100MHz reference input option, or the modulation input.
6. There are problems with the board resetting correctly on startup. Adding a 10uF to the microcontroller reset pin seems to work. (This capacitor is not in the design files, so you'll have to add it yourself). The BNC connectors for FM and phase noise
7. The BNC connectors are too close for the FM and main tuning output. I was able to stuff in an SMA connector, though, and there is enough room for those to fit.

Anyways you guys are free to do whatever you want with my design, just keep it open source though.