We've built a current starved voltage controlled oscillatorCurrent Starved Oscillator

And tuned it so it works only in the linear region, which means it oscillations between 720-840 mV, and outputs 3.6-4.4 GHz. In order to calculate the SQNR, we need the ABW. How is it possible to find the analog bandwidth of the Vco?

I've tried taking the output of the ring oscillator and using a fourier transform of it, and we get a series of pulses at different amplitudes from 25Mhz to 675 Mhz. Do I need to take the two 3dB values from here?


  • 1
    \$\begingroup\$ Taking the output of the oscillator will just show you the harmonic content of that signal. The analog BW is (I think) up to what frequency of the input signal (at VCO_IN) the oscillator can follow. I would plot the output frequency for different input frequencies. For low frequencies the output freq. should follow. Like 1 MHz at VCO_IN should result in a 1 MHz (or 1 us repeating pattern) variation on the output. For high input freqs. it will not be able to follow. \$\endgroup\$ – Bimpelrekkie Aug 30 '18 at 14:19
  • \$\begingroup\$ @Bimpelrekkie thank you for the response! Does it matter what the input voltage is? Should it be from 0 - 2.5V? \$\endgroup\$ – RonaldB Aug 30 '18 at 14:33
  • 1
    \$\begingroup\$ Yes it matters as you want to keep variations small otherwise non linear effects come into play. I would use 1 V DC and then a 100 mV (sinusoidal) variation on top of that. \$\endgroup\$ – Bimpelrekkie Aug 30 '18 at 14:35
  • \$\begingroup\$ How did you get that 1MHZ should lead in a 1us repeating pattern? \$\endgroup\$ – RonaldB Aug 30 '18 at 19:46
  • \$\begingroup\$ Uhm because f = 1 / T so a 1 MHz sinusoid has a period of 1 us. How is that not obvious? \$\endgroup\$ – Bimpelrekkie Aug 30 '18 at 21:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.