2
\$\begingroup\$

can you help me with the basic comprehension of this Voltage Controlled Oscillator scheme (it was used in a PLL)?

enter image description here

I do not understand why there are two varactors instead of one, and the role of the variable capacitance CT.

Reference: http://mwl.diet.uniroma1.it/people/pisa/RFELSYS/L03_VCO_PLL_Phase%20noise.pdf

\$\endgroup\$
0

3 Answers 3

2
\$\begingroup\$

There is no VCO 'reason' to have two varactors, they are functionally in parallel.

We therefore need to guess why the system found it convenient to provide two varactors.

The labelling of the tune lines to the varactors gives us a clue. \$V_T\$ is probably a 'tune' line, used to set the varactor to a coarse setting, static, as there is no t parameter. \$e_v(t)\$ is probably an error voltage back from the PLL, varying in time.

The relative weight of these two varactors can be set by the values of \$C_{D1}\$ and \$C_{D2}\$. A value larger than the varactor capacitance allows more or less the full varactor variation to affect the loop. A value much small than the varactor value attenuates the effective varactor swing at the PLL.

Why make \$C_T\$ a manually adjusted variable? It might be for a coarse pre-setting of the centre frequency of the VCO. However that type of VCO needs a particular range of capacitor ratios to actually oscillate. Perhaps the FET used has large enough parasitic variations that each one needs to be tuned after manufacture.

Noise in a VCO is a parameter that's particularly difficult to guess at from just a circuit diagram. There's a wide range of applications for PLLs, some of which require exquisitely low noise, others of which don't. That might have driven the separation of the varactors, rather than use a single varactor with a composite drive, and a manual adjuster control to 'peak up' the VCO, who knows?

\$\endgroup\$
1
\$\begingroup\$

Having two varactors instead of only one is probably just a design choice for this particular VCO.

Reasons to have two inputs can be that one input is used for the PLL, meaning it is used for the frequency control loop to set the average frequency of the VCO. The other input could then be used to modulate the frequency, for example with an audio signal. Then you'd have an FM signal generator where the average transmit frequency is accurately controlled by the PLL but FM modulation is still possible. The PLL then simply has to be slow enough not to suppress the modulation signal.

So having two VCO inputs isn't needed by the VCO itself, you could connect the two inputs together and treat it as one input. Another option is to remove one of the two input altogether, that would then limit the frequency range of the VCO. However that can easily by restored by using two varicaps (VAR1, VAR2) in parallel.

CT is there to trim the VCO to the right frequency. There are always some unknown (parasitic) capacitances present and CT allows to compensate for those. Normal procedure is to set CT "in the middle" then lock the PLL, that will result in a certain control voltage across the varicaps. Then adjust CT such that the DC voltage across the varicap is as required (for example in the middle of its usable range).

\$\endgroup\$
1
\$\begingroup\$

Having examined and implemented various synthesizer designs(some as discrete, some on silicon), and in some designs I needed to separate the Integral path from the Proportional path, to achieve vastly different Hz/volt requirements, this dual-VCO could provide that flexibility.

Note the Cd1 and Cd2 may have dramatically different values, thus providing dramatically different tuning effects of the two varactors.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.