3.12 VCO Measurement

3.12 VCO Measurement


This document gives some brief and basic procedures and tips on VCO characterization and is divided into the following sections:
Common VCO Specifications
VCO Measurement
Equipment
Calibration
Tuning Range, Output Power and Linearity
Spur Measurement
Frequency Pushing
Phase Noise
Other Measurements
General VCO Measurement Procedure
Test Setup Considerations

References

 

Common VCO Specifications:

  • Frequency tuning characteristic
  • Output power
  • Linearity
  • Tuning sensitivity
  • Frequency pushing
  • Spurs
  • Phase noise
  • Frequency drift with temperature
  • Frequency pulling
  • Tuning speed (or response time)

VCO Measurement:

Most of the measurements use the same equipment setup and the various performance metrics can be determined using the same data. Power supplies are required to bias the VCO and a spectrum analyzer is required for measuring VCO output power and frequency. A power splitter may be required to monitor the power separately using a power meter in conjunction with the spectrum analyzer.

A good first test would be to verify that there is a signal-peak within the expected tuning-range under proper biasing and to determine the tuning range.

Equipment:

  • Spectrum analyzer
  • DC power supplies (current mirror biasing, Vcc, tuning voltage)
  • Power splitter (MiniCircuits)
  • Power meter (Rohde Schwarz)
  • Signal source (for calibration purposes)
  • Bias Tees (Picosecond Pulse Labs)
  • Cables (SMA for the RF, BNC for DC)
  • Various adapters ( Male-Male, Male-Female, etc)
  • Torque wrench

Calibration:

Keep in mind that you will have to calibrate for the effects of the cables if you are concerned about output power. The cables and adapters contribute some attenuation. With a signal generator connected to the input, you can measure the attenuation in the cables on a spectrum analyzer. When using a signal generator, the actual power is known by connecting the signal generator directly to a power meter and reading the value. A subsequent measurement is taken with the RF cable inserted between the signal generator and power meter and the attenuation becomes the difference between the two readings. You can correct for the attenuation simply by setting the offset-value on the spectrum analyzer. The same measurements can be performed using a spectrum analyzer instead of the power meter. With a spectrum analyzer, keep in mind that relative accuracy is better than absolute accuracy.

If a power splitter is used, its insertion loss must also be accounted for. See the calibration section under nonlinear measurements for the procedure.

Tuning Range, Output Power and Linearity:

The frequency of the signal as a function of tuning voltage is measured. The power of the output peak at different tuning voltages can also be measured here, as well as the linearity of the frequency of the output signal as a function of tuning voltage. In addition, the tuning sensitivity can be calculated. This is the slope of the frequency characteristic which is expressed by the change in frequency divided by the change in voltage.

After the biasing is established, the tuning voltage is set and the frequency is measured using the spectrum analyzer. It is convenient to use the peak-function that exists on the spectrum analyzer. For the best resolution, the frequency range on the spectrum analyzer is set so that it just covers the signal. It is important to be consistent in using the same frequency range, at different tuning voltages. The total output power is measured with the power meter for each tuning voltage and frequency.

Spur Measurement:

When measuring the spurs, it is practical to divide the measurement into two parts, one for the harmonics and the other for the non-harmonics.

The harmonics, if there are any, lie on frequencies which are multiples of the fundamental frequency. Use the frequency range on the spectrum analyzer to find the regions and then zoom in.

Measuring non-harmonics is a little more tricky, since they appear on more or less random frequencies. Start close to the fundamental and then go to a wider band to see if there are any spurs at other frequencies.

Frequency Pushing:

Measuring the pushing involves varying the supply voltage and repeating the tuning measurement to determine the effect of supply voltage on VCO frequency. One of the parts for this measurement has already been completed, i.e., the frequency characteristic at the correct supply voltage. Additional measurements can be made at higher (Vcc + 0.5V) and lower (Vcc - 0.5V) supply voltages.

Phase Noise:

Phase noise is specified in terms of dBc or dB relative to a carrier and is displayed only when the signal is far enough above the system noise floor.

Using the delta-function on the spectrum analyzer, the phase noise can be determined by measuring the difference between the main signal peak and the noise at a certain offset frequency.

Other Measurements:
  • Temperature Drift
  • Frequency Pulling
  • Tuning Speed
To complete the characterization of the VCO, measurements of temperature-drift and pulling can be performed, if the equipment is available. Frequency pulling, defined as VCO frequency as a function of change in output-load requires special equipment which allows control of load impedance. In addition, tuning speed (or response time) commonly defined as the time it takes for the output-signal to reach 90% of its final value, when a step change in frequency has been made.

General VCO Measurement Procedure:

The following general procedure can be used to measure your VCO:
  1. Allow equipment to warm up 1/2 hour before measurement.
  2. Calibrate for the loss in the cables and adapters.
  3. Ensure proper biasing and proper ground connection on all supplies. Perform a functional test.
  4. Start with the measurement of tuning range. Before you start decide the frequency range and reference level. They should both be set to values so that the entire tuning range can be measured without requiring further adjustments to the spectrum analyzer.
  5. Using the same setup, perform the pushing measurement.
  6. Measure spurs (non-harmonics). Start with those that are close to the main signal. Use the narrowest span that is possible without losing any spurs. The "Res BW" can be lowered too. After this is done check for other non-harmonic spurs using a wider band.
  7. Measure the harmonics.
  8. Perform phase noise measurement.
  9. Continue with a new chip.
Test Setup Considerations:
  • If both the spectrum analyzer and the power meter are used simultaneously in the measurement, the output signal needs to be divided using a power splitter. Keep this in mind when reading the output power. The type of power splitter will determine, the ratio of the power split.
  • Always ground only in one point. Otherwise loops are created which could add interference via inductive coupling.
  • Try to attach loose cable to something solid, e.g., the probing station.
  • To reduce the noise introduced in the measurement via the power supply, a decoupling capacitor (e.g., electrolytic of about 10uF can be used or a smaller capacitor for high frequency interference) can be inserted between the powerline and ground. Also an inductive choke in series with the cable can be added. If the external noise interference affects the measurement, then the decoupler and choke do not make any noticeable improvement.
  • If your VCO has a differential output but you do not use a balun (on-chip or off-chip) to convert to a single-ended output, both outputs must be biased and loaded with 50 ohms even if only one output is measured.
  • Use a multimeter to monitor current.
Some useful spectrum analyzer functions:

Video Average: Takes X number of measured values and makes an average of these. X can be chosen by the user.
Max Hold: Displays the largest measured power.

Deciding which function to use, depends on the signal characteristics. If the signal is very unstable in frequency then the "Max Hold" is preferable. Since the frequency is changing, an average power at a given frequency would not be accurate in this case. When a signal fluctuates in power (not frequency) then the "Video Average" can be helpful to stabilize the display and make the reading easier. No function is required if the signal is very stable.

References:

  • Hewlett-Packard Company, 1997 Back to Basics Seminar (HP publication number 5965-7008E)
  • http://www.tmo.hp.com/tmo/iia/edcorner/English/sabas.html
  • Hewlett-Packard Company, Spectrum Analysis (HP Application Note 150)
  • Hewlett-Packard Company, Effective Test Methods for Today's RF Devices (HP Publication Number 5963-5191E)
  • Hewlett-Packard Company, Automatic Characterization of Microwave VCO's, HP Application Note 377-2 (HP Publication Number 5952-7988)
  • MiniCircuits, RF/IF Designer's Handbook
  • MiniCircuits, VCO Designer's Handbook http://www.minicircuits.com/appnote/application.htm
  • I. Bahl, P. Bartia, Microwave Solid State Circuit Design, John Wiley & Sons, Inc. 1988, pp.427-481
  • Aeroflex, Phase Noise Theory and Measurement Application Note #1
  • Hewlett-Packard www.hp.com
  • RFGlobalnet www.rfglobalnet.com

Back to Test Guide Table of Contents