The function generator source used in the RIPS is MAX038 function generator IC powered by a ±5 V supply. With the bipolar supply, the chip will output a 1 V peak signal centered on 0 VDC. The device has the ability of delivering a sine wave, square wave or triangle waveform over a wide frequency range. Course frequency set is effected by connecting an appropriate value of capacitance from pin 5 to ground. Using capacitor decade values starting from 200 pF, decade-stepped frequency bands are possible between 1 Hz and 1 MHz. The current into pin 10 provides further adjustment of the frequency. In the RIPS circuit, the combination of reference voltage at pin 1 and the front-panel potentiometer is the current source that sets the frequency within the decade. Voltage levels on pins 3 and 4 determine whether the output waveform is sinusoidal, triangular or square wave.

Surprisingly, minimal troubleshooting was necessary during the appraisal of the RIPS' performance. The following is a summary of the bench data from the evaluation.

Frequency range:

DC output:

• Minimum DC voltage: 17 mV
• Maximum DC voltage: 19.94 V
• Maximum DC current: 1.93 amps

Frequency versus output amplitude response with an OCXO load connected by 12 inches of power cabling:

Measured voltage with 1 kHz ripple, 10 Ohm resistive load,
VDC = 10 V:

Measured voltage with no ripple, 10 Ohms resistive load,
VDC = 10 V:

Measured voltage with 1 MHz ripple, OCXO load,
VDC = 8 V:

The peak amplitude target of 1 V has been met for modulating frequencies up to roughly 300 kHz. Under light resistive loads, the modulating frequency extends up to 2 MHz with one volt peak amplitude. Under moderate reactive loads, such as those encountered with typical oscillators, the range is decreased to about 1 MHz at an amplitude of about 100 mVp-p. The load reactance largely influences the maximum usable frequency and amplitude. At 5 VDC and a 10 Ohm resistive load with no cabling, all three wave shapes appear undistorted up to about 250 kHz. The corresponding sine wave amplitude begins to roll off at about 1.4 MHz.

Load regulation is an indicator of how much the output voltage will change for a change in the load current and is calculated using equation 3. The RIPS regulation with a typical stabilized OCXO load measured 0% within the resolution of the voltmeter. For a fairly heavy load of 1 amp, the regulation degraded to 0.3% with or without ripple modulation.

Line regulation is the ratio of the output voltage change for a change in the input voltage as expressed in equation 4. A modulated output amplitude change was not detected for an input range of 100 Vrms to 130 Vrms. It was noted that the ripple frequency changed a total of 20 Hz for this same input range from a nominal of 10 kHz.

In another test, the output of the RIPS was capacitively coupled into an HP3561 signal analyzer to determine 1 kHz ripple harmonic content versus amplitude. The results of this test are shown in Figures 7, 8, and 9. Notice that disabling the ripple modulation at the front panel does not deactivate the function generator, nevertheless 100 dB of suppression should not pose any problems.

As observed, the 60 Hz line spurs produced by the RIPS are a bit higher compared to what the off-the-shelf supply generates. Keep in mind, however, that filtering at the RIPS output is virtually nonexistent so that the bandwidth can extend up to 2 MHz.

The spur level of an OCXO was characterized using several RIPS frequencies and magnitudes. These plots are shown in Figures 10, 11, 12, and 13. Notice the increase in harmonic content as the ripple modulation amplitude increases. The harmonics of the injected sinusoidal ripple result from two processes. Because of internal nonlinearities, the supply itself generates a certain level of harmonics as a function of the output amplitude for sinusoidal ripple modulation. Second, the modulated oscillator by virtue of its operation is nonlinear and results in its own harmonic generation.

On the whole, the performance goals of the RIPS have been met. As demonstrated by practical use in the lab, this supply will be a useful tool in the characterization of oscillator circuits.

1. Smith, Jack, “Modern Communication Circuits,” pp. 407-408.

2. Lance, Algie L., Seal, Wendell, D., Labaar, Frederick, “Phase Noise and AM Noise Measurements in the Frequency Domain,” “Characterization of Clocks and Oscillators” NIST Technical Note 1337, pp. TN-199 TN-200.

3. Fernandes, Stephen, Brooks, Peter, “Power Supplies” section of the “Handbook of Electronics Calculations for Engineers and Technicians,” p. 10-3.

Robert F. Smith spent many hours breadboarding, assembling and testing the RIPS unit. Robert W. Camp authored the firmware for the frequency counter microcontroller. I am grateful to both for their effort and excellent work.

Mike F. Wacker is a senior design engineer at Vectron International's Mount facility in Holly Springs, PA. He has worked at Vectron since 1989, and has over 21 years of experience in the frequency control field specializing in ovenized oscillator design. An author of numerous published trade articles and application notes, Wacker's recent research focuses on oscillator modeling and low noise OCXOs. He holds a Bachelors of Science in Electrical Engineering with honors from the University of Central Florida.

Mike F. Wacker is a senior design engineer at Vectron International's facility in Mount Holly Springs, PA. He has worked at Vectron since 1989, and has more than 21 years of experience in the frequency control field specializing in ovenized oscillator design. An author of numerous published trade articles and application notes, Wacker's recent research focuses on oscillator modeling and low noise OCXOs. He holds a Bachelors of Science in Electrical Engineering with honors from the University of Central Florida.

Click here for the enhanced PDF version of this article

Previous 1 2 3

RSS    Save to Del.icio.us  Digg This