Test Ideas: Sync sine waves over three decades
This sine-wave generator can synchronize a sine-wave output through three decades of frequency, while maintaining low total harmonic distortion and constant amplitude.
By Alfredo H. Saab and Tina Alikahi, Maxim Integrated Products -- Test & Measurement World, 8/1/2008 2:00:00 AM
The test and calibration of audio-frequency circuits and systems often require a sine wave with an accurate amplitude and frequency and a low total harmonic distortion (THD). Some test applications require that a signal generator have the ability to accurately synchronize the output with an external timing signal.
|
Have a test idea you'd like to share? Publish it here and receive $150. Send us your ideas. Read other Test Ideas. Read Design Ideas published by EDN. |
Simple sine-wave generators can offer varying degrees of performance, but you’ll find it difficult to maintain a low THD with constant amplitude if the sine-wave output must remain locked to a synchronization signal over an extended range of frequencies. The circuit in Figure 1 (see circuit and table below) uses three ICs to generate a low-THD, constant-amplitude sine-wave output synchronized to an external square wave or pulse input signal that spans a three-decade range from 20 Hz to 20 kHz.
The synchronizer IC (74HC4046) is a phase-locked loop (PLL) with one voltage-controlled oscillator (VCO) and three phase/frequency detectors. The best phase/frequency detector to use for maintaining a low THD with constant amplitude is one with a frequency-capture range equal to the VCO frequency range (the maximum frequency minus the minimum frequency). In this case, it is the phase-comparator 2 (PHC 2 out) output.
A 74HC4060 general-purpose binary frequency divider generates at its Q6 output the 74HC4046’s VCO output divided by 64. This divided VCO output feeds back to the 74HC4046’s phase/frequency comparator input (Comp In). When the VCO is phase locked, it synchronizes the pulse input (Sig In) to the divided VCO output. Components C1 and R1 set the VCO’s frequency range from the minimum to the maximum level of the VCO’s input-voltage range.
The MAX297 switched-capacitor low-pass filter that follows the PLL has a cutoff frequency (from analog signal input to analog output) equal to 1/50th of the frequency at its clock input. Any signal with a ratio to the clock frequency lower than 50 is heavily attenuated (Ref. 1). That clock signal is, in this case, the VCO output.
The analog input is the same 74HC4060 Q6 output square wave used for the PLL feedback. Because the clock and signal inputs always have a frequency ratio of 64, the fundamental sine wave of the input signal always falls within the filter bandpass, appearing at the output. No other harmonic component of the input square wave will fall within this bandpass, because the ratio of the clock frequency to frequency is always less than 50.
The fact that the filter’s analog input signal is a 50% duty cycle square wave helps in this application because such a square wave contains only odd harmonics of the fundamental. The lowest-frequency harmonic is then the third, which is well within the filter’s deep-attenuation range (frequency to clock ratio is 21.33 for the 3rd harmonic).
You can frequency-modulate the synchronization signal, but that task forces a compromise between the synchronization-tracking speed (or maximum modulation frequency and depth) and the frequency-locking range. The PLL’s low-pass filter components—R2, R3, and C2—set that range. Modulation speed is limited for the values in the figure because those values are optimized for an extended-frequency locking range.
Figure 1.
This three-IC sine-wave generator, which covers three frequency decades and provides low distortion, can be synchronized with an external signal.
THD and amplitude vs. frequency for circuit in Figure 1
| Frequency (Hz) | THD (%) | Amplitude (VRMS) |
| 20 | 2.775 | 1.470 |
| 50 | 2.650 | 1.472 |
| 100 | 2.525 | 1.472 |
| 200 | 2.250 | 1.473 |
| 500 | 1.002 | 1.473 |
| 1000 | 0.186 | 1.473 |
| 2000 | 0.260 | 1.472 |
| 5000 | 0.330 | 1.473 |
| 10000 | 0.405 | 1.473 |
| 20000 | 0.022 | 1.472 |
| Reference |
|
-
Dear Mr. Whitlocck:
Thank you very much for your attentive reading of our Design Idea. Your guess is correct. The distortion is from the 64th and higher order harmonics of the output (MAX297 clock), against which the design and the layout used for the implementation had no reduction techniques applied.
The instrument used for the THD measurement was an audio range analyzer, and it is its 20kHz LPF which progressively reduces the total THD number as the frequency increases. For what we had in mind at the time, harmonics on the 64th order and above were "don't care" components. We should have indicated both facts in the text, or left the table out.
Best Regards,
Alfredo H. Saab
Alfredo Saab - 2008-21-8 14:09:00 EDT -
Perhaps I'm also missing something. Shouldn't we be using the Q5 output from the 74HC4060 so that the fundemantal frequency of the wanted signal is within the passband of the filter?
F Tham - 2008-20-8 21:12:00 EDT -
Maybe I'm missing something fundamental here, but if the MAX297 tracks the VCO and divider, why does the THD of its output vary by over 100:1 over the audio range? 3% THD is pretty awful for a test generator! I'm suspicious that the THD measurement was made with an unmentioned LPF (with a cutoff just past 20 kHz) in place.
Bill Whitlock - 2008-8-8 00:59:00 EDT
Test Ideas: Pulse generator aids IC testing
09/01/2009Measure Your Filter's Characteristics
04/01/1999Sine/square wave generator
01/31/2004Notch Filter Extends Spectrum-Analyzer Range
01/01/2000Build your own capacitance meter
11/30/2009
-
FLIR offers IR camera for under $3000
-
Don't let the economy compromise quality (Guest commentary)
-
Danaher speeds up restructuring, acquires life-sciences businesses
-
Agilent’s Cover-Extend technology eliminates need for physical test points for in-circuit test
-
So many combinations: Testing a switch-matrix board
View All Webcasts
