There are only a few spurious peaks below the 100 uV peak level due to sampling frequencies in the ADC of the scope. The measured noise floor is about 1/100 th of the LSB noise, about what is expected. The expected reduction in noise is sqrt(20,000) = 141. To get this noise floor, we are effectively averaging over 20,000 measurements in the time domain. This is to be compared to the LSB of 1000 uV. The scope baseline shows a similar noise floor of -40 dBmV, which is an equivalent amplitude of 10 uV.
Qucs file based voltage source generator#
The low-cost function generator has 30 dB larger harmonics and significant sub harmonics. These higher harmonics are more than 70 dB down from the first harmonic, but the harmonic distributions are slightly different. The third harmonic is about -10 dBmV, or 0.3 mV amplitude. The first harmonic amplitude is about 60 dBmV, or 1 V amplitude, exactly as expected. The WaveStation and Digilent generators show very similar harmonic distortion. Calculated spectrum for the three sources and the channel noise. This is always a useful baseline when we are looking into the weeds at the fine details of signals.Īn important best measurement practice is situational awareness- always be aware of the limitations of your instruments and how close your signal’s figures of merit are to the instrument’s limits.įigure 3. In addition to measuring the three function generators, I also measured the noise on a channel with nothing connected. This is about 20,000 samples used to calculate the spectrum.įigure 3 shows the comparison of the four spectra. It’s easy to reverse engineer my settings and know that I must have used a time base of about 10 msec and a sample rate higher than 2 Msamples/sec. I changed the time base scale to get a resolution of 100 Hz and a freq range of 1 MHz.
All modern scopes can take the measured data acquisition buffer and perform an FFT and show the signal in the frequency domain. This is when we can unleash the power of analysis tools to dissect the signals. The time base was set for 125 Msamples/sec or a measurement every 8 nsec.Īt first glance, they all look the same. On this scale, the least significant bit (LSB) is 4 V/4095 = 0.98 mV. The vertical resolution of the scope is 12-bit. The flat line is the voltage on channel 4 with no input, used as a reference to see the system noise level. Measured sine waves from each source on a Teledyne LeCroy HDO8108 scope.
Qucs file based voltage source full#
The scope was set up to measure the signal from each generator, all on the same scale of 4 V full scale, 50 Ohm termination and the same time base of 50 usec full scale.įigure 2. Figure 2 shows the measured signals when each was setup for a 50 kHz sine wave with amplitude of 1 V and no offset. The first step in characterizing the sine wave sources is to look at their voltage output on a scope. Three different function generators spanning in price from $12 to $1200. The high-end function generator is a Teledyne LeCroy WaveStation, very similar to units from Keysight and other high-end vendors.įigure 1. The entire 2-channel scope with 2-channel function generator, costs only $275.
The mid-priced function generator is actually the waveform generator built into the Digilent Analog Discovery scope. The low-cost unit is a common function generator from eBay for $12.
The three sources I had available are shown in Figure 1. This method can be applied to any waveform. Using my scope, I was able to compare these three sine-wave sources and worked out a simple way of comparing each source to an ideal source. Measuring Three Different Function Generators Just how much bang do you get for your buck? I wondered just how much better the sine wave is from the more expensive function generator compared to the lower cost units. I happened to have three different sine wave generators available with a price range of $12, $275 and $1200.
I use a function generator as a source to stimulate circuits and measure their response.
0 kommentar(er)
