The analog oscilloscope
A badly compensated 10:1 probe makes the calibrator’s square wave start at 3.7 V instead of 5 V. The probe trimmer is not a detail: it decides whether the measured amplitude is correct.
01Seeing the shape of a signal
The oscilloscope plots voltage as a function of time. No other instrument shows at a glance the shape, the peaks, the oscillations, the rise times and the transients. Today almost all oscilloscopes are digital, but their vertical channel, their trigger and their probes inherit the architecture of the analog one, and understanding it explains why they are configured the way they are. Practical use is presented in the technician program, in the oscilloscope.
02The cathode-ray tube
In the analog oscilloscope, an electron gun emits a beam that is accelerated with several kilovolts and focused onto a phosphor-coated screen, which glows where the beam strikes it. Two pairs of plates deflect the beam: the vertical ones, with the amplified signal; the horizontal ones, with a ramp that sweeps it from left to right. The deflection is proportional to the voltage between plates, with a sensitivity of a few millimeters per volt, so the plates need amplifiers of tens or hundreds of volts. The brightness of the trace depends on how long the beam spends at each place: an analog scope naturally shows with more intensity what happens more often.
03The vertical channel
The signal goes through a compensated attenuator that sets the volts per division, a preamplifier, a delay line that delays it by about a hundred nanoseconds so that the sweep has time to start and the edge that triggered it is visible, and the final amplifier that drives the plates. The input presents 1 MΩ in parallel with about 20 pF, and the coupling can be DC, AC (with a capacitor that blocks the DC) or ground, to locate the zero.
The vertical channel is an amplifier with limited bandwidth, and that limits the rise time it can display:
04Timebase and trigger
The timebase generates a linear ramp that sweeps the beam in a time set by the seconds-per-division control. For the image to stay still, each sweep has to start at the same point of the signal: that is decided by the trigger, which starts the ramp when the signal crosses a level with a chosen slope. In normal mode it sweeps only if there is a trigger; in auto mode, it sweeps anyway after a time, to show the trace even when there is no signal; in single mode, it sweeps just once. The holdoff prevents re-triggering for a time, to stabilize complex signals.
In X-Y mode the ramp is replaced by a second signal: two sine waves form Lissajous figures, from which the frequency ratio and the phase shift are read.
05The compensated attenuator probe
The 10:1 probe has a 9 MΩ resistor at its tip which, together with the megohm input, divides by ten and multiplies by ten the resistance seen by the circuit. But the input also has capacitance, and the cable’s adds more: without correction, the resistive divider would become a low-pass filter. The solution is an adjustable capacitor in parallel with the 9 MΩ, which forms a capacitive divider with the same ratio:
A 9 MΩ probe with an adjustable capacitor \( C_p \), on a 1 MΩ input with 100 pF between the oscilloscope and the cable. The signal is the 0 to 5 V, 1 kHz square wave from the calibrator; the screen is set to 1 V/div and 0.2 ms/div, with the probe factor already applied.
Even when compensated, the probe loads the circuit: it presents 10 MΩ in parallel with about 10 pF. At 100 MHz that capacitance is a reactance of only 160 Ω. For fast signals in high-impedance circuits, active probes are used, with an amplifier at the tip and less than 1 pF.
06In the laboratory
Connect each probe to the calibrator, adjust the trimmer until the square wave looks flat and record how it looks overcompensated and undercompensated.
Measure the rise time of a fast pulse generator with oscilloscopes of different bandwidth and verify the sum in quadrature.
In X-Y mode, measure the phase shift of an RC network at several frequencies with the Lissajous ellipse and compare with the calculation.
07Common mistakes
- Not compensating the probe when changing it to another channel or oscilloscope. Each input has its own capacitance.
- Using the probe in 1:1 for fast signals. In that position the bandwidth drops to a few megahertz.
- Measuring rise times close to those of the oscilloscope without correcting.
- Connecting the probe ground to a point that is not ground. The oscilloscope ground is tied to the installation’s earth: a short circuit results.
08Self-assessment
What rise time does a 200 MHz oscilloscope have?
\( 0.35 / 200 \) MHz = 1.75 ns.
A 5 ns edge is measured with a 3.5 ns oscilloscope. What is read?
\( \sqrt{25 + 12.25} \approx 6.1 \) ns: 22% more.
What capacitance compensates a 9 MΩ probe on 1 MΩ with 20 pF?
\( C_p = 1 \cdot 20 / 9 \approx 2.2 \) pF.
How does the calibrator square wave look with the probe overcompensated?
With spikes on the edges that decay toward the flat top: the capacitive divider lets more than a tenth through on fast changes.
09Further reading
- Ian Hickman. Oscilloscopes: How to Use Them, How They Work. 5th ed., Newnes, 2001. The oscilloscope from the inside, from the tube to digital ones, with probes and triggering.
- Tektronix. XYZs of Oscilloscopes. Tektronix, technical primer. The manufacturer’s introduction: vertical and horizontal systems, triggering and specifications.
- Albert D. Helfrick and William D. Cooper. Modern Electronic Instrumentation and Measurement Techniques (Spanish edition: “Instrumentación electrónica moderna y técnicas de medición”). Prentice Hall, 1990. The cathode-ray oscilloscope and its circuits.