First oscilloscope lab: probe compensation, DC vs AC coupling and reading V_pp

Principles and common slips of the function generator and oscilloscope lab, using the values the scope displayed.

The first electronics lab moves on from the bench supply and multimeter to the function generator and the oscilloscope. The supply and multimeter part is covered with real data in the ammeter guide; this is the second half. The numbers in the tables are copied exactly as the scope's measurement panel showed them, and the waveform plots are redrawn from the generator settings.

1. What the lab is for

• Compensate a 10x probe and see what an uncompensated probe does to a waveform.
• Read V_pp and frequency of sine, square and triangle waves off the screen.
• Add a DC offset and see the difference between DC and AC coupling.
• Find out which controls leave the readings alone (vertical position, time base) and which ones get in the way.

2. How it works

Probe compensation. A 10x probe divides the signal by ten with 9 MΩ in the probe and the scope's 1 MΩ input. The scope input and the cable also have a capacitance C₂, so with resistors alone the higher-frequency parts of a signal would be cut more. An adjustable C₁ across the 9 MΩ makes the two time constants equal.

Too much C₁ makes the corners of a square wave spike and settle (over-compensated); too little rounds them off (under-compensated). With example values the over-compensated corner jumps to 1.23 V, the under-compensated one to 0.74 V, and both settle to 1 V with a time constant of about 100 µs.

The report's before-compensation screen showed exactly that spike-and-settle shape at every edge, and the corrected screen a flat square wave. A sine keeps its shape and only changes slightly in amplitude, so always check compensation on a square wave.

DC and AC coupling. DC coupling shows the signal as it is. AC coupling puts a series capacitor in the input to remove the mean (the DC part). That capacitor and the input resistance form a high-pass filter, so only signals well above its corner pass unchanged. The corner varies by model but is usually a few hertz to about 10 Hz, which hardly matters at 1 kHz.

A sine's mean equals its offset, so on AC coupling it drops to sit around 0 V. A 30% duty square wave's mean is not its midpoint but 1.2 V, so on AC coupling its levels become lopsided.

3. The setup

Probe the generator output (V_s) and clip every ground lead to the same ground. Putting the same signal on both channels lets you compare channels and probes, which is how a misbehaving channel shows up, as below.

4. Step by step

1. Set the probe switch to 10x and the scope channel's probe setting to 10X as well. If they differ, every reading is off by ten.
2. Clip the tip to the CAL terminal (usually a 1 kHz square wave) and the ground clip next to it, then turn the trimmer until the corners are flat. Do each channel.
3. Set the generator to a 1 kHz, 4 V_pp sine with 0 V offset, and check its output load setting is High-Z, not 50 Ω.
4. Choose a time base that shows at least two periods and a vertical scale that fills at least half the screen. Keep the trigger level between the waveform's minimum and maximum.
5. Turn on the V_pp and frequency measurements, note them, then add a 2 V offset and switch between DC and AC coupling.
6. Repeat with square waves (50% and 30% duty) and a triangle. Also try moving the vertical position and changing the time base, and note what the readings do.

5. The readings

Values the scope showed for the 4 V_pp sine. In the setting column 0 V and 2 V are the offset, DC and AC the coupling, "Moved" the capture with the vertical position moved, and 100 µs/div the one with the time base changed. Everything is at 1 kHz except the 10 kHz row, and the ratio is CH2 ÷ CH1. The time base was 500 µs/div at 1 kHz and 50 µs/div at 10 kHz. One photo was too blurred to read V_pp, so it is left out (—).

Amplitude

Frequency

The square and triangle waves were measured on CH1 only. 30% is the square wave's duty cycle and the triangle's symmetry (the share of the period spent rising); only the 30% square wave had a 2 V offset, the others 0 V. Every one of these captures showed 1.000 kHz; "vs setting" is the difference from the 4 V_pp setting.

6. Reading the results

• Coupling does not change V_pp: with the 2 V offset sine, CH1 read 3.995 V on DC coupling and 4.020 V on AC coupling. Coupling only moves the trace (its mean). The four CH1 readings at 1 kHz span 3.995 to 4.020 V, within 1% of the 4 V_pp setting.
• Only CH2 read low: CH2 was 0.74 to 0.78 times CH1 at 1 kHz and 0.84 times at 10 kHz. A wrong attenuation setting is off by a factor of ten and does not change with frequency, so it cannot explain this alone. A ratio that moves with frequency is typical of a probe that is not fully compensated; the report also does not say exactly where CH2 was connected, so the connection point and the probe contact are candidates too. Clip both probes to the same point, then swap them, to tell whether the probe or the channel is to blame.
• Square waves read above 4 V: 4.186 to 4.240 V for the square waves against 4.010 V for the triangle. V_pp is maximum minus minimum, so any overshoot, ringing or single noise spike at an edge goes straight into it. Probe compensation and a long ground lead shape the edges, so suspect those first. For the height of the flat parts, use an amplitude (top − base) measurement or cursors.
• When frequency shows "?": at 100 µs/div both channels showed "?" for frequency. With 14 divisions across, the screen held only 1.4 periods, too few to time one. At 500 µs/div it showed 7. Show at least two periods to measure frequency.
• Vertical position does not affect the readings: it only moves where the trace is drawn. If the trace runs off the top or bottom, though, V_pp reports the clipped height, so keep the whole waveform on screen.
• Small frequency differences: 1.003 kHz against 1.000 kHz is the resolution of the scope's frequency measurement. When comparing with the generator setting, only the displayed digits mean anything.

7. Common slips

• The probe switch (1x or 10x) and the channel's probe setting disagree, so readings are off by ten.
• The generator's output load is set to 50 Ω while it drives a High-Z scope input, so the screen shows twice the set amplitude; a 50 Ω load on a High-Z setting halves it.
• The trigger level is outside the waveform, so the picture rolls. Check the trigger level again after adding an offset or changing coupling.
• Ground clips go to different potentials. Scope ground is normally tied to mains earth, so clipping it to another point in the circuit shorts that point to ground.
• On AC coupling, a very low-frequency square wave sags across its top, and that gets mistaken for a signal problem. Near the corner frequency the coupling capacitor discharges and the waveform droops.

Convert V_pp to rms or dBV with the dB calculator, work out RC time constants for AC coupling and probe compensation with the RC filter calculator, and check a multimeter's accuracy on the same signal with the multimeter accuracy calculator.