Scopes Need Frequency Response Checks

The frequency-response test is the more common. In this test, the cal lab technician generates a sine wave of constant amplitude at increasing frequencies while measuring the signal’s amplitude on the scope.

This test shows how the scope’s amplitude responds versus frequency. Typically, the technician sets the signal’s amplitude and the scope’s vertical sensitivity to get 6 divisions (±3 divisions from center) on the scope’s grid. He or she measures the peak-to-peak amplitude while increasing frequency until the waveform reaches 0.707 of its original value (4.2 divisions peak-to-peak). That’s the –3-dB point, which is the typical measure of a scope’s bandwidth.

Note that the plot in Figure 1 isn’t the plot of a single-pole low-pass filter. Therefore, the scope’s frequency response doesn’t show the classical flat passband followed by a smooth roll-off. If the lab provides just the scope’s bandwidth, you won’t always know how the scope responds in the pass band. You can, however, request that the cal lab provide the scope’s response at frequencies below the –3-dB point.

Figure 1. During a calibration, a scope’s bandwidth is measured by setting the input signal to 6 divisions. The bandwidth is the lowest frequency where the scope measures the signal’s –3-dB point, 4.2 divisions.

If you have a digital scope and you know the frequency content of your signal, the lab can program the scope with a curve-fit equation to perform a correction that will flatten the instrument’s measurement response. On analog scopes, the technician uses the scope’s display to view the amplitude of the sine waves. Digital scopes, however, can calculate that value and provide a numerical readout.

The bandwidth of a scope is often lower at the scope’s lowest sensitivity settings. That’s because the input amplifier must operate at its highest gain. Table 1 shows the test results from a LeCroy LC584 DSO taken with a 50-V input. LeCroy’s catalog specifies the scope’s bandwidth at 1 GHz with the input set to 50 V, but the test report indicates that the bandwidth rating and the actual bandwidth at the two lowest sensitivity settings is lower. If you request that the lab measure signals on the 2-mV/div or 5-mV/div scales, be aware that all scope’s will have a lower bandwidth at the lowest volts/div settings.

If you use the scope at its 1-MV input impedance, then the bandwidth is also lower than at 50 V. The probe’s impedance limits the measureable bandwidth. To alleviate this problem, many engineers use active probes, which have a higher bandwidth than passive probes, to probe fast circuits.

Using a Pulse Generator
A scope’s bandwidth will also limit its response to a rising edge or falling edge. The lab can get a measure of a scope’s bandwidth by measuring how the scope responds to a rising edge from a pulse generator. The scope’s risetime is the time between when the edge is at 10% and 90% of the pulse’s full amplitude. DSOs have cursors that let a technician get an accurate measure of the risetime.

With an analog scope, the cal lab must measure risetime using the scope’s grid. After measuring the scope’s risetime, the lab technician can estimate the scope’s bandwidth through the following formula:

Bandwidth » 350 ¸ Risetime

So, a 350-ps risetime indicates a bandwidth of about 1 GHz.

Unfortunately, the measured risetime isn’t just the risetime of the scope. No pulse generator can create a pulse with zero risetime, so the pulse generator’s risetime must be subtracted from the measured risetime. The risetimes of the pulse generator and scope don’t add linearly, but a simple formula can provide the scope’s risetime:

  If the pulse generator’s risetime is sufficiently faster than that of the scope, then Tpulse gen becomes insignificant. For example, a manufacturer may calibrate its 1-GHz scopes with a pulse generator having a risetime of 50 ps. That risetime is insignificant compared to the 350-ps risetime of the scope.

When displaying a rising or falling edge, a scope may show ringing indicating an underdamped signal. Or you may see the corners cut off, indicating an overdamped signal.

Scope manufacturers use the pulse response to calibrate the input impedance of a scope’s input amplifiers during final test. They typically adjust input impedance to allow a small amount of ringing, which gets the maximum bandwidth from the scope.

The ringing amplitude is kept to less than 5% of the settled waveform’s amplitude. If the ringing exceeds 5%, then a component on the scope’s front end may need to be replaced. For routine calibration, a calibration lab won’t adjust the scope’s input impedance but will adjust a probe’s capacitance to match the probe’s impedance to that of the scope.     T&MW

FOR FURTHER READING
Customer Guidance Note: Calibration of Oscilloscopes, Wavetek, San Diego, CA, July 1997.
Deyst, John, Nicholas G. Paulter, Tamás Dabóczi, Gerard N. Stenbakken, and Michael T. Souders, “A Fast-Pulse Oscilloscope Calibration System,” IEEE Transactions on Instrumentation and Measurement, Vol. 47, No. 5, October 1998, IEEE, Piscataway, NJ.

Futornick, Ken, “Calibration of Oscilloscope High-Frequency Response,” 1998 Measurement Science Conference Proceedings, Measurement Science Conference, Newport Beach, CA.

Roddis, Richard, “Implementing Automated Oscilloscope Calibration Systems,” 1997 NCSL Workshop and Symposium Proceedings, National Conference of Standards Laboratories,
Boulder, CO.

Using the Fluke 5500A-SC Oscilloscope Calibration Option, Fluke, Everett, WA, www.fluke.com/applications/cal-app4.htm.