示波器，英文叫oscilloscope, 以前也称为oscillograph, 正式名字叫scope, CRO (阴极摄像管示波器), 或DSO (现在的数字示波器), 是电子测量仪器的一种，它能够观察固定变化的信号电压, 并将一个或多个信号的幅度以时间函数的方式进行二维显示，其它的信号比如声音或震动等可以转换成电压信号进行转换并显示。

示波器被用来观察电信号随时间的变化，并在屏幕上以校准好的标尺将电压和时间显示出来，并能够分析被观察的波形的参数比如幅度、频率、上升时间、时间间隔、失真等。现在的数字仪器都可以计算并直接显示这些参数。

示波器通过调整可以将重复的信号连续地在屏幕上显示，存储示波器可以捕捉单次事件并以相对较长的时间进行显示，这样可以观察到不容易直接接收到的快速变化的事件。

示波器被广泛用在科研、医疗、工程以及电信领域，通用的仪器被用在电子设备和实验室工作的维护中，专用的示波器可以用来比如分析汽车点火系统或在医疗中显示心跳的波形等。

The basic oscilloscope, as shown in the illustration, is typically divided into four sections: the display, vertical controls, horizontal controls and trigger controls. The display is usually a CRT or LCD panel which is laid out with both horizontal and vertical reference lines referred to as the graticule. In addition to the screen, most display sections are equipped with three basic controls: a focus knob, an intensity knob and a beam finder button.

The vertical section controls the amplitude of the displayed signal. This section carries a Volts-per-Division (Volts/Div) selector knob, an AC/DC/Ground selector switch and the vertical (primary) input for the instrument. Additionally, this section is typically equipped with the vertical beam position knob.

The horizontal section controls the time base or “sweep” of the instrument. The primary control is the Seconds-per-Division (Sec/Div) selector switch. Also included is a horizontal input for plotting dual X-Y axis signals. The horizontal beam position knob is generally located in this section.

The trigger section controls the start event of the sweep. The trigger can be set to automatically restart after each sweep or it can be configured to respond to an internal or external event. The principal controls of this section will be the source and coupling selector switches. An external trigger input (EXT Input) and level adjustment will also be included.

In addition to the basic instrument, most oscilloscopes are supplied with a probe as shown. The probe will connect to any input on the instrument and typically has a resistor of ten times the oscilloscope's input impedance. This results in a .1 (‑10X) attenuation factor, but helps to isolate the capacitive load presented by the probe cable from the signal being measured. Some probes have a switch allowing the operator to bypass the resistor when appropriate.[3]

Most modern oscilloscopes are lightweight, portable instruments that are compact enough to be easily carried by a single person. In addition to the portable units, the market offers a number of miniature battery-powered instruments for field service applications. Laboratory grade oscilloscopes, especially older units which use vacuum tubes, are generally bench-top devices or may be mounted into dedicated carts. Special-purpose oscilloscopes may be rack-mounted or permanently mounted into a custom instrument housing.

The signal to be measured is fed to one of the input connectors, which is usually a coaxial connector such as a BNC or UHF type. Binding posts or banana plugs may be used for lower frequencies. If the signal source has its own coaxial connector, then a simple coaxial cable is used; otherwise, a specialised cable called a “scope probe”, supplied with the oscilloscope, is used. In general, for routine use, an open wire test lead for connecting to the point being observed is not satisfactory, and a probe is generally necessary. General-purpose oscilloscopes usually present an input impedance of 1 megohm in parallel with a small but known capacitance such as 20 picofarads.[4] This allows the use of standard oscilloscope probes.[5] Scopes for use with very high frequencies may have 50‑ohm inputs, which must be either connected directly to a 50‑ohm signal source or used with Z0 or active probes.

Less-frequently-used inputs include one (or two) for triggering the sweep, horizontal deflection for X‑Y mode displays, and trace brightening/darkening, sometimes called z'‑axis inputs.

Open wire test leads (flying leads) are likely to pick up interference, so they are not suitable for low level signals. Furthermore, the leads have a high inductance, so they are not suitable for high frequencies. Using a shielded cable (i.e., coaxial cable) is better for low level signals. Coaxial cable also has lower inductance, but it has higher capacitance: a typical 50 ohm cable has about 90 pF per meter. Consequently, a one-meter direct (1X) coaxial probe will load a circuit with a capacitance of about 110 pF and a resistance of 1 megohm.

To minimize loading, attenuator probes (e.g., 10X probes) are used. A typical probe uses a 9 megohm series resistor shunted by a low-value capacitor to make an RC compensated divider with the cable capacitance and scope input. The RC time constants are adjusted to match. For example, the 9 megohm series resistor is shunted by a 12.2 pF capacitor for a time constant of 110 microseconds. The cable capacitance of 90 pF in parallel with the scope input of 20 pF and 1 megohm (total capacitance 110 pF) also gives a time constant of 110 microseconds. In practice, there will be an adjustment so the operator can precisely match the low frequency time constant (called compensating the probe). Matching the time constants makes the attenuation independent of frequency. At low frequencies (where the resistance of R is much less than the reactance of C), the circuit looks like a resistive divider; at high frequencies (resistance much greater than reactance), the circuit looks like a capacitive divider.[6]

The result is a frequency compensated probe for modest frequencies that presents a load of about 10 megohms shunted by 12 pF. Although such a probe is an improvement, it does not work when the time scale shrinks to several cable transit times (transit time is typically 5 ns). In that time frame, the cable looks like its characteristic impedance, and there will be reflections from the transmission line mismatch at the scope input and the probe that causes ringing.[7] The modern scope probe uses lossy low capacitance transmission lines and sophisticated frequency shaping networks to make the 10X probe perform well at several hundred megahertz. Consequently, there are other adjustments for completing the compensation.[8][9]

Probes with 10:1 attenuation are by far the most common; for large signals (and slightly-less capacitive loading), 100:1 probes are not rare. There are also probes that contain switches to select 10:1 or direct (1:1) ratios, but one must be aware that the 1:1 setting has significant capacitance (tens of pF) at the probe tip, because the whole cable's capacitance is now directly connected.

Most oscilloscopes allow for probe attenuation factors, displaying the effective sensitivity at the probe tip. Historically, some auto-sensing circuitry used indicator lamps behind translucent windows in the panel to illuminate different parts of the sensitivity scale. To do so, the probe connectors (modified BNCs) had an extra contact to define the probe's attenuation. (A certain value of resistor, connected to ground, “encodes” the attenuation.) Because probes wear out, and because the auto-sensing circuitry is not compatible between different makes of oscilloscope, auto-sensing probe scaling is not foolproof. Likewise, manually setting the probe attenuation is prone to user error and it is a common mistake to have the probe scaling set incorrectly; resultant voltage readings can then be wrong by a factor of 10.

There are special high voltage probes which also form compensated attenuators with the oscilloscope input; the probe body is physically large, and some require partly filling a canister surrounding the series resistor with volatile liquid fluorocarbon to displace air. At the oscilloscope end is a box with several waveform-trimming adjustments. For safety, a barrier disc keeps one's fingers distant from the point being examined. Maximum voltage is in the low tens of kV. (Observing a high voltage ramp can create a staircase waveform with steps at different points every repetition, until the probe tip is in contact. Until then, a tiny arc charges the probe tip, and its capacitance holds the voltage (open circuit). As the voltage continues to climb, another tiny arc charges the tip further.)

There are also current probes, with cores that surround the conductor carrying current to be examined. One type has a hole for the conductor, and requires that the wire be passed through the hole; they are for semi-permanent or permanent mounting. However, other types, for testing, have a two-part core that permit them to be placed around a wire. Inside the probe, a coil wound around the core provides a current into an appropriate load, and the voltage across that load is proportional to current. However, this type of probe can sense AC, only.

A more-sophisticated probe includes a magnetic flux sensor (Hall effect sensor) in the magnetic circuit. The probe connects to an amplifier, which feeds (low frequency) current into the coil to cancel the sensed field; the magnitude of that current provides the low-frequency part of the current waveform, right down to DC. The coil still picks up high frequencies. There is a combining network akin to a loudspeaker crossover network.

This control adjusts CRT focus to obtain the sharpest, most-detailed trace. In practice, focus needs to be adjusted slightly when observing quite-different signals, which means that it needs to be an external control. Flat-panel displays do not need focus adjustments and therefore do not include this control.

This adjusts trace brightness. Slow traces on CRT oscilloscopes need less, and fast ones, especially if not often repeated, require more. On flat panels, however, trace brightness is essentially independent of sweep speed, because the internal signal processing effectively synthesizes the display from the digitized data.

Can also be called “Shape” or “spot shape”. Adjusts the relative voltages on two of the CRT anodes such that a displayed spot changes from elliptical in one plane through a circular spot to an ellipse at 90 degrees to the first. This control may be absent from simpler oscilloscope designs or may even be an internal control. It is not necessary with flat panel displays.

Modern oscilloscopes have direct-coupled deflection amplifiers, which means the trace could be deflected off-screen. They also might have their beam blanked without the operator knowing it. To help in restoring a visible display, the beam finder circuit overrides any blanking and limits the beam deflected to the visible portion of the screen. Beam-finder circuits often distort the trace while activated.

The graticule is a grid of squares that serve as reference marks for measuring the displayed trace. These markings, whether located directly on the screen or on a removable plastic filter, usually consist of a 1 cm grid with closer tick marks (often at 2 mm) on the centre vertical and horizontal axis. One expects to see ten major divisions across the screen; the number of vertical major divisions varies. Comparing the grid markings with the waveform permits one to measure both voltage (vertical axis) and time (horizontal axis). Frequency can also be determined by measuring the waveform period and calculating its reciprocal.

On old and lower-cost CRT oscilloscopes the graticule is a sheet of plastic, often with light-diffusing markings and concealed lamps at the edge of the graticule. The lamps had a brightness control. Higher-cost instruments have the graticule marked on the inside face of the CRT, to eliminate parallax errors; better ones also had adjustable edge illumination with diffusing markings. (Diffusing markings appear bright.) Digital oscilloscopes, however, generate the graticule markings on the display in the same way as the trace.

External graticules also protect the glass face of the CRT from accidental impact. Some CRT oscilloscopes with internal graticules have an unmarked tinted sheet plastic light filter to enhance trace contrast; this also serves to protect the faceplate of the CRT.

Accuracy and resolution of measurements using a graticule is relatively limited; better instruments sometimes have movable bright markers on the trace that permit internal circuits to make more refined measurements.

Both calibrated vertical sensitivity and calibrated horizontal time are set in 1 - 2 - 5 - 10 steps. This leads, however, to some awkward interpretations of minor divisions

These select the horizontal speed of the CRT's spot as it creates the trace; this process is commonly referred to as the sweep. In all but the least-costly modern oscilloscopes, the sweep speed is selectable and calibrated in units of time per major graticule division. Quite a wide range of sweep speeds is generally provided, from seconds to as fast as picoseconds (in the fastest) per division. Usually, a continuously-variable control (often a knob in front of the calibrated selector knob) offers uncalibrated speeds, typically slower than calibrated. This control provides a range somewhat greater than that of consecutive calibrated steps, making any speed available between the extremes.

Found on some better analog oscilloscopes, this varies the time (holdoff) during which the sweep circuit ignores triggers. It provides a stable display of some repetitive events in which some triggers would create confusing displays. It is usually set to minimum, because a longer time decreases the number of sweeps per second, resulting in a dimmer trace. See Holdoff for a more detailed description.

To accommodate a wide range of input amplitudes, a switch selects calibrated sensitivity of the vertical deflection. Another control, often in front of the calibrated-selector knob, offers a continuously-variable sensitivity over a limited range from calibrated to less-sensitive settings.

Often the observed signal is offset by a steady component, and only the changes are of interest. A switch (AC position) connects a capacitor in series with the input that passes only the changes (provided that they are not too slow – “slow” would mean visible). However, when the signal has a fixed offset of interest, or changes quite slowly, the input is connected directly (DC switch position). Most oscilloscopes offer the DC input option. For convenience, to see where zero volts input currently shows on the screen, many oscilloscopes have a third switch position (GND) that disconnects the input and grounds it. Often, in this case, the user centers the trace with the Vertical Position control.

Better oscilloscopes have a polarity selector. Normally, a positive input moves the trace upward, but this permits inverting—positive deflects the trace downward.

This control is found only on more elaborate oscilloscopes; it offers adjustable sensitivity for external horizontal inputs.

The vertical position control moves the whole displayed trace up and down. It is used to set the no-input trace exactly on the center line of the graticule, but also permits offsetting vertically by a limited amount. With direct coupling, adjustment of this control can compensate for a limited DC component of an input.

The horizontal position control moves the display sidewise. It usually sets the left end of the trace at the left edge of the graticule, but it can displace the whole trace when desired. This control also moves the X-Y mode traces sidewise in some instruments, and can compensate for a limited DC component as for vertical position.

Each input channel usually has its own set of sensitivity, coupling, and position controls, although some four-trace oscilloscopes have only minimal controls for their third and fourth channels.

Dual-trace oscilloscopes have a mode switch to select either channel alone, both channels, or (in some) an X‑Y display, which uses the second channel for X deflection. When both channels are displayed, the type of channel switching can be selected on some oscilloscopes; on others, the type depends upon timebase setting. If manually selectable, channel switching can be free-running (asynchronous), or between consecutive sweeps. Some Philips dual-trace analog oscilloscopes had a fast analog multiplier, and provided a display of the product of the input channels.

Multiple-trace oscilloscopes have a switch for each channel to enable or disable display of that trace's signal.

These include controls for the delayed-sweep timebase, which is calibrated, and often also variable. The slowest speed is several steps faster than the slowest main sweep speed, although the fastest is generally the same. A calibrated multiturn delay time control offers wide range, high resolution delay settings; it spans the full duration of the main sweep, and its reading corresponds to graticule divisions (but with much finer precision). Its accuracy is also superior to that of the display.

A switch selects display modes: Main sweep only, with a brightened region showing when the delayed sweep is advancing, delayed sweep only, or (on some) a combination mode.

Good CRT oscilloscopes include a delayed-sweep intensity control, to allow for the dimmer trace of a much-faster delayed sweep that nevertheless occurs only once per main sweep. Such oscilloscopes also are likely to have a trace separation control for multiplexed display of both the main and delayed sweeps together.

A switch selects the Trigger Source. It can be an external input, one of the vertical channels of a dual or multiple-trace oscilloscope, or the AC line (mains) frequency. Another switch enables or disables Auto trigger mode, or selects single sweep, if provided in the oscilloscope. Either a spring-return switch position or a pushbutton arms single sweeps.

A Level control varies the voltage on the waveform which generates a trigger, and the Slope switch selects positive-going or negative-going polarity at the selected trigger level.

To display events with unchanging or slowly (visibly) changing waveforms, but occurring at times that may not be evenly spaced, modern oscilloscopes have triggered sweeps. Compared to simpler oscilloscopes with sweep oscillators that are always running, triggered-sweep oscilloscopes are markedly more versatile.

A triggered sweep starts at a selected point on the signal, providing a stable display. In this way, triggering allows the display of periodic signals such as sine waves and square waves, as well as nonperiodic signals such as single pulses, or pulses that do not recur at a fixed rate.

With triggered sweeps, the scope will blank the beam and start to reset the sweep circuit each time the beam reaches the extreme right side of the screen. For a period of time, called holdoff, (extendable by a front-panel control on some better oscilloscopes), the sweep circuit resets completely and ignores triggers. Once holdoff expires, the next trigger starts a sweep. The trigger event is usually the input waveform reaching some user-specified threshold voltage (trigger level) in the specified direction (going positive or going negative—trigger polarity).

In some cases, variable holdoff time can be really useful to make the sweep ignore interfering triggers that occur before the events to be observed. In the case of repetitive, but complex waveforms, variable holdoff can create a stable display that cannot otherwise be achieved.

Trigger holdoff defines a certain period following a trigger during which the scope will not trigger again. This makes it easier to establish a stable view of a waveform with multiple edges which would otherwise cause another trigger.[10]

Imagine the following repeating waveform:

The green line is the waveform, the red vertical partial line represents the location of the trigger, and the yellow line represents the trigger level. If the scope was simply set to trigger on every rising edge, this waveform would cause three triggers for each cycle:

Assuming the signal is fairly high frequency, the scope would probably look something like this:

Except that on the scope, each trigger would be the same channel, and so would be the same color.

It is desired to set the scope to only trigger on one edge per cycle, so it is necessary to set the holdoff to be slightly less than the period of the waveform. That will prevent it from triggering more than once per cycle, but still allow it to trigger on the first edge of the next cycle.

Triggered sweeps can display a blank screen if there are no triggers. To avoid this, these sweeps include a timing circuit that generates free-running triggers so a trace is always visible. Once triggers arrive, the timer stops providing pseudo-triggers. Automatic sweep mode can be de-selected when observing low repetition rates.

If the input signal is periodic, the sweep repetition rate can be adjusted to display a few cycles of the waveform. Early (tube) oscilloscopes and lowest-cost oscilloscopes have sweep oscillators that run continuously, and are uncalibrated. Such oscilloscopes are very simple, comparatively inexpensive, and were useful in radio servicing and some TV servicing. Measuring voltage or time is possible, but only with extra equipment, and is quite inconvenient. They are primarily qualitative instruments.

They have a few (widely spaced) frequency ranges, and relatively wide-range continuous frequency control within a given range. In use, the sweep frequency is set to slightly lower than some submultiple of the input frequency, to display typically at least two cycles of the input signal (so all details are visible). A very simple control feeds an adjustable amount of the vertical signal (or possibly, a related external signal) to the sweep oscillator. The signal triggers beam blanking and a sweep retrace sooner than it would occur free-running, and the display becomes stable.

Some oscilloscopes offer these—the sweep circuit is manually armed (typically by a pushbutton or equivalent) “Armed” means it's ready to respond to a trigger. Once the sweep is complete, it resets, and will not sweep until re-armed. This mode, combined with an oscilloscope camera, captures single-shot events.

触发类型包括:

• external trigger, a pulse from an external source connected to a dedicated input on the scope.
• edge trigger, an edge-detector that generates a pulse when the input signal crosses a specified threshold voltage in a specified direction. These are the most-common types of triggers; the level control sets the threshold voltage, and the slope control selects the direction (negative or positive-going). (The first sentence of the description also applies to the inputs to some digital logic circuits; those inputs have fixed threshold and polarity response.)
• video trigger, a circuit that extracts synchronizing pulses from video formats such as PAL and NTSC and triggers the timebase on every line, a specified line, every field, or every frame. This circuit is typically found in a waveform monitor device, although some better oscilloscopes include this function.
• delayed trigger, which waits a specified time after an edge trigger before starting the sweep. As described under delayed sweeps, a trigger delay circuit (typically the main sweep) extends this delay to a known and adjustable interval. In this way, the operator can examine a particular pulse in a long train of pulses.

Some recent designs of oscilloscopes include more sophisticated triggering schemes; these are described toward the end of this article.

More sophisticated analog oscilloscopes contain a second timebase for a delayed sweep. A delayed sweep provides a very detailed look at some small selected portion of the main timebase. The main timebase serves as a controllable delay, after which the delayed timebase starts. This can start when the delay expires, or can be triggered (only) after the delay expires. Ordinarily, the delayed timebase is set for a faster sweep, sometimes much faster, such as 1000:1. At extreme ratios, jitter in the delays on consecutive main sweeps degrades the display, but delayed-sweep triggers can overcome that.

The display shows the vertical signal in one of several modes: the main timebase, or the delayed timebase only, or a combination thereof. When the delayed sweep is active, the main sweep trace brightens while the delayed sweep is advancing. In one combination mode, provided only on some oscilloscopes, the trace changes from the main sweep to the delayed sweep once the delayed sweep starts, although less of the delayed fast sweep is visible for longer delays. Another combination mode multiplexes (alternates) the main and delayed sweeps so that both appear at once; a trace separation control displaces them.

DSOs allow waveforms to be displayed in this way, without offering a delayed timebase as such.

Oscilloscopes with two vertical inputs, referred to as dual-trace oscilloscopes, are extremely useful and commonplace. Using a single-beam CRT, they multiplex the inputs, usually switching between them fast enough to display two traces apparently at once. Less common are oscilloscopes with more traces; four inputs are common among these, but a few (Kikusui, for one) offered a display of the sweep trigger signal if desired. Some multi-trace oscilloscopes use the external trigger input as an optional vertical input, and some have third and fourth channels with only minimal controls. In all cases, the inputs, when independently displayed, are time-multiplexed, but dual-trace oscilloscopes often can add their inputs to display a real-time analog sum. (Inverting one channel provides a difference, provided that neither channel is overloaded. This difference mode can provide a moderate-performance differential input.)

Switching channels can be asynchronous, that is, free-running, with trace blanking while switching, or after each horizontal sweep is complete. Asynchronous switching is usually designated “Chopped”, while sweep-synchronized is designated “Alt[ernate]”. A given channel is alternately connected and disconnected, leading to the term “chopped”. Multi-trace oscilloscopes also switch channels either in chopped or alternate modes.

In general, chopped mode is better for slower sweeps. It is possible for the internal chopping rate to be a multiple of the sweep repetition rate, creating blanks in the traces, but in practice this is rarely a problem; the gaps in one trace are overwritten by traces of the following sweep. A few oscilloscopes had a modulated chopping rate to avoid this occasional problem. Alternate mode, however, is better for faster sweeps.

True dual-beam CRT oscilloscopes did exist, but were not common. One type (Cossor, U.K.) had a beam-splitter plate in its CRT, and single-ended deflection following the splitter. Others had two complete electron guns, requiring tight control of axial (rotational) mechanical alignment in manufacturing the CRT. Beam-splitter types had horizontal deflection common to both vertical channels, but dual-gun oscilloscopes could have separate time bases, or use one time base for both channels. Multiple-gun CRTs (up to ten guns) were made in past decades. With ten guns, the envelope (bulb) was cylindrical throughout its length. (Also see “CRT Invention” in Oscilloscope history.)

In an analog oscilloscope, the vertical amplifier acquires the signal[s] to be displayed. In better oscilloscopes, it delays them by a fraction of a microsecond, and provides a signal large enough to deflect the CRT's beam. That deflection is at least somewhat beyond the edges of the graticule, and more typically some distance off-screen. The amplifier has to have low distortion to display its input accurately (it must be linear), and it has to recover quickly from overloads. As well, its time-domain response has to represent transients accurately—minimal overshoot, rounding, and tilt of a flat pulse top.

A vertical input goes to a frequency-compensated step attenuator to reduce large signals to prevent overload. The attenuator feeds a low-level stage (or a few), which in turn feed gain stages (and a delay-line driver if there is a delay). Following are more gain stages, up to the final output stage which develops a large signal swing (tens of volts, sometimes over 100 volts) for CRT electrostatic deflection.

In dual and multiple-trace oscilloscopes, an internal electronic switch selects the relatively low-level output of one channel's amplifiers and sends it to the following stages of the vertical amplifier, which is only a single channel, so to speak, from that point on.

In free-running (“chopped”) mode, the oscillator (which may be simply a different operating mode of the switch driver) blanks the beam before switching, and unblanks it only after the switching transients have settled.

Part way through the amplifier is a feed to the sweep trigger circuits, for internal triggering from the signal. This feed would be from an individual channel's amplifier in a dual or multi-trace oscilloscope, the channel depending upon the setting of the trigger source selector.

This feed precedes the delay (if there is one), which allows the sweep circuit to unblank the CRT and start the forward sweep, so the CRT can show the triggering event. High-quality analog delays add a modest cost to an oscilloscope, and are omitted in oscilloscopes that are cost-sensitive.

The delay, itself, comes from a special cable with a pair of conductors wound around a flexible, magnetically soft core. The coiling provides distributed inductance, while a conductive layer close to the wires provides distributed capacitance. The combination is a wideband transmission line with considerable delay per unit length. Both ends of the delay cable require matched impedances to avoid reflections.

A 24-hour clock displayed on a CRT oscilloscope configured in X-Y mode as a vector monitor with dual R2R DACs to generate the analog voltages. Most modern oscilloscopes have several inputs for voltages, and thus can be used to plot one varying voltage versus another. This is especially useful for graphing I-V curves (current versus voltage characteristics) for components such as diodes, as well Lissajous patterns. Lissajous figures are an example of how an oscilloscope can be used to track phase differences between multiple input signals. This is very frequently used in broadcast engineering to plot the left and right stereophonic channels, to ensure that the stereo generator is calibrated properly. Historically, stable Lissajous figures were used to show that two sine waves had a relatively simple frequency relationship, a numerically-small ratio. They also indicated phase difference between two sine waves of the same frequency.

The X-Y mode also allows the oscilloscope to be used as a vector monitor to display images or user interfaces. Many early games, such as Tennis for Two, used an oscilloscope as an output device.

Complete loss of signal in an X-Y CRT display means that the beam strikes a small spot, which risks burning the phosphor. Older phosphors burned more easily. Some dedicated X-Y displays reduce beam current greatly, or blank the display entirely, if there are no inputs present.

As with all practical instruments, oscilloscopes do not respond equally to all possible input frequencies. The range of frequencies an oscilloscope can usefully display is referred to as its bandwidth. Bandwidth applies primarily to the Y-axis, although the X-axis sweeps have to be fast enough to show the highest-frequency waveforms.

The bandwidth is defined as the frequency at which the sensitivity is 0.707 of that at DC or the lowest AC frequency (a drop of 3 dB). The oscilloscope's response will drop off rapidly as the input frequency is raised above that point. Within the stated bandwidth the response will not necessarily be exactly uniform (or “flat”), but should always fall within a +0 to -3 dB range. One source states that there is a noticeable effect on the accuracy of voltage measurements at only 20 percent of the stated bandwidth. Some oscilloscopes' specifications do include a narrower tolerance range within the stated bandwidth.

Probes also have bandwidth limits and must be chosen and used to properly handle the frequencies of interest. To achieve the flattest response, most probes must be “compensated” (an adjustment performed using a test signal from the oscilloscope) to allow for the reactance of the probe's cable.

Another related specification is rise time. This is the duration of the fastest pulse that can be resolved by the scope. It is related to the bandwidth approximately by:

Bandwidth in Hz x rise time in seconds = 0.35

For example, an oscilloscope intended to resolve pulses with a rise time of 1 nanosecond would have a bandwidth of 350 MHz.

In analog instruments, the bandwidth of the oscilloscope is limited by the vertical amplifiers and the CRT or other display subsystem. In digital instruments, the sampling rate of the analog to digital converter (ADC) is a factor, but the stated analog bandwidth (and therefore the overall bandwidth of the instrument) is usually less than the ADC's Nyquist frequency. This is due to limitations in the analog signal amplifier, deliberate design of the Anti-aliasing filter that precedes the ADC, or both.

For a digital oscilloscope, a rule of thumb is that the continuous sampling rate should be ten times the highest frequency desired to resolve; for example a 20 megasample/second rate would be applicable for measuring signals up to about 2 megahertz. This allows the anti-aliasing filter to be designed with a 3 dB down point of 2 MHz and an effective cutoff at 10 MHz (the Nyquist frequency), avoiding the artifacts of a very steep (“brick-wall”) filter.

A sampling oscilloscope can display signals of considerably higher frequency than the sampling rate if the signals are exactly, or nearly, repetitive. It does this by taking one sample from each successive repetition of the input waveform, each sample being at an increased time interval from the trigger event. The waveform is then displayed from these collected samples. This mechanism is referred to as “equivalent-time sampling”.[14] Some oscilloscopes can operate in either this mode or in the more traditional “real-time” mode at the operator's choice.

Some oscilloscopes have cursors, which are lines that can be moved about the screen to measure the time interval between two points, or the difference between two voltages. A few older oscilloscopes simply brightened the trace at movable locations. These cursors are more accurate than visual estimates referring to graticule lines.

Better quality general purpose oscilloscopes include a calibration signal for setting up the compensation of test probes; this is (often) a 1 kHz square-wave signal of a definite peak-to-peak voltage available at a test terminal on the front panel. Some better oscilloscopes also have a squared-off loop for checking and adjusting current probes.

Sometimes the event that the user wants to see may only happen occasionally. To catch these events, some oscilloscopes, known as “storage scopes”, preserve the most recent sweep on the screen. This was originally achieved by using a special CRT, a “storage tube”, which would retain the image of even a very brief event for a long time.

Some digital oscilloscopes can sweep at speeds as slow as once per hour, emulating a strip chart recorder. That is, the signal scrolls across the screen from right to left. Most oscilloscopes with this facility switch from a sweep to a strip-chart mode at about one sweep per ten seconds. This is because otherwise, the scope looks broken: it's collecting data, but the dot cannot be seen.

In current oscilloscopes, digital signal sampling is more often used for all but the simplest models. Samples feed fast analog-to-digital converters, following which all signal processing (and storage) is digital.

Many oscilloscopes have different plug-in modules for different purposes, e.g., high-sensitivity amplifiers of relatively narrow bandwidth, differential amplifiers, amplifiers with four or more channels, sampling plugins for repetitive signals of very high frequency, and special-purpose plugins, including audio/ultrasonic spectrum analyzers, and stable-offset-voltage direct-coupled channels with relatively high gain.

One of the most frequent uses of scopes is troubleshooting malfunctioning electronic equipment. One of the advantages of a scope is that it can graphically show signals: where a voltmeter may show a totally unexpected voltage, a scope may reveal that the circuit is oscillating. In other cases the precise shape or timing of a pulse is important.

In a piece of electronic equipment, for example, the connections between stages (e.g. electronic mixers, electronic oscillators, amplifiers) may be 'probed' for the expected signal, using the scope as a simple signal tracer. If the expected signal is absent or incorrect, some preceding stage of the electronics is not operating correctly. Since most failures occur because of a single faulty component, each measurement can prove that half of the stages of a complex piece of equipment either work, or probably did not cause the fault.

Once the faulty stage is found, further probing can usually tell a skilled technician exactly which component has failed. Once the component is replaced, the unit can be restored to service, or at least the next fault can be isolated. This sort of troubleshooting is typical of radio and TV receivers, as well as audio amplifiers, but can apply to quite-different devices such as electronic motor drives.

Another use is to check newly designed circuitry. Very often a newly designed circuit will misbehave because of design errors, bad voltage levels, electrical noise etc. Digital electronics usually operate from a clock, so a dual-trace scope which shows both the clock signal and a test signal dependent upon the clock is useful. Storage scopes are helpful for “capturing” rare electronic events that cause defective operation.

Pictures of use

First appearing in the 1970s for ignition system analysis, automotive oscilloscopes are becoming an important workshop tool for testing sensors and output signals on electronic engine management systems, braking and stability systems.

For work at high frequencies and with fast digital signals, the bandwidth of the vertical amplifiers and sampling rate must be high enough. For general-purpose use, a bandwidth of at least 100 MHz is usually satisfactory. A much lower bandwidth is sufficient for audio-frequency applications only. A useful sweep range is from one second to 100 nanoseconds, with appropriate triggering and (for analog instruments) sweep delay. A well-designed, stable trigger circuit is required for a steady display. The chief benefit of a quality oscilloscope is the quality of the trigger circuit.[citation needed]

Key selection criteria of a DSO (apart from input bandwidth) are the sample memory depth and sample rate. Early DSOs in the mid- to late 1990s only had a few KB of sample memory per channel. This is adequate for basic waveform display, but does not allow detailed examination of the waveform or inspection of long data packets for example. Even entry-level (<$500) modern DSOs now have 1 MB or more of sample memory per channel, and this has become the expected minimum in any modern DSO.[citation needed] Often this sample memory is shared between channels, and can sometimes only be fully available at lower sample rates. At the highest sample rates, the memory may be limited to a few tens of KB. Any modern “real-time” sample rate DSO will have typically 5–10 times the input bandwidth in sample rate. So a 100 MHz bandwidth DSO would have 500 Ms/s – 1 Gs/s sample rate. The theoretical minimum sample rate required, using SinX/x interpolation, is 2.5 times the bandwidth.

Analog oscilloscopes have been almost totally displaced by digital storage scopes except for use exclusively at lower frequencies. Greatly increased sample rates have largely eliminated the display of incorrect signals, known as “aliasing”, that was sometimes present in the first generation of digital scopes. The problem can still occur when, for example, viewing a short section of a repetitive waveform that repeats at intervals thousands of times longer than the section viewed (for example a short synchronization pulse at the beginning of a particular television line), with an oscilloscope that cannot store the extremely large number of samples between one instance of the short section and the next.

The used test equipment market, particularly on-line auction venues, typically has a wide selection of older analog scopes available. However it is becoming more difficult to obtain replacement parts for these instruments, and repair services are generally unavailable from the original manufacturer. Used instruments are usually out of calibration, and recalibration by companies with the equipment and expertise usually costs more than the second-hand value of the instrument.[citation needed]

As of 2007, a 350 MHz bandwidth (BW), 2.5 gigasamples per second (GS/s), dual-channel digital storage scope costs about US$7000 new.

On the lowest end, an inexpensive hobby-grade single-channel DSO could be purchased for under $90 as of June 2011. These often have limited bandwidth and other facilities, but fulfill the basic functions of an oscilloscope.

Many oscilloscopes today provide one or more external interfaces to allow remote instrument control by external software. These interfaces (or buses) include GPIB, Ethernet, serial port, and USB.

The earliest and simplest type of oscilloscope consisted of a cathode ray tube, a vertical amplifier, a timebase, a horizontal amplifier and a power supply. These are now called “analog” scopes to distinguish them from the “digital” scopes that became common in the 1990s and 2000s.

Analog scopes do not necessarily include a calibrated reference grid for size measurement of waves, and they may not display waves in the traditional sense of a line segment sweeping from left to right. Instead, they could be used for signal analysis by feeding a reference signal into one axis and the signal to measure into the other axis. For an oscillating reference and measurement signal, this results in a complex looping pattern referred to as a Lissajous curve. The shape of the curve can be interpreted to identify properties of the measurement signal in relation to the reference signal, and is useful across a wide range of oscillation frequencies.

A standard DSO is limited to capturing signals with a bandwidth of less than half the sampling rate of the ADC (called the Nyquist limit). There is a variation of the DSO called the digital sampling oscilloscope that can exceed this limit for certain types of signal, such as high-speed communications signals, where the waveform consists of repeating pulses. This type of DSO deliberately samples at a much lower frequency than the Nyquist limit and then uses signal processing to reconstruct a composite view of a typical pulse. A similar technique, with analog rather than digital samples, was used before the digital era in analog sampling oscilloscopes.

A digital phosphor oscilloscope (DPO) uses color information to convey information about a signal. It may, for example, display infrequent signal data in blue to make it stand out. In a conventional analog scope, such a rare trace may not be visible.

Handheld oscilloscopes are useful for many test and field service applications. Today, a hand held oscilloscope is usually a digital sampling oscilloscope, using a liquid crystal display. Many hand-held and bench oscilloscopes have the ground reference voltage common to all input channels. If more than one measurement channel is used at the same time, all the input signals must have the same voltage reference, and the shared default reference is the “earth”. If there is no differential preamplifier or external signal isolator, this traditional desktop oscilloscope is not suitable for floating measurements. (Occasionally an oscilloscope user will break the ground pin in the power supply cord of a bench-top oscilloscope in an attempt to isolate the signal common from the earth ground. This practice is unreliable since the entire stray capacitance of the instrument cabinet will be connected into the circuit. Since it is also a hazard to break a safety ground connection, instruction manuals strongly advise against this practice.)

Some models of oscilloscope have isolated inputs, where the signal reference level terminals are not connected together. Each input channel can be used to make a “floating” measurement with an independent signal reference level. Measurements can be made without tying one side of the oscilloscope input to the circuit signal common or ground reference.

The isolation available is categorized as shown below:

Overvoltage category Operating voltage (effective value of AC/DC to ground) Peak instantaneous voltage (repeated 20 times) Test resistor

PicoScope 6000 digital PC-based oscilloscope using a laptop computer for display & processing For more details on this topic, see Oscilloscope types § PC-based oscilloscopes. A new type of oscilloscope is emerging that consists of a specialized signal acquisition board (which can be an external USB or parallel port device, or an internal add-on PCI or ISA card). The user interface and signal processing software runs on the user's computer, rather than on an embedded computer as in the case of a conventional DSO.

A large number of instruments used in a variety of technical fields are really oscilloscopes with inputs, calibration, controls, display calibration, etc., specialized and optimized for a particular application. Examples of such oscilloscope-based instruments include waveform monitors for analyzing video levels in television productions and medical devices such as vital function monitors and electrocardiogram and electroencephalogram instruments. In automobile repair, an ignition analyzer is used to show the spark waveforms for each cylinder. All of these are essentially oscilloscopes, performing the basic task of showing the changes in one or more input signals over time in an X‑Y display.

Other instruments convert the results of their measurements to a repetitive electrical signal, and incorporate an oscilloscope as a display element. Such complex measurement systems include spectrum analyzers, transistor analyzers, and time domain reflectometers (TDRs). Unlike an oscilloscope, these instruments automatically generate stimulus or sweep a measurement parameter.

• XYZ of Oscilloscopes, Tek公司提供的64页关于示波器基础的教程
• 示波器基础, Rohde & Schwarz
• 示波器的应用
• 示波器基本指南
• 示波器教学视频
• 数字存储示波器测量基础
• 示波器在线演示

此次提供两款双通道示波器的参考设计，第一款为元器件分销商Arrow Electronics Inc.基于Altera Corp公司的MAX10平台提供的双通道示波器参考设计BeScope，主要技术指标如下：

• 50MHz模拟带宽
• 采用Analog Devices Inc公司的AD9286, 双通道同时采样，每通道采样频率为250MSPS
• 具有三级增益设置的可编程增益放大器
• 输入信号范围±60V(采用10x示波器探头设置, 最低增益)
• 分辨率最低到12mv/LSB(1x示波器探头设置, 最高增益)
• 2.5MHz和5MHz方波信号产生
• 采用25K逻辑单元的FPGA用于控制和数据分析
• 128MB DDR3用于波形存储
• Altera Qsys互连框架方便FPGA逻辑的修改

相关技术资料：

• Arrow基于Altera的MAX10提供的双通道示波器参考设计
• Arrow基于Altera的MAX10提供的双通道示波器参考设计物料清单BOM表
• Arrow基于Altera的MAX10提供的双通道示波器参考设计PDF格式的原理图

Digilent公司推出的Analog Discovery 2

Digilent Analog Discovery 2 是一款基于USB的示波器／多功能仪器，用于测量、可视化、产生、录制以及控制所有种类的混合信号电路。模拟和数字信号的输入／输出可以通过简单的连线同待测的电路进行连接也可以通过提供的BNC适配器和BNC探头进行连接。

主要性能指标:

• 双通道USB数字示波器(1MΩ, ±25V, 差分, 14-bit, 100Msample/sec, 30MHz+带宽 - 使用Analog Discovery带的BNC适配板)
• 双通道任意函数发生器(±5V, 14-bit, 100Msample/sec, 20MHz+带宽 - 使用Analog Discovery带的BNC适配板)
• 立体声音频放大器用以驱动外部的耳机或喇叭
• 16通道数字逻辑分析仪(3.3V CMOS, 100Msample/sec)
• 16通道模式发生器(3.3V CMOS, 100Msample/sec)
• 16通道虚拟数字I/O包括按钮、开关、LEDs – 非常适用于逻辑训练应用
• 两个输入／输出数字触发信号用以连接多个设备(3.3V CMOS)
• 单通道电压表(AC, DC, ±25V)
• 网络分析仪 – Bode, Nyquist, Nichols transfer diagrams of a circuit. Range: 1Hz to 10MHz
• 频谱分析仪 – power spectrum and spectral measurements (noise floor, SFDR, SNR, THD, etc.)
• 数字总线分析仪(SPI, I²C, UART, 串行)
• 两个可编程电源(0…+5V , 0…-5V). 最大可输出电流以及功率取决于Analog Discovery 2的供电选择:
  • 通过USB供电，可以每一路提供最大250mW或总计500mW
  • 通过外部电源适配器每一路可以提供最大700mA或2.1W

Analog Discovery 2 参考手册

基于树莓派的双通道高速ADC的设计