1.. Always use "10 X" probes: they load the DUT (device-under-test) ~ 10 Meg ohms @ ~ 10 pfd. A "1 X" probe offers 1 Meg ohm @ ~ 50 pfd. The designation "10 X" refers to the attenuation of the signal by the probe (not gain). In order to attain such light loading by the scope--while maintaining bandwidth--this tradeoff is required.

2.. Make sure the probes are compensated (adjust trimmer at connector housing) if attaching them to a different scope. This ensures maximum fidelity and bandwidth of the signals being eyeballed.

3.. Use the shortest ground lead or clip-lead possible: the shorter the better! Excessive ground lead length introduces unnecessary inductance and can alter the displayed signal, as well as reducing the scope's effective bandwidth (acts like a lowpass filter).

4.. When Measuring very high frequencies--especially in tight spaces--consider using a RF probe (see figure). Also, there are--so-called--FET or active probes, which are non-loading (almost) wideband probes with built-in amplifiers.

5.. When buying probes for your oscilloscope, make sure the probe is of sufficient bandwidth for your particular scope: the probe is the first-order bandwidth determinant of any scope.

Some scopes have such a wide bandwidth, that no passive probe is able to do it justice, and the only way to use the maximum bandwidth of this type of scope is to drive the scope from a 50 ohm source through a 50 ohm coax, terminated into 50 ohms at the scope's input. In fact, some high performance scopes have a 1Meg ohm/50 ohm termination switch for just such occasions.
 

------

Something that cheap scopes are guilty of, is the fact that you never see the leading edge of the event that triggered the sweep. That is to say, the sweep is triggered by the leading edge of a pulse, and by the time the beam starts across the screen it is now displaying the signal ~ 100 nsec after the event. To solve this little "hand-is-quicker-than-the-eye" problem, the Keebler Elves added wideband analog delay lines between the vertical amplifier and the CRT's deflection plates. So, by the time the beam has started its short trip across Mr. CRT's face, the signal that triggered the sweep is finally dribbling out of that long delay line.

There is a sweep begun whenever the signal reaches the trigger level, regardless of whether it is a rising or falling voltage. This means that--in the case of a sine wave--the scope is just as apt to start the display at zero degree phase, as 180 degree phase. To solve this little oversight, a switch is used to tell the sweep generator which excursion--rising or falling--to trigger on: this is called the "SLOPE." (+) SLOPE (-)

The figure above show several combinations of settings.

Let's say I wanted to eyeball the 188th horizontal scan line, and I had dialled in both the "Delay" and the "Sweep" intervals: as soon as the field sync pulse (VD) occurs the delay sweep starts to time out for 11,885 usec (187 x 63.5), at which point it triggers the scope's sweep generator. The sweep runs for ~ 65 usec--displaying one complete horizontal scan line (the 188th).

Other features of the oscilloscope that are available, are left for the user to discover in the user's manual: you know, that thing you read when "all-else-fails!"