the timebase
The timebase is the oscilloscope's horizontal control: it decides how much time fits across the screen. Picture an old chart recorder where paper scrolls under a pen — the timebase is the speed of that paper. Wind it fast and the screen covers a tiny sliver of time, so a single fast edge spreads out wide and you can study its shape. Wind it slow and the screen spans a long stretch, so a slow heartbeat-like signal shows many cycles at once. Setting it well is the difference between seeing a useful picture and seeing a meaningless blur or a single frozen line.
It is calibrated in time per division — the duration of one horizontal grid square. A typical screen is ten divisions wide, so the whole window is ten times that setting. At 1 ms/div the screen shows 10 ms end to end; a 1 kHz signal has a period of 1 ms (period = 1/frequency), so you would see ten complete cycles march across. Want to examine just one cycle in detail? Speed up to 0.1 ms/div and that single cycle now fills the whole screen. The rule of thumb: to count cycles, slow down; to inspect an edge, speed up.
Why it matters: the timebase is half of every scope measurement (the vertical scale is the other half). You read a signal's period and frequency straight off it — count the horizontal squares for one cycle and multiply by the time per division. Two honest notes. The timebase only freezes a repeating wave into a still picture when the trigger is set correctly; a good timebase with a bad trigger still gives you a smear. And a digital scope draws from samples, so if you slow the timebase too far on a fast signal the screen can show a misleading, aliased shape that is not really there.
A signal shows one full cycle spanning four horizontal divisions. The timebase is 50 us/div, so one cycle lasts 4 times 50 us = 200 us. The frequency is 1/200 us = 5 kHz. Reading frequency off the screen is just counting squares and a little arithmetic.
Time per division sets the window — slow to count cycles, fast to study an edge.
Many beginners get a flat line and assume the circuit is dead, when really the timebase is just far too fast or too slow to show that signal. Sweep the timebase across its range before declaring anything broken.