What is the purpose of the AC setting on oscilloscopes?

[Plow]

If I set the scope to AC and measure say a DC square wave it pulls the waveform down and inaccurately tries to average it out between positive and negative.

If I set it to DC it correctly measures only positive voltage when fed DC, but also seems to correctly measure the + - oscillations in AC?

I'm sure there's a reason for the AC setting, what is it?

[JJ_023]

From my understanding is primarily to filter out the DC component of your Signal.

[ataradov]

It is useful if you want to observe 10 mVpp noise on a 15 V power rail. With a DC mode, you will never see those 10 mV, since you will need to set really high vertical range and they will be in the ADC noise.

With AC, you can set high vertical gain and get full resolution.

[Andy Chee]

Consider a TTL square wave from 0V to 5V.

If you use the DC setting, the display will correctly show 0 to +5V.
If you use the AC setting, the display will incorrectly show -2.5V to +2.5V

HOWEVER

Let's say you're probing the RF stages of a superheterodyne radio.  In this case, there are probably all sorts of DC bias voltages everywhere that you want to ignore, because you're interested in the AC signals only.  Signals might be varying from 5.13V to 5.15V in one section, and 9.13V to 9.15V in another. 

In this case, AC setting on the oscilloscope is extremely useful.  It eliminates the 5V and 9V DC component and allows you to see the 0.03V to 0.05V signal (which will be displayed as -0.1V to 0.1V when AC setting is used).

I'm sure others will chime in with different use cases.

[David Hess]

It is useful if you want to observe 10 mVpp noise on a 15 V power rail. With a DC mode, you will never see those 10 mV, since you will need to set really high vertical range and they will be in the ADC noise.

With AC, you can set high vertical gain and get full resolution.

Some oscilloscopes have a DC offset function which allows high sensitivity DC measurements far from zero volts.  Effectively this extends the range of the vertical position control, although it is implemented differently.  On some oscilloscopes the offset and vertical position controls are the same thing.

Some old oscilloscopes took this to an extreme.  Tektronix had several models which operated at 10mV/div with an offset of +/- 100 volts, so it was like having a vertical position control with a range of 20,000 divisions.

[Psi]

If you looking at, for example, a 0.1V AC waveform on top of a 30V DC signal (so like a sinewave between 30.0V and 30.1V) and you want to enlarge that sinewave, then you might try adjusting the Volts per Div so 0.1V range fills up the entire screen.
But if you try that in DC mode the waveform would be like 3 meters off the top of the scope screen.
You'd have to increase the V/Div to see the signal again and you don't want that, you want to zoom in on just the AC part.
If the DC voltage was small you might be able to change the vertical adjustment knob to bring it back on screen, but that wont work with such a huge DC voltage difference to the AC voltage you want to look at. The range of the vertical adjustment knob is pretty limited. You can't keep on turning and turning it. It has a max.

So instead you use AC mode to remove this 30V DC from the signal entirely and effectively move this 0.1V AC waveform from 30V back down to the center of the scope around 0V, Now you can adjust the V/Div to zoom in on it without that problem.

[Peabody]

I don't know how professional scopes handle this, but on my small battery-powered hand-held scope AC mode is achieved by switching the input so it runs through a capacitor.  As a result, depending on what you're measuring, the waveform may be modified somewhat compared to what a true DC offset would show.

[Terry Bites]

The AC switch eliminates the DC component of the waveform.
In AC input mode a waveform will centre on the Y axis according to the "area under the curve" ie The average of the plus and minus bits.
A squarewave isn't DC, even if it goes from zero to some + volts its still AC.

[David Hess]

I don't know how professional scopes handle this, but on my small battery-powered hand-held scope AC mode is achieved by switching the input so it runs through a capacitor.  As a result, depending on what you're measuring, the waveform may be modified somewhat compared to what a true DC offset would show.

That is how old oscilloscopes do it, either with a manual switch or a relay because it has to support the full input voltage range.

Modern oscilloscopes usually do it a different way which only requires low voltage switching.  The high speed buffer is always AC coupled, and a low voltage low frequency DC restorer circuit is used to make a 2 path amplifier.  Now a low voltage low frequency switch can be used to select AC or DC coupling.

[EPAIII]

I don't know how professional scopes handle this, but on my small battery-powered hand-held scope AC mode is achieved by switching the input so it runs through a capacitor.  As a result, depending on what you're measuring, the waveform may be modified somewhat compared to what a true DC offset would show.

That's pretty much how it is done in most scopes. The size of the capacitor and the input impedance of the amplifier that followed were chosen to provide a fairly low frequency cut-off point so most AC waveforms were not distorted to any significant degree. I imagine that a fair amount of consideration was given to base band audio and video signals which started around 15 or 20 HZ. The 3 db down point was probably significantly lower, perhaps around 2-5 HZ.

I imagine that there are any number of variations on how the signal path(s) were implemented by different manufacturers, but the capacitor is probably king of the hill where DC removal is needed. Many, perhaps most scopes that I have seen use some kind of resistor divider at or near the input. This allows the active circuitry to be designed with relatively lower DC Voltage power rails. So a common arrangement was resistor attenuators followed by a capacitor that can be bypassed and then the input amplifier. The attenuators provided a relatively constant level to the amplifier, simplifying it's design. It is possible that the switchable capacitor could precede the resistor attenuators.

[radiolistener]

AC coupling mode removes DC component from the signal. In such way there is no need to use DC offset tuning to see signal ripple.

Also when signal has high DC offset, there is no way to use DC offset knob to see signal ripple, because oscilloscope DC offset tuning is limited. Using AC coupling helps to solve it.