How to probe audio amp output with a scope?

[Zbig]

Hello everyone

For no particular reason other than hoping to see some WinAMP-style audio signal visualizations on my scope, I thought I'll just stick a scope probe across one of the amp's speaker outputs. Then I had some second thoughts. Is it okay to do just that or perhaps it's not the wisest thing to do? I mean having in mind that the scope has its probe ground tied to the mains ground and the audio signal is bipolar, am I in a risk of doing any harm to my amp? Does it depend on the way the amp is powered from mains and the amp class?

[Strada916]

Not sure about modern scopes but I was always told to isolate the scope from mains and ground when in doubt. 

[Psi]

Connect the scope to your amps audio line output (L or R channel).
At least with that it's obvious which is GND and which is signal, and if you short a line level signal nothing bad will happen.

Isolating a scope is usually a bad idea, it works but can be quite dangerous.
If you attach the scope ground clip to say 220V for example, then you're entire scopes metal case is now at 220V
It's much better to isolate the device under test, it can be tricky though if you have something that's wired up to lots of other things (like an amp wired to TV,PC etc.) as they will all be providing a earth path to the amp.

[fcb]

Watchout if you scope an active speaker - they often use a bridge driver, so both sides swing (if you like).

[G7PSK]

Just isolate the probes with a couple of capacitors 50 or so uF  would do. You only really need a capacitor in the earth lead, I tried a one meg resistor in the earth lead once but for some reason it caused the trace to become curved, so that the sine wave I was looking at appeared in an arc across the screen, I have no idea if this would be the case with a DSO as I only have a CRO but certainly you can look at audio with the scope probes isolated with a couple of capacitors.

[Zbig]

Thank you for the hints, guys.

[807]

For 100% isolation, you could always connect a microphone to the scope! 

[mzacharias]

I repair audio equipment for a living the past 30 years. Most home audio units are common grounded, and this is easily confirmed by using a multimeter to check for near zero ohms between speaker GROUND connection and the chassis, and also to the input ground if there is any doubt about them being the same.
A load resistor, typically 8 ohms at 20 watts or more, is connected to the speaker terminals and the 'scope negative lead goes to chassis, and the (usually x1) probe tip or oscilloscope input can be directly connected to the POSITIVE speaker output terminal, or at the resistor itself. A x10 probe is not necessary at this frequency but may be used if desired.
If the output of the amp is not common grounded, you may still connect the 'scope ground to the chassis and one positive probe to a speaker terminal, and if so desired you may use two probes in differential mode (inverted on one) to measure each channel.
In this case the load resistors must have all separate wiring and not be grounded to the scope chassis.
I keep a dedicated 'scope with load resistors directly wired to 5-way dual binding post jack-to-BNC adapters directly plugged into the channel 1 and 2 inputs of my 'scope. Somewhat dangerous for the vertical amplifiers if something went massively wrong, but I have measured amplifiers in excess of 300 watts per channel in this manner with no problems. Suitable 'scopes for this have vertical ranges of 10 or even 20 volts/division. Some older B&K model go to 20 volts/division. The Philips 50mHz I'm presently using goes to 10 volts/division. 5 volts per division is acceptable but you will have to take it off cal to look at full power of larger amps and still keep the full waveform on screen.

[staze]

I repair audio equipment for a living the past 30 years. Most home audio units are common grounded, and this is easily confirmed by using a multimeter to check for near zero ohms between speaker GROUND connection and the chassis, and also to the input ground if there is any doubt about them being the same.
A load resistor, typically 8 ohms at 20 watts or more, is connected to the speaker terminals and the 'scope negative lead goes to chassis, and the (usually x1) probe tip or oscilloscope input can be directly connected to the POSITIVE speaker output terminal, or at the resistor itself. A x10 probe is not necessary at this frequency but may be used if desired.
If the output of the amp is not common grounded, you may still connect the 'scope ground to the chassis and one positive probe to a speaker terminal, and if so desired you may use two probes in differential mode (inverted on one) to measure each channel.
In this case the load resistors must have all separate wiring and not be grounded to the scope chassis.
I keep a dedicated 'scope with load resistors directly wired to 5-way dual binding post jack-to-BNC adapters directly plugged into the channel 1 and 2 inputs of my 'scope. Somewhat dangerous for the vertical amplifiers if something went massively wrong, but I have measured amplifiers in excess of 300 watts per channel in this manner with no problems. Suitable 'scopes for this have vertical ranges of 10 or even 20 volts/division. Some older B&K model go to 20 volts/division. The Philips 50mHz I'm presently using goes to 10 volts/division. 5 volts per division is acceptable but you will have to take it off cal to look at full power of larger amps and still keep the full waveform on screen.

Is there any reason, if you're just looking at the waveform, you couldn't just connect the scope probe to the "red" speaker output, and the ground to the "black" speaker connection? The scope is generally going to be 1Mohm, so you're not going to over-tax the amp...

Just curious... I'm working on a stereo receiver right now, and this is how I'm hooked up (well, I have a BNC to banana adapter, and am running wire directly from the speaker terminals, to the banana jacks), omitting the probe.

Thanks!

[mzacharias]

Connecting the scope ground to the speaker ground works just fine PROVIDED the ground is really ground.

A bridged output amp would not be common grounded. Traditionally very few home receivers are bridged, but more and more of the newer digital amplifiers are essentially bridged ( a "carrier" on both wires, signal only on the positive). Why they feel they have to do this is beyond me.

It's just safer to connect the scope ground to chassis ground until you know for sure. Use an ohmmeter to check the grounds if you have any doubt.

[staze]

Connecting the scope ground to the speaker ground works just fine PROVIDED the ground is really ground.

A bridged output amp would not be common grounded. Traditionally very few home receivers are bridged, but more and more of the newer digital amplifiers are essentially bridged ( a "carrier" on both wires, signal only on the positive). Why they feel they have to do this is beyond me.

It's just safer to connect the scope ground to chassis ground until you know for sure. Use an ohmmeter to check the grounds if you have any doubt.

Makes sense... and yeah, bridged doesn't make a whole lot of sense... but hey, I'm no audio engineer. Maybe it's just easier/cheaper than running ground all over the place?

Anyway, I did confirm that the negative speaker terminal is the ground, but even so, I clipped to the ground lug on the back of the amp.

Any thoughts on my question here: https://www.eevblog.com/forum/chat/stereo-repair/msg268851/#msg268851 ?

[broo0d]

What happens if we use a battery powered oscilloscope, do we still need to connect the ground clip to the amp chassis instead of the - ?

[rs20]

What happens if we use a battery powered oscilloscope, do we still need to connect the ground clip to the amp chassis instead of the - ?

If you have an isolated jack oscilloscope (like most battery powered oscilloscopes are), you can safely connect the ground clip of the probe to the -.

[broo0d]

What happens if we use a battery powered oscilloscope, do we still need to connect the ground clip to the amp chassis instead of the - ?

If you have an isolated jack oscilloscope (like most battery powered oscilloscopes are), you can safely connect the ground clip of the probe to the -.

By saying "isolated jack oscilloscope" do you mean isolated from mains earth?

[rs20]

By saying "isolated jack oscilloscope" do you mean isolated from mains earth?

Isolated from anything else connected to the audio amp, including mains earth yes.

[broo0d]

By saying "isolated jack oscilloscope" do you mean isolated from mains earth?

Isolated from anything else connected to the audio amp, including mains earth yes.

Thank you very much, this information would really help.

[tggzzz]

Not sure about modern scopes but I was always told to isolate the scope from mains and ground when in doubt. 

And then, if something isn't behaving in the way you guess, the scope's chassis and exposed metal can be at high voltages.

For how to do it properly, read the professional references in the "Praxis" section of https://entertaininghacks.wordpress.com/library-2/scope-probe-reference-material/

[tggzzz]

Isolating a scope is usually a bad idea, it works but can be quite dangerous.

Yes - but sometimes it doesn't even work, e.g. if you connect the ground to a point that is sensitive to capacitance or noise.

[migueldelag]

I repair audio equipment for a living the past 30 years. Most home audio units are common grounded, and this is easily confirmed by using a multimeter to check for near zero ohms between speaker GROUND connection and the chassis, and also to the input ground if there is any doubt about them being the same.
A load resistor, typically 8 ohms at 20 watts or more, is connected to the speaker terminals and the 'scope negative lead goes to chassis, and the (usually x1) probe tip or oscilloscope input can be directly connected to the POSITIVE speaker output terminal, or at the resistor itself. A x10 probe is not necessary at this frequency but may be used if desired.
If the output of the amp is not common grounded, you may still connect the 'scope ground to the chassis and one positive probe to a speaker terminal, and if so desired you may use two probes in differential mode (inverted on one) to measure each channel.
In this case the load resistors must have all separate wiring and not be grounded to the scope chassis.
I keep a dedicated 'scope with load resistors directly wired to 5-way dual binding post jack-to-BNC adapters directly plugged into the channel 1 and 2 inputs of my 'scope. Somewhat dangerous for the vertical amplifiers if something went massively wrong, but I have measured amplifiers in excess of 300 watts per channel in this manner with no problems. Suitable 'scopes for this have vertical ranges of 10 or even 20 volts/division. Some older B&K model go to 20 volts/division. The Philips 50mHz I'm presently using goes to 10 volts/division. 5 volts per division is acceptable but you will have to take it off cal to look at full power of larger amps and still keep the full waveform on screen.

Hi, saw your post, just trying to get things right, i made a diagram, could you confirm if ok? My idea is to have the speakers connected while the scope gets their signal. Thank you (: