Why most oscilloscopes ground its reference to earth?

[parachutes30]

Usually a isolated tranformer is needed when we use oscilloscope in a circuit non isolatedly derived from mains power. So why oscilloscope is designed as its reference is gounded to earth? just because BNC is metal and for safety?

 

[c_mitra]

You use an isolation transformer as an insurance of your personal safety. Nothing more.

However, consider a circuit that has a ground different from the mains ground. So to measure the potential at any point in the circuit, you need to connect your probe ground to the circuit ground and the test point to the probe point. Your scope ground is referenced to the power plug earth point (they are connected via the chassis) and the ground of the test circuit is at some other potential with respect the the earth ground. You cannot short them because if they are at different potentials, a large current will flow and there will be lots of headache for you!!

Now consider you are running your test circuit via an isolation transformer. In such a case, the ground in the test circuit is at an undefined potential (simply floating) and if your probe ground is connected to the test ground, there will be no current and no danger. (There is simply no current path).

Best way is to use dual probe; one probe you connect to the ground and the other probe you connect to the test point. Set your scope input to A-B mode (you will see only one trace on the scope). But please check the common mode rejection number for the scope inputs.

You say "Usually a isolated tranformer is needed when we use oscilloscope in a circuit non isolatedly derived from mains power" which is not correct. If you are sure that they share the same ground and you are not interchanging the live and ground points, you can go ahead. But take care always when you are dealing with mains voltages.

 

[danadakk]

The prior post mentions using A-B mode to create a differential measurement.

If you measurement requires high CM rejection the scope may not be able to meet
that in that mode. Especially at higher frequencies than line rates. Usually one in that
case uses a differential probe specifically designed for high performance.

https://www.tek.com/document/primer/abcs-probes-primer

https://www.keysight.com/us/en/assets/7018-03057/data-sheets/5990-8576.pdf

https://www.newark.com/pdfs/techarticles/tektronix/FFM.pdf

Regards, Dana.

 

[parachutes30]

thanks! oh.... I want to know why oscilloscope connect its reference ground of channel to the chassis/earth? I think all the measurement or control circuits of oscilloscope's channel is already on the secondary side of its only power transformer and it's already isolated from the mains power. Just making reference ground connected to the DUT works well....

You use an isolation transformer as an insurance of your personal safety. Nothing more.

However, consider a circuit that has a ground different from the mains ground. So to measure the potential at any point in the circuit, you need to connect your probe ground to the circuit ground and the test point to the probe point. Your scope ground is referenced to the power plug earth point (they are connected via the chassis) and the ground of the test circuit is at some other potential with respect the the earth ground. You cannot short them because if they are at different potentials, a large current will flow and there will be lots of headache for you!!

Now consider you are running your test circuit via an isolation transformer. In such a case, the ground in the test circuit is at an undefined potential (simply floating) and if your probe ground is connected to the test ground, there will be no current and no danger. (There is simply no current path).

Best way is to use dual probe; one probe you connect to the ground and the other probe you connect to the test point. Set your scope input to A-B mode (you will see only one trace on the scope). But please check the common mode rejection number for the scope inputs.

You say "Usually a isolated tranformer is needed when we use oscilloscope in a circuit non isolatedly derived from mains power" which is not correct. If you are sure that they share the same ground and you are not interchanging the live and ground points, you can go ahead. But take care always when you are dealing with mains voltages.

--- Updated Aug 9, 2020 ---

oh... My DUT is already isolated from the mains, but I get poor ringing wave in one channel measuing a low voltage when a fast rising (5ns) high voltage pulse occurs on the second channel on a normally eathed oscilloscope.
some guy recommend me to use a isolated power for oscilloscope. it seems that he want break the ground loop, but maybe a differential probe is what i need.

The prior post mentions using A-B mode to create a differential measurement.

If you measurement requires high CM rejection the scope may not be able to meet
that in that mode. Especially at higher frequencies than line rates. Usually one in that
case uses a differential probe specifically designed for high performance.

https://www.tek.com/document/primer/abcs-probes-primer

https://www.keysight.com/us/en/assets/7018-03057/data-sheets/5990-8576.pdf

https://www.newark.com/pdfs/techarticles/tektronix/FFM.pdf

Regards, Dana.

 

[KlausST]

Hi,

but I get poor ringing wave in one channel measuing a low voltage when a fast rising (5ns) high voltage pulse occurs on the second channel on a normally eathed oscilloscope.

If this is cross talk or GND current through the jigs then no additional supply can cure this.

some guy recommend me to use a isolated power for oscilloscope

Using an isolated power supply for the scope creates a safety problem. Thus I don't recommend this.
Any metal piece on the ocsilloscope may carry dangerous voltage. Screws, case, interfaces (USB, RS232...), connectors, handle...

Yes, I know it is done many times. Hopefully your "some guy" feels responsible in case of an accident...but usually those persons deny it...

Klaus

 

[c_mitra]

That is more complex than this...

Consider an ideal transformer that has primary fed from 220VAC and has a single secondary (say) 10 VAC.

The secondary is floating: the two free terminals of the secondary has potential U1 and U2. However, they are undefined (because they are floating).

Of course you know U2-U1 is 10VAC. That simply means that these two terminals are floating but are not independent.

What is the common mode voltage? Usually we define the common mode voltage as (U1+U2)/2. As you see, the common mode voltage is also undefined.

Can you measure the voltage at the two terminals independently (say with a multimeter or a scope)? The answer is No! You can measure only potential difference. Absolute potential has no meaning whatsoever.

So you connect one terminal to the scope ground (so the floating fellow is now grounded)- there is no fear of any danger.

And connect the probe tip to the other terminal: you see a nice 10 VAC on the screen.

Disconnect the ground connection of the probe from the transformer terminal: you shall see noise on the scope (more likely the power line 50Hz pickup or something similar).

I hope this is now clearer.

 

[FvM]

oh... My DUT is already isolated from the mains, but I get poor ringing wave in one channel measuing a low voltage when a fast rising (5ns) high voltage pulse occurs on the second channel on a normally eathed oscilloscope.

You can be pretty sure that the problem won't be solved by isolating the oscilloscope ground. Similar common mode noise will be also observed with an isolated ground due to the unavoidable ground capacitance. It can be reduced in some cases by using ferrite toroids as common mode chokes for the probe cables, but ultimately you have to use good differential probes. Another approach is by using oscilloscopes with individually isolated inputs like this instrument https://cleverscope.com/news/cs448/

Here's a photo of the common mode choke

 

[c_mitra]

"My DUT is already isolated from the mains, but I get poor ringing wave in one channel measuing a low voltage when a fast rising (5ns) high voltage pulse occurs on the second channel on a normally eathed oscilloscope." The ringing may be real or not, but by now you must have realized that the problem is more complex that you originally thought.

Yes, the problems you mention are common when you reach high frequency. By the way, what is the meaning of high frequency?

In the present context, the high frequency range or limit is described by your measuring instruments. Also you need to focus on ALL stray couplings (not only to grounds but also intercomponent) that may be present. The truth is that we and you do not know but we can guess. But first thing that needs to be attended to is to limit all high frequency components that the measuring instrument cannot handle. Yes, I know that the 5ns rise time (from a signal generator) looks very nice when tested on the same scope. But the stray coupling are your problem, not the scope's! Many signals have this bad habit of jumping from one place to another.

I am not an expert and wise men have spoken but ground problems are often more complex than they appear. Your case is only an example.

 

[smithdonald078]

The external case must be grounded but input not at all

 

[c_mitra]

The external case must be grounded but input not at all

The external case is grounded for personal safety. This is connected to the ground pin of the power plug.

The signal input is most commonly a grounded shell BNC connector; some scopes has an isolated connector and the ground is connect to the circuit ground via a 100R or 1K resistor via a switch. If the switch is pressed, the input ground is directly connected to the circuit ground.

No meaningful measurements can be carried out if the input ground is left unconnected. ALL voltage measurements need a reference point. That is not the case for current measurements.

From a practical viewpoint, a floating potential has no significance; only potential differences has.

Of course you can use A-B mode without using the ground connection at all; but this comes with a penalty.

 

[schmitt trigger]

My personal experience:

Not all isolation transformers are created equal.
Some cheap stuff, the primary and secondary are wound one on top of each other on E-I cores.
The capacitance between those can be significant, and while at 50/60 Hz the leakage is very low, at higher frequencies it could be several milliamps.
I once measured one which had over 900 pF between windings! That is only 3.5k ohm reactance at 50 Khz!

There are better construction techniques and shielding that improve upon that, but the transformer becomes very expensive, very quickly.

If you plan on doing off-line SMPS work, consider purchasing a differential probe. It is worth every single Dollar/Euro/Pound/Yen you pay for it.

 

[std_match]

As c_mitra says, it is for the personal safety.

The ideal solution is that every scope input channel is isolated from the other channels and ground, but this makes the scope much more expensive. With a normal scope, it is better to have a fuse blown when you connect a probe ground to a "bad" place than to electrucute yourself when you grab the ground lead for one of the other channels. For such measurements, get diff-probes or a scope with isolated input channels. Please note that most diff-probes aren't isolated, but they have a large common-mode range and a high resistance (MegaOhms) to the scope ground.