O'scope ground probe

Thread Starter

cng2018

Joined Jun 6, 2018
5
Hello,
There is a spark when I connected O'scope Ground Probe to AGND pin on the circuit board (DUT). I then measured between AGND and chassis GND is about 28Vac by Fluke meter. So there is a float voltage between those points.
In order to avoid that spark, can I put a 2-pronge outlet from O'scope input voltage to isolate the ground? If I want to keep the 3-pronge on the outlet, what should I use on my Probe to avoid the spark?
Thanks for your time
 

dendad

Joined Feb 20, 2016
4,451
Some more info would help.
What is the DUT in question?
Is the 28VAC because of design, a fault, or is it the capacitively coupled mains from a switch mode PS?
Do you have a dual channel scope?
If so, you can ground the probes to chassis and hook one channel to the AGND, and the other to the signal, and switch to differential input.
 

DickCappels

Joined Aug 21, 2008
10,153
It would be far safer to ground the DUT before applying power or connecting any test equipment.

The 28V you see is probably the result of a small leakage current through the radio frequency interference (RFI) filter in the power supply for the DUT.
 

Thread Starter

cng2018

Joined Jun 6, 2018
5
Hello,
Thanks for all response.
1. The Model of Scope is TDS 210.

2. The 28VAC is the design. According to the schematic, 75Vac [Line and Neutral] is applied to isolate primary Transformer. The low side of Transformer is AGND where it is asked to connect ground probe to AGND, and another probe is asked to connect at different Test Point to measure the waveform. But like I said, when connected ground probe to AGND, it is sparked. So I checked there is a 28Vac between my ground probe and AGND. Test Procedure is asked to connect Chanel 1 to Test point and AGND, not Chassis gnd ( note: Chassis ground and ground probe is the same ground)
 

DickCappels

Joined Aug 21, 2008
10,153
So...your DUT's analog ground is elevated to +28 volts, and you want to know if it is safe to float your scope at +28V. +28V is considered a low enough voltage to be safe for a person to come in contact without much danger provided the person is not standing in a puddle of saltwater or something similar.

It is probably safe to float your entire scope at +28V. What, if anything does Tektronix say about the paractive?

Back in the 1970's we would occasionally float out old tube-type scopes on the AC power line for various reasons. We stood on rubber mats, kept "one hand in out back pocket" and did not do that if there was nobody else in the lab. But that was back then with those old scopes.
 

ebp

Joined Feb 8, 2018
2,332
Try measuring the current between the scope ground and AGND using your meter, but with the voltage ranges and an external resitor:

Start with a resistor of about 10k ohms. Connect it across the meter (if you have one, a dual banana plug/jack with a way of connecting wires to it is a good way to do this). Set your meter for volts and measure between the scope ground and AGND. If the voltage is the same as it was without the resistor, you must stop at this point. If the voltage is reduced, calculate the current though the resistor. If the voltage is significantly reduced, you can then test with a lower value resistor, such as 1k ohms and again calculate the current. It may be the same as it was with 10k or it may increase. Be careful, because if the current goes up you may exceed the power rating of the resistor. As long as it isn't too much above the rating this is OK for the few seconds it will take to make the measurement.

If the current flow is more than a milliamp or two, and it sounds like it will be, then the safe way to proceed is to figure out how that current can be flowing.

In general it is dangerous to disconnect an oscilloscope's safety ground from a good earth ground. This can allow the entire oscilloscope chassis to rise to a voltage that could be dangerous, making it unsafe to touch the oscilloscope. I have done this, but fully understanding the risks and being extremely careful. If there are other people working around you I advise against it in the strongest terms. As Dick says, 28 volts is generally considered to be "safe" (especially if it is DC), but getting into the habit of disconnecting the scope safety ground to avoid doing something the right way is a very bad thing.

You also need to consider the possibility that what you think is earth ground for the oscilloscope is not. If you are in your usual working environment you probably do know for sure your ground is OK, but if you are working somewhere else there may be a problem. What you find as a 28 volt difference at one time might turn out to be a hundred volts some other time, and then things really do get dangerous.

75 VAC is rather unusual. Where is it coming from?
 
Last edited:

ebp

Joined Feb 8, 2018
2,332
Yes, a differential measurement would probably work (certainly would with an appropriate differential probe, less well by invert-and-add in the scope), however the TS is apparently following written instructions for how to connect the oscilloscope and something unexpected is happening. That unexpected thing is a very serious problem and should not be casually worked around.
 

rsjsouza

Joined Apr 21, 2014
383
A spark may happen even with very low energy differential (28V but with micro amps of current) - as ebp has suggested, measuring the current with a series resistor is a great way to know how much energy you have between your oscilloscope ground probe and AGND.

You mentioned your design has a 28Vac transformer tap connected to the AGND net, but where is the other side of the transformer? Is it connected to the chassis ground? If so, I would be extremely careful in connecting the oscilloscope ground probe as you will, in essence, short the output of this transformer section (and the current measured by the method above will be significant). If the schematics show a completely isolated chassis ground, then you may have an assembly or mechanical fault that is inadvertently connecting the chassis to a given net in the circuit.

At any rate, the differential method will work for you in any case. Also, a schematic diagram would be helpful to better understand this scenario.
 
Top