Floating voltage measurements with oscilloscopes?..ok with floating power source?

[grizedale]

Hello,

The following doc, on page 2, states that floating measurements should NOT be taken using an oscilloscope and passive probe because a current will flow through the scope probes reference terminal.......

Fundamentals of Floating Measurements and Isolated Input Oscilloscopes:-
**broken link removed**

.............but surely most power supplies that supply test equipment are isolated, so it simply will not matter to which node one connects the scope probe "ground" clip?...and no such current will flow in the scope probe wire

 

[bhav]

yup u r right

 

[godfreyl]

.....but surely most power supplies that supply test equipment are isolated, so it simply will not matter to which node one connects the scope probe "ground" clip?

Page 4 explains why that's a bad idea.

 

[grizedale]

But surely Page 4 says nothing about a floating power source supplying the circuit under test?

 

[godfreyl]

No, there's no problem if the oscilloscope is grounded and the circuit under test is floating. When you said "surely most power supplies that supply test equipment are isolated", I thought you were referring to the oscilloscope. Sorry for the misunderstanding.

 

[grizedale]

Sorry about my unfortunate choice of descripture.

Anyway, So i have a TDS210 scope that's plugged into the mains, and having its probe's ground clip clipped to any node of a small circuit that's being supplied by an isolated power supply.

I am mostly scoping between one node thats at 10V and another node thats at 240V (that voltage being with reference to the circuits gnd PCB trace).

.....so i am clipping the scope ground clip to the 10V node, and the probe tip to the 240V node....
-this seems to have caused no problems, however, once i got the probe connections the wrong-way-round and clipped the probe ground clip to the 240V node, and the probe tip to the 10V node.....

.....as soon as i touched the probe gnd clip to the 240V node there was a cracking sound, as if the probe gnd clip had suddenly discharged a capacitance.

Why did this happen?..and which capacitance (stray) was discharged?
would this have damaged the scope or the scope probe?

...and do i really need a differential probe? (hopefully not because they are $300)

 

[godfreyl]

.....as soon as i touched the probe gnd clip to the 240V node there was a cracking sound, as if the probe gnd clip had suddenly discharged a capacitance.

Why did this happen?..and which capacitance (stray) was discharged?
would this have damaged the scope or the scope probe?

I don't really know, but it sounds like there is a small capacitance between the isolated supply's output and ground. I doubt that little discharge caused any harm, as it just went through the ground wiring, not any circuitry.

I would certainly have an allergic reaction to a price of $300 for a probe. That raises the question of how hard it would be to make one. I'm pretty sure a rudimentary high voltage differential probe could be put together for a few dollars. It wouldn't be overly accurate, but good enough for a lot of tasks where high accuracy and speed aren't important.

 

[jiripolivka]

Oscilloscopes that have their chassis connected to ground cannot be used to observe "floating" voltages; in principle, "floating" means that one or another point of the measured voltage source has a voltage above ground.

The only good option is to use a battery-operated oscilloscope, or, an isolation transformer in AC power line. The transformer capacitance between its primary and secondary can affect the measured voltage, and you will need to know certainly that its insulation can stand the "floating" voltage. If a battery-powered oscilloscope is used, again a caution is to be exercised as the "floating" voltage will exist between the tested circuit and the ground in the laboratory; now the oscilloscope ground (like the probe ground lead and clip) will carry the tested-circuit voltage above laboratory ground. High voltage >30 V is dangerous, and the complete oscilloscope circuit must be handled with an extreme caution.

I would advise against trying a differential probe as such probes can only "float" e.g. 10-20V above laboratory ground. Recently, Fluke and others introduced DVMs with wireless transfer of measured voltage to a remote indicator module. Such devices can help to measure floating voltages at a high-voltage "virtual ground". I am not sure how fast voltages can be measured and displayed on a connected PC.

 

[grizedale]

jiripoliivka:

I am sorry but i dont understand you there.

All oscilloscopes that are plugged into the mains have their chassis connected to earth ground.



When i was at Panasonic test department, it was standard procedure to have the scope lugged into the mains, but the circuit under test fed through an isolation trxfomrer, so oviously the circuit-U-T was floating, and that was Panasonic standard procedure so it could not be wrong?

And surely a diff probe is ok for floating voltage measurements, that is what they are designed for?

 

[jiripolivka]

jiripoliivka:

I am sorry but i dont understand you there.

All oscilloscopes that are plugged into the mains have their chassis connected to earth ground.



When i was at Panasonic test department, it was standard procedure to have the scope lugged into the mains, but the circuit under test fed through an isolation trxfomrer, so oviously the circuit-U-T was floating, and that was Panasonic standard procedure so it could not be wrong?

And surely a diff probe is ok for floating voltage measurements, that is what they are designed for?

1. All scopes plugged to the mains have their chassis connected to earth ground. If they use a standard AC cable with the grounding line, then most have their chassis connected to this AC ground and the probe ground lines. As the built-in AC/DC power supply is otherwise insulated, you can disconnect the AC grounding line and with a caution you can connect the probe ground to a floating point having up to ~100 V against the laboratory ground. A better option is to use the isolation AC transformer but you must be sure its insulation can hold the floating voltage above ground.

2. My best option is to use a battery-powered scope, again with a caution to that its probe ground could stay above laboratory ground.

3. If you have the described option of using an isolation transformer at the INPUT, it is possibly the best one, the input isolation transformer can only be used for the designed purpose; it is not suitable for any signal pattern at a low or high frequency, only to the one to be measured. Also such input transformer is the safest option if designed correctly,.

4. The differential probes are made for differential-voltage measurement, which may be floating at a SPECIFIED voltage above ground. Such specified "common mode" voltage is usually low, like 10 to 30 V. Most probes are used for a high-sensitivity measurements, for the display of 10 mV/ div. patterns. Special differential probes like the Panasonic input transformer can be made but such devices rarely conform to a typical scope input specification of DC to say 20 MHz or more.

 

[FvM]

I don't agree with the initial premise.

.............but surely most power supplies that supply test equipment are isolated, so it simply will not matter to which node one connects the scope probe "ground" clip?...and no such current will flow in the scope probe wire

- most D.U.T. have multiple connections, e.g data lines, other grounded test equipment
- even if the oscilloscope is the only external connection, you still have stray capacitances of power supply and enclosure. Connecting a grounded probe for an actually differential measurement injects common mode interferences in the D.U.T and potentially messes up the measurement results
- from a safety viewpoint it's usually unwanted to short the safety transformer's isolation by a grounded probe, particularly grounding a "live" voltage

 

[jiripolivka]

Everything depends on two important things:

1. How large is the floating voltage above ground
2. What is the frequency range and typical differential voltage swing desired to be displayed on the scope.

The INPUT differential transformer or probe can be designed to conform to the common-mode voltage while allowing to observe "typical" differential signal in a tested circuit.

For extreme cases, there are battery-operated isolation opamps; for extremely high common-mode voltages a differential preamplifier can be connected to the tested circuit, and its output sent by an optical fiber to a photodetector connected to the scope. All such devices must be designed to specific situations.

 

[grizedale]

jiripolivka:
sorry but i do not comprehend this part........

the input isolation transformer can only be used for the designed purpose; it is not suitable for any signal pattern at a low or high frequency, only to the one to be measured

........i dont see what you mean by talking of signals and isolation transformers, a mains isolation transformer is not for signals, its for mains.


I have worked at many companies ( big and small) that simply fed the circuit under test through an isolation transformer and then scoped it with an oscilloscope plugged into thte mains.

-the suggestion that this is wrong is to me, surely incorrect, i simply cannot believe that the huge companies where i worked , did not know what they were doing.....their engineers were 30 (sucessful) years experienced plus.

Once we had a contractor come in and say we were doing it wrong, but he was never able to specifically tell us what was the problem...............i think he was just making it up.

 

[jiripolivka]

jiripolivka:
sorry but i do not comprehend this part........



........i dont see what you mean by talking of signals and isolation transformers, a mains isolation transformer is not for signals, its for mains.


I have worked at many companies ( big and small) that simply fed the circuit under test through an isolation transformer and then scoped it with an oscilloscope plugged into thte mains.

-the suggestion that this is wrong is to me, surely incorrect, i simply cannot believe that the huge companies where i worked , did not know what they were doing.....their engineers were 30 (sucessful) years experienced plus.

Once we had a contractor come in and say we were doing it wrong, but he was never able to specifically tell us what was the problem...............i think he was just making it up.

I am sorry you feel offended by my statements. It is not my intention. I wanted to point out that if you want to use a scope to display a signal pattern between two points under a voltage floating above laboratory ground, you can either isolate the scope power supply from the ground, or use a battery-operated scope, or, use an INPUT transformer, or an isolation amplifier, a fiber-optical isolation device.

All the above devices have specific problems by which the scope cannot be used to its full capability which is displaying small signals as mV/div. over a wide frequency range, DC to say 20...50 or >100 MHz. Depending on the CMR voltage and signal parameters, you can design a specific device meeting your conditions. Most such designs limit the scope capabilities.