Oscilloscope and grounding

[Cifrocco]

While trying to view signals on a switching power supply (the type used in computers, but a smaller version) I blew the power supply's fuse twice when I attached the scope's ground alligator clip to what I thought were ground points in the power supply.

I guess the scope's ground is not isolated from the AC power outlet? It's a model GOS-622G from Instek. Is there a safe way to view all the voltages in the power circuit?

 

[williB]

yeah , you gotta use a two pronged ( no ground) plug adapter on your scope..

 

[Klaus]

The safest way would be connecting the scope to the mains via a 1:1 isolation transformer.

 

[JimB]

Disconnecting the earth (ground) from the mains supply to the scope is a BAD idea, infact it is dangerous.

What you need in this situation is either a "differential probe", or use both channels of the scope and set the Y amplifier to A+B and one of the channels to Invert.
This has the effect of displaying the difference between the two inputs.

Now connect the earth clip to a TRUE earth point on your PSU under test and use the probes to connect to the circuit which can be above ground potential.

For lots of good info about using oscilloscopes, have a look at the Tektronix website, www.tek.com.
They have a selection of "primers", one called "ABC of Probes" gives information relating to this problem.

JimB

 

[eblc1388]

The safest way would be connecting the scope to the mains via a 1:1 isolation transformer.

Hardly. Removing the scope grounding connections, either by using a two-pronged plug or using isolation transformer to power the scope is a dangerous practice. The whole metal work of the scope can be at line voltage when one connects the probe to AC circuitry under test.

The safest method is to connect the circuit one wants to test, i.e." the switching power supply" via a 1:1 isolation transformer and use the scope as normal, with its earth lead connection intact. If this cannot be done owing to physical limitation, then proceed as what JimB had suggested.

 

[Nigel Goodwin]

Disconnecting the earth (ground) from the mains supply to the scope is a BAD idea, infact it is dangerous.

This is an often discussed subject!.

Basically it's down to personal preference, and the type of danger you perceive.

Personally I consider an earthed scope EXTREMELY DANGEROUS, I disconnect the earth lead on ALL my scopes, and mark them accordingly. This is common practice in the service trade, not only by engineers, but also by large multi-national electronics companies.

But you really need to know the reasons for things being earthed (which is only a good idea in some cases, and under some failure conditions), and the potential dangers involved in being earthed or not.

'Cifrocco' has already found one of the dangers in an earthed scope, and has survived a possibly fatal situation.

But as I mentioned above, it's all down to understanding what is going on, and why it's done that way - you SHOULDN'T disconnect the earth on ANYTHING unless you know exactly why you're doing it. If you don't have that knowledge, you shouldn't be messing about with isolation transformers either! - isolation transformers aren't an instant safety cure for all evils, they merely shift the danger slightly (whilst giving a false impression that it's now safe!) - you can be killed just as easily with an isolation transformer, just in a slightly different way!.

 

[JimB]

This is common practice in the service trade, not only by engineers, but also by large multi-national electronics companies.

This practice should be picked up and stopped by regular PAT Testing which is madatory for large companies, under HSE legislation.

'Cifrocco' has already found one of the dangers in an earthed scope, and has survived a possibly fatal situation.

I would take the opposite view here. "Cifrocco" has found the dangers of working on a piece of equipment which he does not understand.

In general, I am opposed to un-earthed mains powered equipment.
Even under normal no fault conditions, there is a small leakage current through the interwinding capacitance of the mains transformer. This results in the chassis of the unearthed equipment "floating" at about half mains voltage.
The source impedance of this 120v (in the UK) is very high and not dangerous to the average human. But if you brush the back of your hand against it, you will feel a tickle which can be disconcerting if you are not expecting it.
However, the floating 'scope is probably lethal to most small signal semiconductors, particularly cmos based types.

JimB

 

[Oznog]

The ground is there for your safety because any point you hook the probe ground to goes to the metal case. If that point has a voltage on it AND is referenced to ground, you could get shocked touching the scope controls.

I use a battery powered digital Fluke scope. It is quite handy, the case and controls are all plastic and rubber so there is no grounding issue provided the voltage is not high enough to start a breakdown somewhere. The DC transformer powering it when not running off the batteries is of course isolated too (provided the breakdown limit is not exceeded).

 

[Nigel Goodwin]

The ground is there for your safety because any point you hook the probe ground to goes to the metal case.

As I've already made clear, it all depends on your point of view!.

Your right hand on the earthed scope, and your left hand on a live wire is EXTREMELY dangerous - if the scope isn't grounded that danger is removed.

It's VERY debatable which is safest, a fully earthed environment, or a fully floating environment (no earth at all). Both have good points, and both have bad points, and both have been in and out of favour over the years. But, as I've also previously made clear, in either case a clear understanding of how and why it's done that way is required to make either of them as safe as possible.

 

[williB]

after i mistakenly connected the scope ground to the + 5V on a pc powersupply , i went searching for my two pronged plug , and a new scope ground wire ..

 

[Dean Huster]

And in one of the other discussions, I was one who had disagreed with Nigel.

"As I've already made clear, it all depends on your point of view!.

Your right hand on the earthed scope, and your left hand on a live wire is EXTREMELY dangerous - if the scope isn't grounded that danger is removed."

Your right hand on a scope floating at the live wire potential and your left hand on ground is EXTREMELY dangerous - if the scope IS grounded, that danger is removed. And there's the opposite point of view. Ground is all around us and I'd rather have everything I commonly touch to be grounded rather than at an unknown potential. That way, I know what the rules are before I get into the game.

Every top-notch manufacturer of electronic test and measurement equipment (Tektronix, Agilent, Fluke, Keithley, R & S, Gould, Philips, etc.)always ground their equipment via the power cord as do all manufacturers of computers, medical equipment and non-isolated AC-powered power tools. You only see the two-wire cords on el-cheapo products or those that are effectively double-insulated.

I worked for Tek for several years, and if an instrument came in with the ground on the power cord missing, even if deliberate, we were required to replace the cord at the customer's expense because of corporate liablility issues.

Every military organization in the U.S. requires ALL equipment operated by the mains to be grounded and not doing so or defeating the ground is a punishable offense.

The U.S. National Electrical Code requires ALL conduits, service panels, motors, enclosures, etc. to be grounded.

You have ground all around you and to let an oscilloscope "float" is courting disaster.

One of these days, maybe I'll convert you, Nigel!

Dean

 

[eblc1388]

When one gets killed simply by plugging in an oscilloscope into the mains, after someone delibrately removed the earthing connection, it will be a case for manslaughter.

Suddenly you will find all the corporate bosses have never know of such practices and without fail they can all produce document with clear instruction warning against such practice and everyone have read and understand the warning. The document even bear the signatures of the culprit.

 

[Dean Huster]

That's quite true. Some of our customers were x-ray repair companies, some independent, some factory-related, but all had the nasty practice of cutting the grounds (earths) of the power cords on their (typically) Tek model 434 scopes and floating them several hundreds of volts, even sometimes pushing one K, above ground. First of all, the power transformer in the scope isn't rated for that kind of chassis-to-line potential difference. Second of all, it makes that scope box lethal to touch, for x-ray machine voltages also have quite high current capabilities -- it's not just static voltage.

The x-ray techs would yell and scream if we replaced the power cord on their scopes. After all, it's standard practice in their company and industry to float the scopes like that and they'd just cut the ground prong off any new cord we installed within seconds of getting their scope back. Wasn't my problem. I'd give them a letter written by the Tek corporate attorneys for the CEO (and anything less than a CEO was unacceptable) of their company to sign, relieving Tek of all liability for allowing such an instrument without ground to leave the service center. Never have we had a CEO willing to do such a thing.

And you can bet that if an x-ray tech was ever injured or killed by floating a scope like that, the company attorneys could come up with documentation and company policy forbidding such techniques.

After this became a big issue with x-ray companies, who were also valuable customers, Tek developed an add-on accessory that allowed the scope to remain grounded while the accessory, double-insulated, would do the floating.

Dean

 

[eblc1388]

Thanks for sharing the view of the equipment manufacturers.

This is a public forum and when there is a safety issue involved, I think we should look after the vulnerables and the uninitiated to explain to them why such practice is wrong and the risks involved.

Despite what is said, people often choose to take the risk as its often the quickest and easiest way to get the job done.

I and my scope? Guilty as charged.

 

[awright]

Oscilloscope Grounding

While I agree that lifting the ground on instruments is a dangerous practice, I think the discussion has gotten a little off the trail here.

Cifrocco said he blew the power supply's fuse twice by connecting his 'scope's ground lead to what he THOUGHT were ground points in his test object. Perhaps the solution to his problem is to correctly locate the ground of his test object, not to create a dangerous situation by making his 'scope chassis capable of being connected to hot test points without blowing a fuse.

awright

 

[eblc1388]

Re: Oscilloscope Grounding

Hi awright, welcome.

Perhaps the solution to his problem is to correctly locate the ground of his test object, not to create a dangerous situation by making his 'scope chassis capable of being connected to hot test points without blowing a fuse.

Ground can be easily located on a switching power supply. It is simply the earth wire connection.

However, I wonder whether one can troubleshoot the high voltage switching circuit waveforms with a grounded scope without differential probes setup.

The high voltage circuit in the primary of the switching supply has no ground reference.

 

[Nigel Goodwin]

Re: Oscilloscope Grounding

While I agree that lifting the ground on instruments is a dangerous practice, I think the discussion has gotten a little off the trail here.

There seems a general 'understanding' that earthing something makes it safe? - this isn't TRUE - earthing something 'increases' it's safety under CERTAIN CIRCUMSTANCES ONLY. Under other circumstances it 'decreases' safety.

Notice I carefully said 'increases' and 'decreases', earthing something (or not) doesn't MAKE it safe - it just changes percentages!.

Cifrocco said he blew the power supply's fuse twice by connecting his 'scope's ground lead to what he THOUGHT were ground points in his test object. Perhaps the solution to his problem is to correctly locate the ground of his test object, not to create a dangerous situation by making his 'scope chassis capable of being connected to hot test points without blowing a fuse.

He was working on a circuit with a live chassis, so there is no 'ground point' - this is commonly the reason for not having a grounded scope.

There are four possible options?.

1) Don't use a scope (very cheap, as you don't need a scope, but you probably can't repair the item?)

2) Disconnect the scope ground (very cheap, assuming you already have a scope?), but be fully aware of what you are doing. Under these particular circumstances it's far safer without a ground on the scope!.

3) Power the item under test through a mains isolating transformer, leaving the scope grounded. The item under test is then floating, and is grounded by the scope earth lead. This is usually the preferred method of working, but requires the expense of an isolation transformer.

There are a couple of problems with this method (apart from the extra cost). Firstly, it tends to lull you into a false sense of security, secondly, not all items work correctly in this way - you sometimes get really weird effects? - I've no idea why?.

4) A combination of 2) and 3), using an isolation transformer AND a non-grounded scope - effectively working in a groundless environment - this is my personal prefered method, which I use at work. At home I use method 2).

As I've said repeatedly, UNLESS YOU KNOW EXACTLY WHAT YOU ARE DOING, AND WHY?, you shouldn't be working on live mains sections with a scope, either grounded or not!.

 

[Cifrocco]

Re: Oscilloscope Grounding

While I agree that lifting the ground on instruments is a dangerous practice, I think the discussion has gotten a little off the trail here.

Cifrocco said he blew the power supply's fuse twice by connecting his 'scope's ground lead to what he THOUGHT were ground points in his test object. Perhaps the solution to his problem is to correctly locate the ground of his test object, not to create a dangerous situation by making his 'scope chassis capable of being connected to hot test points without blowing a fuse.

awright

Yes, that's it! I'd like to be able to get waveforms of all points of the circuit from beginning to end so I can understand better how a switching power supply works. I haven't returned to trying it again yet, but I would like to. Not sure if it is best to connect the power supply circuit through an isolation transformer or try to use the X and Y channels of my scope, having never attempted either before.

One of the first things I did when I got my scope was measuring (viewing) my home's AC line. Nothing blew up, I guess I was lucky to place the ground clip on the neutral side of the line.

 

[awright]

Oscilloscope grounding

Yes, Cifrocco, if you were just blindly connecting the two leads from one channel of your 'scope to the two 120 volt power lines, you had exactly a 50% chance of getting fireworks and a 50% chance of seeing a nice, clean waveform. (Those are terrible odds.)

I speak from experience, having shorted the hot line to ground as a green engineer 40 years ago while demonstrating to a newcomer (6 months behind me) how to use a 'scope. That was an indelible lesson. Since the 'scope was grounded (absolutely mandatory in an IBM lab), there were fireworks and a fuse that had to be replaced and some acute embarrassment, but no injuries and no hot 'scope chassis left sitting on the bench, waiting for the next person to come along and grab it.

And, yes, the proper method of measuring the waveform between two possibly ungrounded points in a circuit is to make a differential measurement, which almost any two-channel 'scope is equipped to do. Also, it is a good skill to learn and not difficult.

All you need is a 'scope with two well balanced, identical channels and a pair of identical probes. Identical probes because any difference in probe resistance or capacitance will degrade the CMRR (Common-Mode Rejection Ratio) built into the scope by the manufacturer. "Common-Mode" refers to the signal that is common to both inputs - that is, the signal you want to reject. You want to look at the "Differential-Mode," the signal that is different between the two inputs.

A high CMRR means that if a large signal is applied to both inputs and the channel selector is set to A-B (usually selecting "ADD," and inverting the B channel), there will be a small (ideally no) deflection on the display. That is, B was perfectly subtracted from A inside the 'scope. You have to be careful here because even though there is no or a small deflection on the display, there is still possibly a large signal at the inputs and you risk damage to the 'scope if you crank up the sensitivity and overload the input in an attempt to see a signal.

CMRR decreases with increasing frequency due to the difficulty of matching the capacitances identically throughout the two (A and B) amplifiers. That's why you first establish that the instrumentation is capable of the measurement you want to do by connecting both probes to the same test point and observing that you get adequate common mode rejection with the signal frequency and amplitude you are actually hoping to measure. Note that you still need a good ground connection, say, to the chassis of the test object, even if you are measuring between two hot test points.

While an isolation transformer is a very good safety device for testing objects that may be defective, I disagree with using one to "float" a 'scope so you can connect the 'scope chassis to one of the hot test points. Success of this technique depends a lot on the impedance at the test point because you are hanging a relatively large hunk of metal with significant capacitance to its surroundings on one of the test points, possibly having an affect on the signal you want to measure (ignoring for the moment the safety aspect of having the 'scope body at some potential off ground). Probably OK (but not safe) for a low impedance point in a power circuit. Probably not OK for making a differential measurement in an amplifier or oscillator or timer or digital circuit.

Have fun!

awright

 

[eblc1388]

Re: Oscilloscope Grounding

Not sure if it is best to connect the power supply circuit through an isolation transformer or try to use the X and Y channels of my scope, having never attempted either before.

Sorry Cifrocco this might sound discouraging. Differential does not mean X & Y channels of your scope.

If you are not sure exactly what you are doing, I (probably Nigel and others) would advise you not to mess around with hot AC circuits. There a million other applications that you can probe and learn to use an oscilloscope.

You can always come back to probing hot AC circuits when you are experienced and know exactly what dangers are involved.