Oscilloscope Floating using UPS

[tggzzz]

The only logical answer is to learn from the experts that base their recommendations on solid engineering theory and practice.

In this case, suitable starting points are Tektronix'
"Fundamentals of Floating Measurements and Isolated Input Oscilloscopes"
https://www.tek.com/en/documents/application-note/fundamentals-floating-measurements-and-isolated-input-oscilloscopes
and
"Floating Oscilloscope Measurements … And Operator Protection"
http://www.tek.com/dl/51W_10640_1.pdf
both of which discuss various techniques.

For other starting points, see the "Praxis and Safety" refs at https://entertaininghacks.wordpress.com/library-2/scope-probe-reference-material/

Thank you for sharing these links and information it is greatly appreciated and has contributed to my knowledge base.  I have read both links as well as 90% of the links on "Scope Probe Reference Material".  I will finish the rest within the next day or so.

You are welcome; it is always a pleasure to provide pointers to someone that is prepared to think and learn.

Don't forget that there is no such thing as "ground" or "earth"; they are merely simplifications that have some utility in some circumstances, but are misleading in others.

Don't forget that capacitors, resistors, and inductors don't really exist in a practical or theoretical sense.
Not practical, since every resistor you buy also comes with "free" capacitance and inductance.
Not theoretical, since the the R,L,C (and earth/ground) concepts are simplifications of the way electromagnetic fields interact. Very useful simplifications in many circumstances, but they are of diminishing use at microwave frequencies and when trying to model EMI/EMC effects in terms of "parasitic" R,L,C.

The latter is of particular importance when considering "floating" components, since interactions will occur even when there is no direct connection.

For example, I remember touching the front flap of a Eurocard rack, and getting a slight transient electric shock. The engineer and technician correctly pointed out that the flap was insulated (plastic hinges and catch), and so it couldn't give me a shock (incorrect!). They were so confident of that they wouldn't touch it themselves. Basically, since it was electrically isolated, it was floating at ~120V, halfway between live and neutral.

[JJ_023]

Despite us not understanding one another on this particular subject I would gladly welcome your help or anybody's when it comes down to building a differential probe.  So if that time arises you can expect a PM from me.  Thank you for the offer, hopefully it will still be valid.

Of course, I may even ship you a PCB for the probe I built.

That would be great.  Thank you.  If you have a Gerber file that would probably be easier.  What I would really be interested in is a detailed explanation of how your circuit works and hopefully answers to whatever questions that generates. 

To be honest before I attempted that I would probably do a modification to a current MicSig DP10007 HV Differential Probe that is outlined in the following post.

https://www.eevblog.com/forum/testgear/new-eevblog-hv-probe/150/

[JJ_023]

You are welcome; it is always a pleasure to provide pointers to someone that is prepared to think and learn.

All I can say is I love to learn I do it professionally and as a hobby.   If you are willing to provide pointers then I am willing to listen and ask questions. 

Don't forget that there is no such thing as "ground" or "earth"; they are merely simplifications that have some utility in some circumstances, but are misleading in others.

To me ground is a reference point.  Having said that.  What would be some examples that are very misleading?

Don't forget that capacitors, resistors, and inductors don't really exist in a practical or theoretical sense.
Not practical, since every resistor you buy also comes with "free" capacitance and inductance.

This I am aware of.  Resistors capacitors and inductors all have inductance capacitance and resistance each on their own.
 

Not theoretical, since the the R,L,C (and earth/ground) concepts are simplifications of the way electromagnetic fields interact. Very useful simplifications in many circumstances, but they are of diminishing use at microwave frequencies and when trying to model EMI/EMC effects in terms of "parasitic" R,L,C.

I have very little experience and very modest understanding of microwave frequencies.  This is something I would probably venture into further down the road.

Right now my main concern is getting an understanding of analog, especially towards power supply design concepts.  Right now I was reading a bunch of stuff on current sensing using shunts and DAC’s.  And just a better understanding of op amps overall.  When I have a bit of time I will post on that subject asking for clarification.

The latter is of particular importance when considering "floating" components, since interactions will occur even when there is no direct connection.

For example, I remember touching the front flap of a Eurocard rack, and getting a slight transient electric shock. The engineer and technician correctly pointed out that the flap was insulated (plastic hinges and catch), and so it couldn't give me a shock (incorrect!). They were so confident of that they wouldn't touch it themselves. Basically, since it was electrically isolated, it was floating at ~120V, halfway between live and neutral.

I totally understand where you are coming from.  Having done a considerable amount of Heliarc welding which uses high-frequency starts I have seen and experienced unexpected events. 

One of which baffles me to today.  I got shocked by the high-frequency with the ground clamp lying on a concrete floor nowhere near the weldment.  Wearing thick boots and full PPE with no perspiration. 

Still not sure how that happened and it has happened more than once on different machines.

[Smokey]

...
But if one is determined to "float the scope,", a (good) isolation transformer is the best choice.  You also want to make sure there are no LAN, USB, etc. cables connected to your scope.

Ethernet should be fully galvanically isolated with transformers.  USB is float-breaker though, ya.

[johansen]

fairly certain my old TEK 475 oscope survived being floated at +/- 170vdc at 120KHZ.. no problem, with a cut ground pin.

doesn't mean the waveform was accurate, due to the resistance, inductance, and coupling of the 10 foot long oscope leads... but, it didn't blow up anything.

[magic]

Is this a typo ?  if you omit those and both devices are earthed

I meant: if you omit grounding clips but the DUT and the scope are both earthed, then the scope is still grounded to DUT ground through mains earth cables.

I can definitely see the dangers associated with the negative of rectified mains.  What would be the purpose of taking such a measurement with an oscilloscope when you can take one with a DMM?

Observing switching waveforms, diagnosing repetitive shutdown issues, monitoring ripple on primary controller's low voltage power supply.

Usually DMM checks are enough to find what's wrong with a simple SMPS and often a few part swaps do the job without further damage, but I wanted to be particularly pedantic about verifying everything when working on more complex or valuable PSUs.

1st of all that was just a possible application given as a possible scenario.  In that particular situation both the UPS and oscilloscope would be outside the car.   Also since the scope is going to attach to a grounding point on the car I am not sure I understand your point.

My point is that no one is known to have tried, but if the scope is grounded to the car (rather than used as a poor man's differential probe*) then the result hopefully won't be too bad.

It must be said that both the scope and the car have some capacitance to earth so they aren't truly floating. For example, if there is high frequency noise between the UPS output and its chassis, the chassis is effectively earthed capacitively and noise can be pushed out through the output cable. Then coupled to scope ground, which is the car's ground. All circuits in the car are powered by its battery/alternator and referenced to the car's ground and they might be expected to track the noise too, but they may not if they have their own capacitance to earth and only a high impedance connection with the rest of the car. So it's possible that UPS switching noise could be found all of a sudden on some unshielded analog sensor signal, for instance.

* The thing about diff probes is that they have high and equal input impedance on both IN+ and IN- and they are designed to have no noise sources on IN-. A floated scope is obviously not a diff probe - IN- has high capacitance to ground and possibility of PSU noise.

[tggzzz]

fairly certain my old TEK 475 oscope survived being floated at +/- 170vdc at 120KHZ.. no problem, with a cut ground pin.

doesn't mean the waveform was accurate, due to the resistance, inductance, and coupling of the 10 foot long oscope leads... but, it didn't blow up anything.

I'm fairly certain I've walked into the road without looking... no problem.

With your 475, what were the voltages across the transformer windings and "nearby" PSU components. Were they rated for that voltage? Were they subtly damaged if the rating was exceeded?

[tggzzz]

You are welcome; it is always a pleasure to provide pointers to someone that is prepared to think and learn.

All I can say is I love to learn I do it professionally and as a hobby.   If you are willing to provide pointers then I am willing to listen and ask questions. 

Don't forget that there is no such thing as "ground" or "earth"; they are merely simplifications that have some utility in some circumstances, but are misleading in others.

To me ground is a reference point.  Having said that.  What would be some examples that are very misleading?

Two grounds/earths might not be at the same potential.

Consider a transmission line where the speed of light means that one end could change potential and it will take a noticeable time for that to reach the other end. N.B. a metre is 5ns, i.e. >> than modern logic transition times

Consider two pieces of industrial equipment, both grounded but to different phases or via 100m long cabling. Any resistance plus current => voltage difference.

Consider the advice as to what to do if caught in a field in a lightning storm: keep your feet together. That's because if there is a nearby lightning strike you don't want the electric currents in the earth to "decide" that your legs are a low resistance.

Consider the "ground" on an aircraft which has been flying through dust/clouds and picked up a static charge relative to the runway.

Of course not all those are relevant to you, but they do serve to make it clear that not all ground/earth points are at the same potential.

I have very little experience and very modest understanding of microwave frequencies.  This is something I would probably venture into further down the road.

Sensible decision

If you do, you will come to realise that the electric power is transmitted through the dielectrics (in electromagnetic fields), not along the conductors (by electrons) Yes Virginia, you aren't in Kansas any more (Oops, mixed cultural references )

I totally understand where you are coming from.  Having done a considerable amount of Heliarc welding which uses high-frequency starts I have seen and experienced unexpected events. 

One of which baffles me to today.  I got shocked by the high-frequency with the ground clamp lying on a concrete floor nowhere near the weldment.  Wearing thick boots and full PPE with no perspiration. 

Still not sure how that happened and it has happened more than once on different machines.

Realising what you don't know is the key starting point for improvement The older you get and the more experienced you become, the more you realise that you don't know very much.

Those that close their eyes to what they might not know can be dangerous.

Have fun, safely.

[MrAl]

I always see people get super paranoid about these kinds of measurements yet they are done every day without incident.

Years ago I used the method of cutting the ground pin off of the scopes line cord plug for literally 100's of measurements.  Many people I knew did the same thing, over and over and over and over and over again.  None of them died.  None of them even got shocked.  These were all CRT type scopes.

The key point here is you just have to pay attention to what you are doing when ...

Exactly the same can be said about walking into the road without looking.

(And why did you bring up the display technology (CRT)? That is completely and utterly irrelevant)

Hello,

You have no idea what you are talking about.
How can I be so sure?
I already mentioned this but you, as usual, fail to understand.
I have worked with a LOT of tech's who ALL used scopes with the ground pin cut off, and NOT A SINGLE FRIGGIN ONE OF THEM EVEN GOT A SHOCK.  Let me repeat that so MAYBE you will read it and understand this time.

NOT A SINGLE FRIGGIN ONE OF THEM EVEN GOT A SHOCK.
NOT A SINGLE FRIGGIN ONE OF THEM EVEN GOT A SHOCK.
NOT A SINGLE FRIGGIN ONE OF THEM EVEN GOT A SHOCK.
NOT A SINGLE FRIGGIN ONE OF THEM EVEN GOT A SHOCK.
NOT A SINGLE FRIGGIN ONE OF THEM EVEN GOT A SHOCK.
Not ever, over years and years of working with them in several different companies.  That's three, four, five companies.  Plus, we're talking about a LOT of technicians, and who were not reprimanded in any way by the upper management.  In fact, they were instructed to do it this way.
How can this be?  It be, period.

I am not trying to force anyone to do this, just stating some facts about this.  If a given reader does not like this idea, that's ok with me.  If you feel it is dangerous, then don't do it.  Don't like it?  Then don't do it.  But don't spread misinformation.

Also, mentioning a CRT scope may be irrelevant to you, but not to me.  You're welcome to talk about any kind of scope you like however, feel free.

Now if you want to go ahead and keep bitching about this, go right ahead, and I'll start to ignore what you have to say also.

[magic]

What I find most amusing is that the same guy talks about unpowered fixed-wing flying coffins.
Now, this is a dangerous activity which actually kills people in the real world

[xavier60]

This might not be useful to know. Some mains powered equipment that have a SMPS, will run on DC.
My HP 54645a DSO will run on 85VDC at 0.9A.
There is a chance that this could damage some equipment.

[Ian.M]

OSHA standard 1910.334 - Use of equipment, section (3) Grounding-type equipment specifically prohibits cutting the ground pin:

Attachment plugs and receptacles may not be connected or altered in a manner which would prevent proper continuity of the equipment grounding conductor at the point where plugs are attached to receptacles. ...

and later in the section, explicitly prohibits the use of adapters that interrupt the ground conductor.

Assuming this was in the USA, MrAl and his 'tech' friends exposed the company to potentially massive liability, and should all have received a formal written warning for the first offence of interrupting a grounding conductor, escalating to retraining and even dismissal for repeated offenses.

Even ignoring the fact that OSHA regulations are written in blood, i.e. are codified in response to serious or fatal industrial accidents, or serious negligent harm to employee health, to attempt to protect all employees in similar industries, so should be obeyed even if (and especially if) you do not understand the rationale behind them, the fines OSHA  would have imposed if even one disgruntled (ex) employee had reported cutting ground pins, + the disruption from the resulting surprise inspection to confirm the report, would almost certainly have exceeded the cost of providing isolated differential probes to all techs that actually needed to make floating scope measurements.

Cutting ground pins, like drunk driving, may once have been socially acceptable even though it was and is prohibited/illegal.  It is no longer acceptable, and any colleague or supervisor that proposes or performs it is an active danger to your personal safety.  If management wont listen and act, gather evidence and report it to OSHA or your equivalent national workplace safety authority.

[shapirus]

NOT A SINGLE FRIGGIN ONE OF THEM EVEN GOT A SHOCK.

I remember somebody mentioning (in a similar thread a year or so ago) a case of somebody getting a lethal shock from an ungrounded scope. I don't remember exact details, but I think it happened when that person was alone, so apparently he touched something grounded and the potential present on the scope's housing or controls at the same time, since that's the only way you can get a shock in this context, and nobody was around to save him.

Yes, of course using the scope without grounding is dangerous. But the dangers will only stay a possibility and never cause actual harm, unless (or should I say "until"?) the operator, and those who stay close to the device, fail to observe a particular safety precaution. Similarly, the hand grenade with the pin removed is dangerous, but will not cause harm as long as you keep the lever pressed.

When it comes to employment, insurance, and employer's liability (wherever the latter two are a thing), it's easier to mandate that all the scopes in the company must be grounded at all times than to make sure no human errors are made (which is impossible anyway), and cheaper to simply buy differential probes for those areas of work where they are required.

The difference between the "strong no" and "why, possible, as long as you know what you're doing" positions, I guess, is the difference between the people who received education or training on this subject in countries where earthing in the electric networks has been widespread for a very long time, and the rest.

There are scopes which don't have the earth pin in their mains plugs at all. They work. I have one of them (an old soviet-made CRT scope).
However, I'm not positive that it doesn't have a dedicated wire and clip to be connected to mains earth *in addition* to the main plug, that has to be checked. I imagine that earthing would have been present at the work places at the factories for which those scopes were made in the first place, probably by way of earthing every metal part of the building's constructions and work places furniture etc., so maybe the scopes were supposed to be connected to that. I will check that some time later, as the scope isn't with me right now.

[pdenisowski]

Many of the major scope manufacturers have specific technical briefs advising that instruments not designed for floating operation should never be floated, so why are we second guessing them here in a beginners subforum?

I can tell you that all of our (R&S) internal training and documentation about floating and isolated measurements very strongly discourage floating the scope. 

Like, bold, capital letters, in red with lots of exclamation points

Sure, it can be done.  And I may or may not have done it myself on one or more occasions.  But an isolation transformer and/or differential probe is cheap insurance and doesn't make safety (as) dependent on operator skill.

[pdenisowski]

I’ve actually used a fluke handheld.  They do perform well.  I guess a better question would be how are they accomplishing this if they are truly different from a mains powered oscilloscope running essentially on a battery?

I know some of the better scopes do not have common grounds but that is usually not on a common case occurrence from what I understand.  Is this a correct assumption?

With the exception of truly handheld scopes (like the R&S RTH or similar models), the grounds of all scope channels are tied to each other and to instrument ground.  This is true for all of our benchtop scopes.

I'm not aware of a modern, non-handheld / battery powered scope that isn't constructed this way (all grounds connected together).

[pdenisowski]

The only logical answer is to learn from the experts that base their recommendations on solid engineering theory and practice.

In this case, suitable starting points are Tektronix'
"Fundamentals of Floating Measurements and Isolated Input Oscilloscopes"
https://www.tek.com/en/documents/application-note/fundamentals-floating-measurements-and-isolated-input-oscilloscopes

Although they are a competitor , I have to give a shout-out to Tek for also stating that floating the scope is "an unsafe and dangerous practice and should never be done!"

[tggzzz]

I always see people get super paranoid about these kinds of measurements yet they are done every day without incident.

Years ago I used the method of cutting the ground pin off of the scopes line cord plug for literally 100's of measurements.  Many people I knew did the same thing, over and over and over and over and over again.  None of them died.  None of them even got shocked.  These were all CRT type scopes.

The key point here is you just have to pay attention to what you are doing when ...

Exactly the same can be said about walking into the road without looking.

(And why did you bring up the display technology (CRT)? That is completely and utterly irrelevant)

Hello,

You have no idea what you are talking about.
How can I be so sure?
I already mentioned this but you, as usual, fail to understand.
I have worked with a LOT of tech's who ALL used scopes with the ground pin cut off, and NOT A SINGLE FRIGGIN ONE OF THEM EVEN GOT A SHOCK.  Let me repeat that so MAYBE you will read it and understand this time.

NOT A SINGLE FRIGGIN ONE OF THEM EVEN GOT A SHOCK.
NOT A SINGLE FRIGGIN ONE OF THEM EVEN GOT A SHOCK.
NOT A SINGLE FRIGGIN ONE OF THEM EVEN GOT A SHOCK.
NOT A SINGLE FRIGGIN ONE OF THEM EVEN GOT A SHOCK.
NOT A SINGLE FRIGGIN ONE OF THEM EVEN GOT A SHOCK.
Not ever, over years and years of working with them in several different companies.  That's three, four, five companies.  Plus, we're talking about a LOT of technicians, and who were not reprimanded in any way by the upper management.  In fact, they were instructed to do it this way.
How can this be?  It be, period.

I am not trying to force anyone to do this, just stating some facts about this.  If a given reader does not like this idea, that's ok with me.  If you feel it is dangerous, then don't do it.  Don't like it?  Then don't do it.  But don't spread misinformation.

Some people have been killed. Your shouted assertions don't change that.

Here’s an anecdote about someone that "was WELL AWARE of the danger and took a lot of precautions to prevent shock . . . Bottom Line: He died anyway". https://groups.io/g/TekScopes/message/10754

Also, mentioning a CRT scope may be irrelevant to you, but not to me.  You're welcome to talk about any kind of scope you like however, feel free.

Why, exactly, is the display type relevant to the points being made? Repeated assertions are not a reason.

Now if you want to go ahead and keep bitching about this, go right ahead, and I'll start to ignore what you have to say also.

From your responses (i.e. not answers) you have already started that

[tggzzz]

I've just received a new-old toy[1], a THS720A handheld mains/battery scope with isolated inputs (Sorry, pdenisowski, not R&S )

At first glance the BNC connector looked surprisingly dirty, but the OP might be interested in the reason it isn't.



As you can see, the "outer shield" is plastic, not metal.

[1] no I don't need it. TEA syndrome.

[JJ_023]

Is this a typo ?  if you omit those and both devices are earthed

I meant: if you omit grounding clips but the DUT and the scope are both earthed, then the scope is still grounded to DUT ground through mains earth cables

That's what I thought was the implication, but I wanted to make sure.

I can definitely see the dangers associated with the negative of rectified mains.  What would be the purpose of taking such a measurement with an oscilloscope when you can take one with a DMM?

Observing switching waveforms, diagnosing repetitive shutdown issues, monitoring ripple on primary controller's low voltage power supply.

Usually DMM checks are enough to find what's wrong with a simple SMPS and often a few part swaps do the job without further damage, but I wanted to be particularly pedantic about verifying everything when working on more complex or valuable PSUs.

This makes a lot of sense thank you for the elaboration.  When you take such measurements what are some of the proportions that you take?  I assume you utilize a differential probe.

1st of all that was just a possible application given as a possible scenario.  In that particular situation both the UPS and oscilloscope would be outside the car.   Also since the scope is going to attach to a grounding point on the car I am not sure I understand your point.

My point is that no one is known to have tried, but if the scope is grounded to the car (rather than used as a poor man's differential probe*) then the result hopefully won't be too bad.

It must be said that both the scope and the car have some capacitance to earth so they aren't truly floating. For example, if there is high frequency noise between the UPS output and its chassis, the chassis is effectively earthed capacitively and noise can be pushed out through the output cable. Then coupled to scope ground, which is the car's ground. All circuits in the car are powered by its battery/alternator and referenced to the car's ground and they might be expected to track the noise too, but they may not if they have their own capacitance to earth and only a high impedance connection with the rest of the car. So it's possible that UPS switching noise could be found all of a sudden on some unshielded analog sensor signal, for instance.

Everything you said makes sense to me.  As I explore this particular topic further my understanding is that a lot of the battery powered scopes are very similar to mains powered scopes (I'm talking about reputable manufacturers).  Very few of them have isolated inputs. 

So if that is the case how is utilizing a mains oscilloscope under battery power from a UPS any different in terms of measurement capability under a particular scenario (let's say testing automobile circuits)?  I understand that there will be extra capacitance and other things to contend with.  Both scopes should have the same disadvantages, or am I missing something?

* The thing about diff probes is that they have high and equal input impedance on both IN+ and IN- and they are designed to have no noise sources on IN-. A floated scope is obviously not a diff probe - IN- has high capacitance to ground and possibility of PSU noise.

This I am aware of.  I've actually done a bit of reading on differential probes.  Thank you for pointing that out though.

[JJ_023]

fairly certain my old TEK 475 oscope survived being floated at +/- 170vdc at 120KHZ.. no problem, with a cut ground pin.

doesn't mean the waveform was accurate, due to the resistance, inductance, and coupling of the 10 foot long oscope leads... but, it didn't blow up anything.

I'm fairly certain I've walked into the road without looking... no problem.

With your 475, what were the voltages across the transformer windings and "nearby" PSU components. Were they rated for that voltage? Were they subtly damaged if the rating was exceeded?

From what I read from some of the suggested reading on floating a scope is you actually in danger the oscilloscope specifically with the transformer.  I too wondered what the thresholds would be for damage.  I guess like anything else it would depend on the design of the oscilloscope.

[JJ_023]

Two grounds/earths might not be at the same potential.

Consider a transmission line where the speed of light means that one end could change potential and it will take a noticeable time for that to reach the other end. N.B. a metre is 5ns, i.e. >> than modern logic transition times

Consider two pieces of industrial equipment, both grounded but to different phases or via 100m long cabling. Any resistance plus current => voltage difference.

Consider the advice as to what to do if caught in a field in a lightning storm: keep your feet together. That's because if there is a nearby lightning strike you don't want the electric currents in the earth to "decide" that your legs are a low resistance.

Consider the "ground" on an aircraft which has been flying through dust/clouds and picked up a static charge relative to the runway.

Of course not all those are relevant to you, but they do serve to make it clear that not all ground/earth points are at the same potential.

I understand all of the examples that you gave.  I am totally aware that 2 grounds can have different potential.  And as I stated before to me a ground is a reference point. So what am I missing or are we on the same page?

Realising what you don't know is the key starting point for improvement The older you get and the more experienced you become, the more you realise that you don't know very much.

Those that close their eyes to what they might not know can be dangerous.

Have fun, safely.

I will take that as a gentleman's way of saying you have no clue how I got shocked by th HF. lol

[JJ_023]

Also, mentioning a CRT scope may be irrelevant to you, but not to me.  You're welcome to talk about any kind of scope you like however, feel free.

From a neutral stance could you please elaborate on why the CRT scope would make a difference?  I am curious to see what the logic behind that is.

Thank you.

[JJ_023]

This might not be useful to know. Some mains powered equipment that have a SMPS, will run on DC.
My HP 54645a DSO will run on 85VDC at 0.9A.
There is a chance that this could damage some equipment.

To somebody more experienced this might make sense.  Could you please elaborate in more complete thoughts so that I could understand this better. 

Thank you.

[JJ_023]

  But an isolation transformer and/or differential probe is cheap insurance and doesn't make safety (as) dependent on operator skill.

In your opinion which would be better?  An isolation transformer on the DUT or Utilizing a differential probe (please don't say both). 

I have read most of the literature that was recommended from Tektronix on the subject matter.  I was wondering what your recommendation is since each has its advantages over the other.

Also in which case would one be preferred over the other?

[JJ_023]

I've just received a new-old toy[1], a THS720A handheld mains/battery scope with isolated inputs (Sorry, pdenisowski, not R&S )

At first glance the BNC connector looked surprisingly dirty, but the OP might be interested in the reason it isn't.



As you can see, the "outer shield" is plastic, not metal.

[1] no I don't need it. TEA syndrome.

I can tell it's plastic.  What do the probes look like if you do not mind?