DSO powered through isolation transformer, is it safe?

[Chriss]

Hi!
I just interesting what you think about to use a bench dso which is powered through an isolation transformer?

I always use an isolation transformer when i dealing with some repair which is connected to the main.

Also my dso is powered through that isolation transformer too every time I use them.

I just thinking what could be go wrong in my setup?
Could I also kill my scope? like if he would be connected to the main and the gnd clip hucked up somewhere wrong and got the magic smoke...

I don't wanna talk about user errors when they/we use the scope in improper way like  measuring main voltage  direct on the main connector etc.

Just how safe is really a measuring device when it is powered through an isolation transformer and an smps or an old brick adapter lets say 220V/12V...

Let's say a dmm could also be powered to that same isolated main output with a smps charger and it should be also much safer then if it would be connected through the same smps adapter but directly to the wall main socket.

Or let's say an olda adapter 220V/9V could be used to power the dmm when the battery is taken out.
That would be also galvanic separated from the main 220V and should not lead to any problem with the negative
terminal of the dmm...

What you think about?

Thank you very much.

[tautech]

Simple rules
Always maintain an earthed mains supply to a DSO.
Use an isolation transformer on the DUT only.
Understand risk.
Use differential probes when/as necessary.

[nctnico]

Use differential probes only because once you connect a grounded oscilloscope (or any other piece of test equipment) to the DUT, the DUT is grounded again but any protection from a GFI is gone. My advice: always keep the DUT grounded, use a differential probe with the right CAT ratings and only hook up probes when the DUT is disconnected from mains.

[Rerouter]

Also the other option, often called the poor mans differential, using math or a hardware function to do channel 1 - channel 2, this way as long as you don't exceed the ratings of the probes you don't need to isolate things,

The other part may sound stupid, but may be worthwhile to you, If you think your going to touch something, have your setup powered through an RCD and have ground straps for your wrists and a non conductive mat to stand on and a non conductive bench to work on, the goal is that if you touch something bad, that it will go through one of those straps first, and not through your heart, it takes about 20-30mA flowing through you to trip the RCD, and if your trusting an RCD with yourself, do not go dollar store bingo, get something decent and test it multiple times for its trip current before you trust yourself to it.

Why the grounding, well if you end up touching active a neutral, an RCD won't trip, it only cares about an inbalance, if you provide a ground, It will always trip, it sounds counter intuitive, but letting the current flow from say your finger to your wrist is 1000x better than letting it flow from one hand to another and not tripping.

[Ice-Tea]

The other part may sound stupid, but may be worthwhile to you, If you think your going to touch something, have your setup powered through an RCD and have ground straps for your wrists and a non conductive mat to stand on and a non conductive bench to work on, the goal is that if you touch something bad, that it will go through one of those straps first, and not through your heart, it takes about 20-30mA flowing through you to trip the RCD, and if your trusting an RCD with yourself, do not go dollar store bingo, get something decent and test it multiple times for its trip current before you trust yourself to it.

Why the grounding, well if you end up touching active a neutral, an RCD won't trip, it only cares about an inbalance, if you provide a ground, It will always trip, it sounds counter intuitive, but letting the current flow from say your finger to your wrist is 1000x better than letting it flow from one hand to another and not tripping.

I'm going to respectfully disagree with some of the above:
- installing "a" RCD/GFI/GFCI" is not good enough. The common ones around here trip at 300mA, which will kill you if it takes the scenic route. Use a 30mA or 10mA (I have one on my bench) but be aware you'll spend a pretty penny for that.
- Don't strap yourself directly to ground. If there's no route to earth, the RDC won't trip that's true but it also won't matter because without current you won't die anyway. This is why ESD straps (must) have a large value resistor inside: enough to bleed away voltages, not enough to allow for large, harmfull currents to run.

[mzzj]

The common ones around here trip at 300mA, which will kill you if it takes the scenic route. Use a 30mA or 10mA (I have one on my bench) but be aware you'll spend a pretty penny for that.

That's weird. 30mA GFCI's are the most common around here, metric ton of them everywhere and they cost pennies*. 300mA GFCI's are for fire protection and not used that often.
*30mA plug-in GFCI's start at about 10 euros, brand name 1-ph rail mount ones maybe 20 euros.

But GFCI protection is "complicated", depending on waveform and whatnot they might trip inadvertently or not at all.

[Ice-Tea]

Plug-ins are not used at all here (and not in any neightbouring countries either afaik, except maybe UK?)). A 300mA unit is ussually in the distribution cabinet, a 30mA is added for additional safety in "wet areas". If I remember correctly, 30mA is still OK for price, 10mA stinks. Still got one, though.

[mac.6]

Strange that 30mA GFCI are not mandatory in Belgium, in France they are for every circuit.

[mzzj]

Plug-ins are not used at all here (and not in any neightbouring countries either afaik, except maybe UK?)).

Typically used on old installations in here when there is no permanently installed one. IE outdoors, industrial setups etc.
Been to some factories that require plug-in GFCI for all portable equipment you bring along.

I believe they exist around there too but maybbe not used commonly.
https://www.techniekwebshop.nl/brennenstuhl-aardlekschakelaar-1290660-4007123641130-bren-aardlek-schakelaar-1-2-polen-meet-230v-nominale-nominale-meetstroom-16-ampere.html
https://www.obelink.nl/exin-stekker-aardlekschakelaar-230v-met-stopcontact.html
https://www.stabilo-fachmarkt.de/fehlerstrom-schutzschalter-30ma/a-915715/
https://karese.de/Fehlerstrom-Schutzschalter--CEE-Personenschutzschalter.html

[Chriss]

Huh!
So much info...  Thanks for your time...

What is "DUT"?
Sorry I don't understand that shortening due to language barrier...

The biggest problem is ( if I understand anything in the right way ) the GND probe from the let's say DSO or the DMM,
because if we connect the GND clip from the scope probe to a live wire that is a direct short and we got the BIGO!

Of course this is the situation when our DSO or DMM probe GND clip is grounded to earth wire of the main source of our
wall socket.

That's ok.

If we measure on a device which is powered through an isolation transformer, the device under investigation has his "own word" of power source and is no more related to the "world" of the power source of the measuring tool.
In our cas a DSO or DMM powered from the wall socket.
So, these two world's are separated and can't be any voltage differences between them because they are completely separated.

Let's say the gnd clip of my dso probe is connected to earth and if I power on my device for investigating through an isolation transformer and touch anything with my gnd clip of my dso, I'm expect nothing will happen.
I'm I right?

Taking to much current from a circuit because of the limitation of the measuring unit like measuring over 100mA on a DMM which is let's say limited to take Imax=50mA that is another story I think.

If I'm correct or if not pls correct me.
There are two ways to protect a measuring instrument from burnout is:

1. separate galvanic the measuring unit power source from the power source of the device under investigation or
the opposite way.

2. stay in the safety range with the measuring quantity for what the measuring unit is build/configured.
Let's say, don't try to measure more current with the DMM that the DMM can handle.

I know this is in nutshell but this are maybe the major and the more fundamental safety rules what is to take care.
I have to say, for the rule No. 0 is does the technician who deals with this kind of stuff should be educated enough to know what he/she is doing.
This is a no question...


Thanks.

[nctnico]

DUT=device under test

The rest of your post is irrellevant. Just buy a differential probe for your oscilloscope. This is much easier and safer compared to thinking about how to make an isolation transformer somewhat safe. Isolation transformers are thing of the past. Just don't go there. If you are really interested in the how and why then search this forum because it is a recurring topic.

[bugi]

... it also won't matter because without current you won't die anyway.

This "it is current that kills" mantra is something I have read often, yet as often it brings to my mind someone stating (I wish I could remember where, but it was someone who actually knows/knew stuff) that "there is no safe current and there is no safe voltage". And he also gave examples for both parts of that claim, like specific waveforms/frequencies applied in certain way that had practically near zero current, but still could kill, and very tiny voltage levels that in a specific way could be just as dangerous. (Note, he did include e.g. situations where skin gets pierced etc., not just touching with a dry finger.)

And he was not trying to make everyone to wear full-body rubber suits or such, but just to emphasize that the simpler mantra was not quite correct.

I do agree that limiting the current is still a good way to reduce dangers, even if it does not eliminate them completely.

[Ice-Tea]

There's this urban legend where one would measure the resistance of your body (hand to hand) with a DMM (9V battery inside) and where you would pierce the skin with the tips, end up in a vein and the salty blood would conduct well enough to do serious damage. I suppose there's always ways but my point was that strapping an arm directly to earth would make it less safe, not more.

[Chriss]

The most tricky part what about I worry most of time is, when I have to make some measurement with my dso on a car and I hook up the ground clip from the dso to the gnd of the car, I mean the minus pole of the battery.

I know, these two power sources are completely isolated from each other but somehow I'm worry what if something is wrong on the earth connection in my main connector and some AC voltage goes through the dc wiring of the car?
I'm worrying because every time I connect the ground clip of my scope on the negative terminal of the battery
I connected the car directly to the earth connection of the main electrical installation of the house.

Maybe I should use a GFCI in my connector which I would use every time I use my scope?

Some time I have to go to the customer and I don't know how is the electric installation made at that place and
even I don't know does they have any extra protection more than simple fuses...
 
Actually, how safe is to connect a bench dso to lets say a car electronic to measure the crankshaft sensor?

I think a differential probe would not make any sense to use in this case, or?

[Ice-Tea]

If its for working on a car, have you considered a portable scope? Like this:

https://www.ebay.de/itm/143267265997

(yes, that is my listing)

Probes are isolated, additional care has been taking to increase safety of your floating measurement (BNC is shrouded, GND (?) clip is shrouded,...)

[BravoV]

Its safe as long you don't make probing mistake, meaning 99.99% right is not enough, as even 0.01% mistake or "once", will toast your expensive equipment, your limb or your life.

Probing low voltages like 3.3, 5, 12 or +/- 15 Volt are totally different game, as the tolerance for probing mistake is very relaxing.

[DDunfield]

Simple rules
Always maintain an earthed mains supply to a DSO.
Use an isolation transformer on the DUT only.

If you are doing this so that you can clip your ground lead to reference some HV signal, then be aware that the word "safe" does not apply just because you isolated the DUT.

As soon as you clip your ground lead to something that is not ground, the DUT will now become grounded, but not at the point where the manufacturer intended it to be grounded, but at the HV point you are referencing.  This will cause all of the "normally ground" points on the DUT such as connectors and often the whole chassis/enclosure to be raised to the offset of the HV signal you are measuring. Note that this includes "normally close to ground" points as well, such as low voltage supplies and circuitry - for most DUTs this means that pretty much everything except the point where you clipped the ground lead will become dangerous.

- If you isolate the DUT, exposed "grounds" on the DUT become hazardous.
- If you isolate the scope, exposed "grounds" on the scope become hazardous.

The wisdom of "always isolate the DUT" comes from the days when scopes were big grounded metal boxes and almost always had a lot more exposed ground than most DUTs.

I'd say it tends to be more 50-50 with a modern scope, and which is less hazardous (notice I didn't say "safer") with depends on the exact case.  In any case, neither is a good idea and the only proper advise to give is "don't do it".

There are much better ways scope HV referenced signals:
  - Properly isolated (rated) battery powered scopemeter
  - Differential probes (must be rated for HV offset you intend to work with)
  - Math (subtract) function to obtain relative HV readings using two channels referencing to ground.

Note that unless you have differential probes, or a scope(meter) specifically with isolated channel grounds, then your scope should follow it's equivalent of the "when probing HV keep one hand in your pocket" rule. For the scope, it is "when clipping the ground side lead to HV keep other channel ground leads in their pocket (accessory bag)" - remove them if possible, don't connect them to anything as they too will be at HV potential.

Dave

[Chriss]

Ok, let's separate things like this:

1. scenario when we wish to measure with a bench DSO a low voltage DUT like let's say an Arduino stuff, or a crankshaft sensor on the car. All this DUT are low voltage devices and powered from battery. ( I won't put here now a situation when the DUT is powered from a bench power supply which could be a tricky stuff if the negative terminal on the bench power supply is grounded.)

2. scenario
when we wish to measure a HV DUT which is powered from mains like lets say an SMPS.

Of course I used also handheld scopes which are battery powered but they are all so small compared to my bench DSO and all that tiny buttons on there are really a pain in the a_s.
I really wondering how to convert my bench DSO in a way so I can use it more safer in circuit testing like in the 1. scenario.

I was thinking to mod my power stage of my DSO in a way so I could hook up a battery an if I'm working on a car I could use a separated battery. That would be the same as I would use a handheld scope.

The second option would be to use a portable GFCI cable where my scope is connected, but that is not a handy device
if some HV is traveling through the ground wire (...for some reason the main fuse is not reacting or there is no GFCI installed in the house) which is connected to the DUT through the ground clip of the scope...
The GFCI will cut the live wire and so shut down my scope but won't protect the DUT also.

I always go back and watch Daves video here:


He talks also about two scenario, the 1. and the 2. scenario, where the DUT is isolated through a battery powering
or some sort of isolated oldstyle power adapter with transformer inside  but for sure the ground of the secondary stage of the power adapter must be really isolated and not connected to ground.

If I use Daves introducing then a car should be also the same DUT.
Actually there is also everything powered through his battery and has no way referenced to the main earth connection.
So, why should that be dangerous to use on a car a bench scope?
I see only one reason:
If some AC HV could travel through the earth wiring.
This could make some damage on the battery powered DUT.

 

[CDaniel]

Beside the operator safety , the DSO ground is tied to mains earth ( the third mains plug pin ) , so any attempt to measure something not insulated from mains with the normal probes will destroy the circuit or the DSO .
Thats why you must use a separation transformer for the DUT . If you use this transformer for the DSO ( thats like cutting the earth wire) , nothing will blow up but you can still electrocute yourself ...

[Chriss]

Based on your writing, if the scope is on AC main powered and the DUT is powered through an isolation transformer nothing can happen to the scope and to the operator too.
This is also my thinking.

When I use my scope to measuring around a car, which is battery powered and floating and not referenced back to earth in no way.
In this scenario I'm and my DUT which is the car in this case are also safe.

Is something still wrong in this?

[DDunfield]

Thats why you must use a separation transformer for the DUT . If you use this transformer for the DSO ( thats like cutting the earth wire) , nothing will blow up but you can still electrocute yourself ...

The idea that isolating the DUT makes it so you can't "electrocute yourself" is a huge fallacy. Please see a few posts back.

To the OP .. I don't see a reason that grounding the chassis of a car should cause trouble. The car is an isolated system and perfectly safe to touch. Unless there is some ground referenced power source applied to it, no current should flow between it and an external ground.

In cases like this, your multimeter is your friend. It will show you if there is any potential difference between your scope ground and what you are thinking of connecting it to. When the DUT is a big hunk of metal (like a car) you may read some induced AC, but measureing across a suitable resistance to ground will let you know if any significant current would actually flow.

Attaching your scope ground to HV (such as in an SMPS) in an attempt to measure something not ground reference is another story however and the only proper advice is "don't do it". Use one of the several much less hazardous techniques described in this thread and elsewhere.

Dave

[DDunfield]

Based on your writing, if the scope is on AC main powered and the DUT is powered through an isolation transformer nothing can happen to the scope and to the operator too.
This is also my thinking.

Picture this - you want to see the waveform across an inductor in a HV power supply of a DUT. This inductor sits +HV above ground on the DUT, which has a grounded metal box, grounded connectors, a few large grounded heatsinks, perhaps some setscrews in adjustment knobs attached to grounded potentiometer shafts etc

To be "safe" your put this whole thing on an isolation transformer, and also to be "safe" you make sure your scope is duitifully grounded.

Then, to measure across the inductor you clip your scope ground lead to one side of it.

You've just externally grounded a point +HV away from what the DUT manufacturer calls ground.  What do you think happens to all those exposed "ground" points on the DUT.

Dave

[Chriss]

Is this what you mean?


In this case I don't think does something can go wrong, cos the secondary side is floating.
The complete device is floating.

What I don't like is, whenever I tap with my alligator clip to any part of the circuit, I direct connected that
circuit to main earth.

[CDaniel]

For big HV potentials and other hazards there are of course DSO with insulated inputs ... for normal working in the primary side of a consumer product  power supply  a separation transformer is good enough .

[DDunfield]

In this case I don't think does something can go wrong, cos the secondary side is floating.
The complete device is floating.

It depends on what you mean by "something can go wrong".

If you mean: Will a significant unexpected current flow through the scope ground lead possibly damaging the DUT and/or the DSO?

Then the answer is "No". You are correct in that the isolation of the transformer will prevent current from the secondary side flowing to ground.


If you mean: Are connectors, controls, heat-sinks, possibly chassis/metal-enclosure, as well as all ground referenced low-voltage circuitry of the DUT that I have become accustomed to "knowing" are safe to touch now sitting at >300v from earth ground, greatly increasing the chances that accidentally touching them with my body or tools will be "unpleasant"?

Then the answer is "Yes" - as soon as you connect the ground lead of a grounded scope, the DUT is no longer floating but in fact grounded at a point where it shouldn't be.

What I don't like is, whenever I tap with my alligator clip to any part of the circuit, I direct connected that
circuit to main earth.

Exactly this. But it's more than just "doesn't seem right" .. with HV, it means that anything circuit connected that's not in the HV path has become hazardous (possibly lethal) relative to mains earth.
Contrary to the popular EE legend, Isolating the DUT does NOT make it "safe".

Dave