is it safe to use an isolation transformer that has a floating output for a PC?

[justflow111]

Is it safe to use an isolation transfomer that has a floating output and ground, for a desktop pc + monitor? Does it mean if there is a fault there is no grounding and I can get shocked? Or if I touch another grounded device? I can't find a concrete answer thank you to everyone!!

[BillyO]

Why do you want to do this?  Do you mean the ground is passed through?

Laptops are floating devices, but they are low voltage and designed to be that way.

I'd say that generally it would not be a great idea, especially if it is an older PC that has a grounded metal cabinet.

[studiot]

Is it safe to use an isolation transfomer that has a floating output and ground, for a desktop pc + monitor? Does it mean if there is a fault there is no grounding and I can get shocked? Or if I touch another grounded device? I can't find a concrete answer thank you to everyone!!

I see no problem with this provided you ground any external metalwork  -  I note you have a floating output and a ground connection.

I actually ran a Compucolor II that way for years.

[BillyO]

So, the ground is not connected.

Edit:  I just looked at the manual.  It has a schematic that shows the ground is connected through.  If you do not modify it, then you will be fine.  The ground will work as it should and your set-up will be safe.


Again, why do you want to do this?

[tooki]

It's a modern desktop gaming pc powered by classic switch mode PSU.

Why would you want to run this from an isolation transformer? The power supply itself already provides galvanic isolation.

[golden_labels]

Welcome to the forum.

This sounds like a XY problem to me. Explain, what is your original goal.


In virtually all PCs we may encounter nowadays, both the case and the signal GND are coupled to power lines through Y-capacitors. These capacitors form a voltage divider and the case voltage is halfway between the lines. This is why PC cords are required to be plugged into earthed sockets. If they’re not, we get up to 115 V_AC to earth on the case.⁽¹⁾

The Y-capacitors have impedance too high in 50–60 Hz range to pass current high enough to hurt a person. But the current is sufficient to be felt or in some cases even painful, and it may lead to secondary incidents with more serious consequences. It may also affect electronic devices connected between the case and earth.⁽²⁾

With an isolation transformer and no earth connected, the case is floating relative to earth and so are both power lines. But not floating relative to each other, and now we have three voltages to deal with. As opposed to one with plain, old isolation transformer scenario. Furthermore, they only remain floating if neither of them is connected to any earth-referenced device.

⁽¹⁾ 60 V_AC for North America.
⁽²⁾ With this additional voltage fed into them. That’s separate from the effects of  missing proper, earthed shielding.

[Sorama]

Maybe it’s because of a slightly different frequency when running on ups in battery
mode.

[golden_labels]

I'm running some sensitive programs on the pc, and I'm experiencing some delay on it that is really in micro seconds but crucial and devastating for my task. The delay disappears when I run the pc connected to a UPS that is running on battery mode and unplugged from the wall (battery mode), and also when I run the pc in a different location.

And what is your reasoning process that explains the difference?

Also, and that is a separate question, how did you manage to measure microsecond differences for the task? Did you use Wald’s or other CI calculation method? What’s the null hypothesis you used? If the latter two sentences leave you confused, I am 99% sure you’re just “seeing things” and over-interpreting random variation.

[negativ3]

The only question I have is why?

This sounds vaguely similar to a PC having "interference induced lag" issues some time ago.

[golden_labels]

Hostile behavior will not earn you anything. If you reject help, there is little that can be given to you.

If you understand basic electricity: I gave a complete overview above.

If you do not understand basic electricity: no, it is not safe.

[justflow111]

Gotchu, thank you for the overview. Bless <3

[studiot]

I am totally mystified by this thread.
I am seeing quotes of replies that I can't see.

[justflow111]

hey studiot, what do you mean by this "provided you ground any external metalwork"?

[negativ3]

I am totally mystified by this thread.
I am seeing quotes of replies that I can't see.

Schrödinger's quotes?

[ledtester]

I am totally mystified by this thread.
I am seeing quotes of replies that I can't see.

Could be related to the server problems going on right now.

As I understand it, there are at least two machines serving the site and it is possible that their databases are not in sync. For instance, if you get directed to one of them you'll get the "Bad Gateway" error page, but then you refresh and get served by the other one and the page comes up normally.

[MrAl]

Is it safe to use an isolation transfomer that has a floating output and ground, for a desktop pc + monitor? Does it mean if there is a fault there is no grounding and I can get shocked? Or if I touch another grounded device? I can't find a concrete answer thank you to everyone!!

I've never used an isolation transformer for a desktop PC, but I would think it would work.
I have used a regular AC transformer to isolate the line from my measuring equipment and that works fine too.

Most small transformers used for commercial sale have good isolation, but you can test it first.  There are strict rules about isolation in small transformers like we use to charge the phone and other stuff.  It's always good to test first though as there has been unforeseen leakage causing actual deaths.  For example, don't trust one while taking a bath when you need isolation from ground.

[studiot]

I am totally mystified by this thread.
I am seeing quotes of replies that I can't see.

Could be related to the server problems going on right now.

As I understand it, there are at least two machines serving the site and it is possible that their databases are not in sync. For instance, if you get directed to one of them you'll get the "Bad Gateway" error page, but then you refresh and get served by the other one and the page comes up normally.

I've definitely been experiencing that problem over the past couple of days.

Thanks for the heads up.

[MariuszD]

I'm running some sensitive programs on the pc, and I'm experiencing some delay on it that is really in micro seconds but crucial and devastating for my task. The delay disappears when I run the pc connected to a UPS that is running on battery mode and unplugged from the wall (battery mode), and also when I run the pc in a different location.

The power source for your PC does not affect computational performance or interrupt latency. If you are observing such effects, something is broken and needs to be fixed.

[studiot]

hey studiot, what do you mean by this "provided you ground any external metalwork"?

I am in England where the mains is different from America.

My isolation transformer I referred to was to supply American voltages to an imported American computer, sourced from British mains.
The stepdown transformer had isolated primary and secondary windings in a metal case and the British mains ground (earth) was connected to the case and available as you described for your setup.
So I could ground external metalwork on the Compucolor to this, whilst maintaining the supply isolation integrity.

Does this help ?

[SteveThackery]

I'm running some sensitive programs on the pc, and I'm experiencing some delay on it that is really in micro seconds but crucial and devastating for my task. The delay disappears when I run the pc connected to a UPS that is running on battery mode and unplugged from the wall (battery mode), and also when I run the pc in a different location.

Hey, hang on, this is starting to sound like that nutter some months ago in Ukraine where he said he could see visual glitches when game playing due to some entirely magical forces in the environment. Do you remember that? He was on about isolating the computers from the mains supply, IIRC.

@justflow111: if you want consistent sub-microsecond latency then a PC is hardly the right tool for the job. Please tell us exactly:

1/ What are these "sensitive programs", and what do they do?

2/ Exactly how do you know that these microsecond delays are occurring?

3/ Exactly what effect do these delays have?

Please explain in detail the processing chain from the source of the signal to the output, and explain where in the chain these delays are occurring and what problem they cause.

[golden_labels]

Hey, hang on, this is starting to sound like that nutter some months ago in Ukraine where he said he could see visual glitches when game playing due to some entirely magical forces in the environment. Do you remember that? He was on about isolating the computers from the mains supply, IIRC.

I don’t recall that one myself (perhaps better for my brain!). But this person here has already made their own reputation. Not because of what they asked about, but because of the interaction style.

The question itself doesn’t have to be a nutjob case. It might’ve been honest, just coming from a greatly misguided person. Not much is needed. We all tend to see patterns and connections where there is none. Without sufficient knowledge to act as a harness, there is nothing to stop one from falling into the abyss of absurd theories. At this point all that is needed is overconfidence and we have a comedy like this.

I suspect this may be the case.⁽¹⁾ Suspicion is not enough, so I still gave the case some chance and asked questions that could rectify it. Either by showing that indeed there is something I am missing, or guiding OP and showing where they made the mistake. Justflow111 decided to reject this option, and by the manner of interaction define their image on the forum in the worst possible way. I believe this thread should just pass into the history.

⁽¹⁾ XY problem, no clear basis in knowledge to link the solution to the original problem, claiming an observation that generally isn’t easy to make.

[SteveThackery]

The question itself doesn’t have to be a nutjob case. It might’ve been honest, just coming from a greatly misguided person.

I refer you to my three questions in Reply #19. Until we've got solid answers to each of them, I remain deeply sceptical. Read this again:

I'm running some sensitive programs on the pc, and I'm experiencing some delay on it that is really in micro seconds but crucial and devastating for my task. The delay disappears when I run the pc connected to a UPS that is running on battery mode and unplugged from the wall (battery mode), and also when I run the pc in a different location.

"Some sensitive programs"; "some delay on it that is really in micro seconds but crucial and devastating for my task". Doesn't that sound like vague, hand-wavy bullshit to you?

This is where the alarm bells ring: " The delay disappears when I run the pc connected to a UPS that is running on battery mode and unplugged from the wall (battery mode), and also when I run the pc in a different location." Note the start of the magical thinking.

Finally, observe how the OP has removed several of his posts, including the one I've quoted.

Naaah, this doesn't smell real to me.

[golden_labels]

I refer you to my three questions in Reply #19. Until we've got solid answers to each of them, I remain deeply sceptical. Read this again:

Similar to mine in reply #7: “how did you manage to measure microsecond differences for the task? Did you use Wald’s or other CI calculation method? What’s the null hypothesis you used?”

"Some sensitive programs"; "some delay on it that is really in micro seconds but crucial and devastating for my task". Doesn't that sound like vague, hand-wavy bullshit to you?

It does sound questionable, which is exactly why I asked them about the details. And why I raised my doubts: “If the latter two sentences leave you confused, I am 99% sure you’re just ‘seeing things’ and over-interpreting random variation.”

This is where the alarm bells ring: " The delay disappears when I run the pc connected to a UPS that is running on battery mode and unplugged from the wall (battery mode), and also when I run the pc in a different location." Note the start of the magical thinking.

I did, which is why I wanted to know how they came to such conclusions: “And what is your reasoning process that explains the difference?”

[SteveThackery]

Ah, found it:

https://www.eevblog.com/forum/beginners/unstable-devices-and-visual-smoothness-problems-in-digital-equipment/

It's a heck of a long thread, but to me, at least, it has the same vibes as this one.