SMPS low voltage side earthed

Thread Starter

munuainen

Joined Jun 13, 2014
11
I'm running a project on an old 310W LiteOn computer PSU that I'm planning to convert into a bench supply. There's a plenty of material online about doing this, and I've pretty much done everything I want atm, run initial tests etc.

There's still one question about the way the supply's been built. I'm sure there are people on this forum with background on commercial SMPS design who can reliably answer this: Low voltage side ground is directly connected to mains earth. Will there be a safety or functional issue in isolating these two from each other?
 

MikeML

Joined Oct 2, 2009
5,444
...
There's still one question about the way the supply's been built. I'm sure there are people on this forum with background on commercial SMPS design who can reliably answer this: Low voltage side ground is directly connected to mains earth. Will there be a safety or functional issue in isolating these two from each other?
If there was, every desktop computer made would have a problem...

I would not like for my bench supply to have its 0V terminal connected to earth ground...
 

Thread Starter

munuainen

Joined Jun 13, 2014
11
If there was, every desktop computer made would have a problem...
I wonder the purpose the manufacturer has originally designed such connection for.

I would not like for my bench supply to have its 0V terminal connected to earth ground...
Neither would I as I think about it. One careless move with +V banana plug -> contact with the PSU chassis -> short circuit. PSU must have its overcurrent protection switch the supply off in such case but it sucks nevertheless.

Also in my living room where the workbench is, there are no earthed mains outlets. So, through Y-type line bypass caps I've got half the AC line voltage via PSU chassis in both my LV ground and V+. Not that there are many earthed discharge paths in the same room, just a radiator and a cable TV connection that I can think of, but still...if there's absolutely no harm in separating them, why not do that.
 

MaxHeadRoom

Joined Jul 18, 2013
28,702
If the P.S. was used to supply auxiliary circuits for a permanent project, or connected to a PC port etc, I would say it it is OK, and often normal for the common to be connected to earth ground.
But in the case of a Bench power supply it is preferable to keep it isolated from Earth ground, then you have the option to do this based on a particular project.
Max.
 

crutschow

Joined Mar 14, 2008
34,470
Typically bench supplies have their outputs floating with a third ground terminal connected to the earth safety ground from the AC outlet. That way you can ground either terminal or not as desired.
 

Thread Starter

munuainen

Joined Jun 13, 2014
11
Why would one then want to earth the output? I take it must be a safety issue, such as offering protection from accidental connection with live line voltage or so? I guess it's just a bit hard for me to imagine as I'm likely ever going to use the supply to feed floating and otherwise powerless circuits.
 

Attachments

crutschow

Joined Mar 14, 2008
34,470
Why would one then want to earth the output? I take it must be a safety issue, such as offering protection from accidental connection with live line voltage or so? I guess it's just a bit hard for me to imagine as I'm likely ever going to use the supply to feed floating and otherwise powerless circuits.
Yes, in most of my circuit testing the ground terminal on the power supply was unused. Only in rare cases have I found that it helps to reduce circuit noise in sensitive circuits, by grounding one of the power pins.
 

MaxHeadRoom

Joined Jul 18, 2013
28,702
I'm running a project on an old 310W LiteOn computer PSU
Why would one then want to earth the output?
The P.C. P.S. common and Hence the M.B. ground plane is earth grounded in desk top/tower P.C.'s.
Which also means that peripheral equipment that is fed or directly referenced/coupled to the P.C., either by USB, P.P. or other means is also referenced to earth ground, this assists in reducing noise and other spurious signals that may be statically or other wise introduced from damaging the P.C.
Max.
 
Last edited:
Top