Earthed enclosures and offline SMPS PCBs?

[Faringdon]

Hi,
Do you agree, that when you have an offline SMPS in a metal enclosure, then the SMPS PCB must obviously have y caps to the earthed enclosure.
Do you agree, that the connection of earth, to the metal enclosure, must be from the earth wire that comes in with the mains cable, and that that earth wire, must connect to the metal enclosure, as near as possible to the point where the mains cable comes in through the enclosure?

Alternatively, the earth wire from the mains cable, can connect to the earthed enclosure at a point as near as possible to the mains cable connector on the  offline SMPS PCB.
Would you agree?...and which of the above two places is best?

[tggzzz]

Hi,
Do you agree, that when you have an offline SMPS in a metal enclosure, then the SMPS PCB must obviously have y caps to the earthed enclosure.
Do you agree, that the connection of earth, to the metal enclosure, must be from the earth wire that comes in with the mains cable, and that that earth wire, must connect to the metal enclosure, as near as possible to the point where the mains cable comes in through the enclosure?

Alternatively, the earth wire from the mains cable, can connect to the earthed enclosure at a point as near as possible to the mains cable connector on the  offline SMPS PCB.
Would you agree?...and which of the above two places is best?

One interesting comment on that...
https://www.eevblog.com/forum/chat/keeping-hold-of-electronics-engineers/msg5047744/#msg5047744

And the preceding post offers some context.

[floobydust]

What is the product's method of protection against electrocuting people? Is it reinforced or double-insulated, or is the secondary side PE grounded?
You leave out this critical piece of information which also determines spacings you ask about.

The PE connection to the enclosure must be able to trip the branch circuit breaker, compared to a PCB trace or wire fusing- leaving the enclosure hazardous live.
So a 15A current capability and certifiers I have seen inject 15A for an hour and then inspect.
edit: unless the product has fusing on Hot/Line, then I think the current handling for PE gets reduced to clearing that fuse. It depends on the mechanical of the enclosure and power entry.

[Faringdon]

Thanks, yes those are great safety points, and of course, the earthing has the important safety aspect.....in this post, from an EMC viewpoint, how would it best be done?

{I know that all will forgive this question, that absolutely none of your industrial secrets are being revealed here....that the entire industry in whatever country you are from, has already been outsourced to (mostly) China........its already gone...the secrets no more. [RANT bit over]}

[Gyro]

...
{I know that all will forgive this question, that absolutely none of your industrial secrets are being revealed here....that the entire industry in whatever country you are from, has already been outsourced to (mostly) China........its already gone...the secrets no more. [RANT bit over]}

If only!

It never was an 'industrial secret', just basic EMC and safety consideration.

[mtwieg]

I'm not aware of any standard that mandates the earth wire from the cable has to connect directly to the chassis/enclosure for class I equipment. Though if the connection is not direct, safety standards often get picky about how the indirect connection is completed. For example, using dedicated fasteners and wires, colors on the wire and fasteners, labelling, etc. Consult the relevant standards for your region.

[Faringdon]

Thanks, i would agree, and from an EMC viewpoint, the earth from the mains cable should connect to the chassis as close as possible to the mains input connector of the offline SMPS...........rather than connecting to chassis at the point where the mains cable enters the metal chassis.......this is to reduce the loop area for common mode noise that is going back down the mains  cable's earth.

Its an exteremly common fault that companies think that the mains cable  earth should  (from an EMC viewpoint) connect to chassis at the point where the cable enters the metal chassis...though this is often the cheapest/easiest place to connect it.

The purpose of common mode filtration is mainly to stop emissions from getting coupled out to the "surrounding" earth..........if you are going to do that, you have to be able to "Invite" the emissions to firstly just not be an emission, and come in by  line, and then leave by neutral...so thats what the common mode choke does.......stops it becoming common mode....but some will inevitably escape....and so you want these emissions to go down the mains earth cable, rather than going to the "surrounding" earth.....because if they go down the mains earth cable, then they at least encounter the common mode choke and Y cap filtration.......so this is why the earth connection needs to be as near to the offline SMPS mains input connector as possible.

Also, you often go to places and  its told that the offline SMPS is only supplied by line and neutral, so no common mode choke or y cap is needed...since "there cant be any common mode noise, since theres no earth connection"....but as you know, the earth is in the surroundings, and common mode filtration is still just as needed......y cap across the isolation transformer and common mode choke at input.

[mtwieg]

Hard to follow what you're saying without some example schematic with it... but I'll say that IMO when deciding how to connect the protective earth from the AC mains cable/receptacle, the main concern should be reliability of the connection (to the entire chassis, not just the circuit board), not whether it's in the optimal position to minimize loop area or whatever. I'd prioritize mitigating shock hazards over EMC.

[jonpaul]

FTTS NO "WE" DO NOT AGREE

GET A LIFE

j

[Faringdon]

I'd prioritize mitigating shock hazards over EMC.

Thanks, yes , but ayk, it can be done and shock well mitigated.

[floobydust]

I'd prioritize mitigating shock hazards over EMC.

Thanks, yes , but ayk, it can be done and shock well mitigated.

Safety trumps EMC in a design, do you agree?

[Faringdon]

Safety trumps EMC in a design, do you agree?

Thanks, ayk, you can easily have both.
The safety one is straightforward to get..the EMC one not so straightforward.

The fact is, that when you have a SMPS PCB in a metal enclosure, then that SMPS will often not work, or be compromised in its operation, due to the metal enclosure acting as an antenna and inducing noise into the SMPS PCB. As such, it is essential to connect the ground (or other "quiet node") of that PCB to the metal enclosure...this can be done  capacitively or directly. Often an entire plane of the PCB is dedicated to chassis ground...and that chassis ground copper pour is connected to the metal enclosure via Y caps.

This can be seen in vicorpower module app notes. Pg 88 onwards shows the chassis ground copper pour in the SMPS PCB..its coloured yellow.....and entire plane dedicated to it....right under the power module....omit it, and noise may well kill the operation of the module.

https://www.vicorpower.com/documents/design_guides/DG-DCM-Design-Guide-VICOR.pdf

[mtwieg]

The fact is, that when you have a SMPS PCB in a metal enclosure, then that SMPS will often not work, or be compromised in its operation, due to the metal enclosure acting as an antenna and inducing noise into the SMPS PCB. As such, it is essential to connect the ground (or other "quiet node") of that PCB to the metal enclosure...

That's not why the conductive enclosure/chassis is normally connected to earth. When a device has a conductive enclosure/chassis, we connect it to earth for the purpose of preventing shock hazards.

The idea that a SMPS enclosure could act as an antenna, and could disrupt the SMPS, doesn't make any sense.

[Faringdon]

The idea that a SMPS enclosure could act as an antenna, and could disrupt the SMPS, doesn't make any sense.

I worked on a 300W  push pull once....and it would not work unless its Y capacitors were connected to the metal chassis.....you connected them, it worked...you disconnected them...it stopped working.
It also worked without the y caps, if you directly soldered the chassis screw restrings to the cct ground.

The SMPS had been designed by an ex-engineer...we were told to get  a few of them  working.....it was after going  home time, and we were all on unpayed overtime...then i noticed this guy sneakily  "accidentally" soldering the cct ground to the chassis screw restrings....flicking the solder across the solder resist .....he knew what he was doing....getting it working any old how, as quickly as poss so he could go home.

He did indeed go home, and muggins here undid it and got it working by connecting in the y caps instead....plus a few other tweaks that needed doing.

Ayk, a lot of electronics doeesnt make immediate sense....take common mode noise...its due to the emissions coupling to external earth or to the metal enclosure...but then you solve it by coupling circuit ground to the metal enclosure with a y cap...so youre solving the problem of coupling to the enclosure...by coupling to the enclosure(!)