1. PF Insights is off to a great start! Fresh and interesting articles on all things science and math. Here: PF Insights

SMPS grounding

  1. Assuming a 2 pin connector for an SMPS power supply, what is the role of the R and C that are connected to Case ground. Usually R is in Megaohms and C around 0.1uF,2KV.
    IS there a risk of shock when the case is touched?
     
  2. jcsd
  3. berkeman

    Staff: Mentor

    If the product has a 2-prong power plug (in the US at least), it must meet "double-insulation" requirements for isolating the AC Mains voltages from any SELV (safety extra-low voltage) areas, like things that people can touch. When you meet the double-insulation requirements, then you don't have to Earth ground your metal enclosure, for example.

    But to help minimize unwanted EM Interference radiated by your product, you may need to make a "ground" connection from your circuitry to your metal enclosure. This can be done with a capacitor (for low impedance at RF frequencies), and you will often parallel up a bleed resistor with the RF cap. The bleed resistor helps to keep static charge from building up between the metal case and the circuitry inside of it.
     
  4. The Power supply in question is a off line SMPS with 24VDC output. If I dont connect the ground pin, the Voltage between Neutral and Ground wall socket is about 100V. I am running it on 220VAC.
    This seems too high.
     
  5. berkeman

    Staff: Mentor

    What country are you in? Can you post a picture of this power supply and it's safety agency approval label?

    You said in your OP that the power supply has a 2-prong power plug, but in this post you talk about a "ground pin". What ground pin? On the 2-prong plug, or in the wall socket?

    And you are saying that the wall outlet has that large voltage between the Neutral pin and Ground pin? In the US that would be a fault, and quite dangerous. Do you have access to a wall socket electrical tester like the one shown here?

    http://www.liberty-hi-railers.com/images/outlet_tester.jpg
    [​IMG]
     
  6. Country is India.
    The Power supply is Cosel KHNA120F-24
    It has all the typical compliance certificates - CE, UL, TuV

    I was just curious how the grounding was done in a 2 pin power supply vs a 3 pin power supply.
    The Ground pin I am referring to here, is the wall socket ground pin.

    The wall socket voltages look normal. Ground to Neutral about 2VAC, Line to Neutral about 219VAC.
    For the Power supply in question, if I use a 2 pin plug(instead of 3 pin) and measure the voltage between the Wall Ground and Power supply neutral, I get about 100VAC.
    Is this normal?
    The Neutral and Ground should be shorted at the utility box. right?
     
  7. berkeman

    Staff: Mentor

    So with nothing plugged into the wall socket, you get about 2Vrms from Neutral to Ground (sounds reasonable). But when you plug your 2-prong power supply cord into the outlet, that voltage chages to 100Vrms?
     
  8. Yes about 100Vrms.
    I just found out the reason - it's the Y capacitor at the input of the SMPS.
    The Y caps are acting as voltage divider, which is why I get about half of the input voltage.
    Aren't the Y Caps a shock hazard?
     
  9. berkeman

    Staff: Mentor

    That makes no sense; I must be misunderstanding you. If Neutral and Ground are tied at the breaker panel, how can plugging in a 2-prong plug into Hot and Neutral impress a large 100Vrms AC voltage between Neutral and Ground in the wall socket?

    This also makes no sense. You don't have a Y capacitor in a power supply input that has only a 2-prong plug. What would the bottom of the Y cap connect to?
     
  10. The Power supply has 3 prong plug. In this case, the Earth pin is not connected (purposfully).
    The Y caps are connected to the Power supply case.
     
  11. berkeman

    Staff: Mentor

    That is dangerous. Do not "cheat" 3-prong equipment by disconnecting the ground pin. That can create a serious shock hazard. There are better ways to get isolation from Hot and Neutral -- you can use an Isolation Transformer, for example.
     
  12. I had to do it purposefully just to double check.
    The first time it happened was when a tech ignored the Earth terminal and used a 2 pin prong.

    In this case, how would Isolation transformer help. It has Earth ground connection too.
     
  13. berkeman

    Staff: Mentor

    Why do you want to purposely disconnect safety ground from your SMPS?

    An isolation transformer gives you an isolated Hot & Neutral that are not galvanically connected to Earth in any way. They are used in testing TVs for example, which generally have 2-prong plugs and often have a "Hot Chassis". When you power the TV with an isolation transformer, you can connect your oscilloscope ground to the chassis and make your measurements normally. If you did that without the Iso Transformer, you would blow up your 'scope probe and maybe part of your 'scope.
     
  14. I tried connecting to a Variable transformer, thinking it should be similar to isolation transformer. But I still got 100Vrms between PS Neutral and transformer Earth ground.
     
  15. berkeman

    Staff: Mentor

    A variac is not isolated.

    Why are you wanting to cheat the power cord ground?
     
  16. jim hardy

    jim hardy 4,723
    Science Advisor
    Gold Member
    2014 Award

    If this is the power supply
    http://www.cosel.co.jp/en/products/pdf/SFE_KH.pdf

    i see two terminals for AC in and no earth (chassis ) terminal
    one terminal labelled "PE" ( : maybe Power Enable? Spec sheet says elsewhere it has remote on-off.... no hint how to use it though)

    and a case described only as oops wrong line [STRIKE] FR-4 which wikipedia claims is fiberglass.[/STRIKE] PBT which Dupont claims is great stuff
    So i suspect it's "double insulated" as per Berkeman's first post with no provision to earth its chassis
    and those capacitors are the reason for the 10 ma current from input to output on AC isolation test. As well as your 100 volts to earth. A lower impedance meter should give lower readings, have you a cheap analog one?

    You have the actual specimen - could this be right ?

    old jim
     
    Last edited: Jun 24, 2014
  17. berkeman

    Staff: Mentor

    PE is Protective Earth = GND.
     
  18. jim hardy

    jim hardy 4,723
    Science Advisor
    Gold Member
    2014 Award


    Aha ! That makes sense now .
    The block diagram shows PE tied into the input noise filter
    which doubtless includes the junction of the across-the-line capacitive voltage divider
    thereby providing a short circuit for RF interference normal mode
    and a return to earth for common mode, keeping both out of the power lines

    and an "earth" for whatever metal needs to be connected there.

    The high megohm resistor is there to prevent static voltage buildup. You'll find one also in those "static wrist guards" sold for handling electronic parts on assembly lines. It is low enough to drain static yet high enough to prevent current that you could feel, the ones i'm familiar with used a ten meg...

    0.1uf at 60 hz 120volts should pass no more than ~4 milliamps
    which you can definitely feel but it won't hurt a lot
    and i doubt you'd notice unless your fingers were moist
    might they be even smaller, 0.01 uf?
    Shock risk would be signiificant only if something else were wrong.


    Thanks guys !
     
  19. dlgoff

    dlgoff 3,079
    Science Advisor
    Gold Member
    2014 Award

    Interesting. Never ran across these Protective Earth or Noiseless Earth symbols.

    [​IMG]
     
  20. berkeman

    Staff: Mentor

    Interesting. I hadn't heard of Noisless Earth before. :smile:

    I ran into PE on some European customers' devices that I helped debug a while back (EMC issues).
     
  21. jim hardy

    jim hardy 4,723
    Science Advisor
    Gold Member
    2014 Award

    PE is new to me too, and pleasantly so.
    I do like and encourage use of "earth" for earth ground and "common" for powr supply return as opposed to "ground" . You've both heard me beat that drum around here plenty.
    I'm glad to see "Protective Earth" used because it leads the mind directly to what is the purpose of the connection.

    IMHO, stateside EE curricula need a one hour course on the IEEE Green Book . It is embarrassing how many graduate EE's don't understand the basics.

    old jim
     
Know someone interested in this topic? Share a link to this question via email, Google+, Twitter, or Facebook

Have something to add?

0
Draft saved Draft deleted