Hi Ken:


>   I wanted to pose the group a question in regards to spacing requirements for
>   the U.S.  Many UL standards allow for the exception of components (such as
>   semi conductors, switches, etc) to not meet spacing requirements (as they
>   usually have different standards they meet anyways)   The question is what
>   about PCB trace spacing for those components, such as a capacitor in a
>   switching power supply to ground.  High voltage, the components is an
>   exception, but what about the pads under the component, are they required to
>   meet the spacing requirements, or would that fall under the component
>   exception?  

You've brought our attention to the anomaly that a
component must meet its spacing requirements, and
that the end-product (i.e. PWB) must meet its spacing
requirements, even when the component requirements
are less than that of the end-product.

>From an engineering point of view, this is nonsense.

Nevertheless, our standards have such requirements,
and our certification houses must enforce such 
requirements.

There are several options:

1)  Design the PWB to meet the spacing requirements
    by trimming edges off circular pads, or by bending
    the component leads.

2)  Test by short-circuiting the spacing and observing
    the results (i.e., no shock or fire or damage to 
    basic insulation).  This is especially appropriate 
    for semiconductors since, by definition, they 
    alternate or vary between open and nearly short.

3)  Remember that "spacings" is a special case of 
    insulation, either air insulation (clearance) or a 
    surface insulation (creepage).  Only those insulations
    that are required by the safety standard (i.e., basic,
    supplementary, or reinforced) are subject to the
    spacings requirements.  Typically, these insulations
    only exist between primary and ground, and between
    primary and secondary.  Typical products do not have
    components between primary and ground and between
    primary and secondary except those specifically 
    rated for such use such as Y-caps, transformers, and
    opto-isolators (and which therefore meet the spacing 
    requirements of the end-product).

    Some standards may require pole-to-pole spacings in
    primary circuits.  In this case you must measure the
    voltage and then determine the spacing from a table.  
    A typical SMPS has lots of low-voltage control
    circuits with respect to the negative rail.  So, you
    can lump all those circuits together as not requiring
    insulation from each other.  Then, they can be taken
    as a whole and spaced from the positive rail.  That
    will generally only leave the bulk capacitor and the
    switching transistor(s) as requiring spacings.  And,
    the snubber circuit, which can be considered a voltage
    divider so that the spacings across any individual
    snubber component need not be the full voltage across 
    the snubber.  Etc.

It really makes no sense to require a PWB to have greater
spacings than the component itself.  It further makes no
sense to enforce spacings across capacitors (that are not
Y capacitors) and semiconductors and similar components.
Internally, these components do not have insulations that
are equivalent to their terminal spacings or to the PWB
spacings.  So, why require a higher level of insulation
than the device itself can provide?  Short-circuiting of
the component will tell the story of whether the circuit
is safe; if safe, then the spacing is inconsequential to
the safety of the product.


Best regards,
Rich





-------------------------------------------
This message is from the IEEE EMC Society Product Safety
Technical Committee emc-pstc discussion list.

To cancel your subscription, send mail to:
     majord...@ieee.org
with the single line:
     unsubscribe emc-pstc

For help, send mail to the list administrators:
     Jim Bacher:              jim_bac...@mail.monarch.com
     Michael Garretson:        pstc_ad...@garretson.org

For policy questions, send mail to:
     Richard Nute:           ri...@ieee.org

Reply via email to