Snubber Network

[[email protected]]

I've been reading http://www.hagtech.com/pdf/snubber.pdf and decided
it would be interesting to lay down on my power supply board with a
view to investigating when i had the time. However i'm left a little
confused as to where to place the RC network. The PSpice model in
figure 7 seems to suggest across the diode to GND effectively and the
text talks about the equivalent of a full wave bridge rectifier having
both diodes in series (which of course makes sense), which then places
the RC across both diodes to GND. However other texts seem to suggest
the RC should go across each diode? Can anyone give me any clues on
this one?

Thanks
Andrew

 

[Paul Mathews]

I've been readinghttp://www.hagtech.com/pdf/snubber.pdfand decided
it would be interesting to lay down on my power supply board with a
view to investigating when i had the time. However i'm left a little
confused as to where to place the RC network. The PSpice model in
figure 7 seems to suggest across the diode to GND effectively and the
text talks about the equivalent of a full wave bridge rectifier having
both diodes in series (which of course makes sense), which then places
the RC across both diodes to GND. However other texts seem to suggest
the RC should go across each diode? Can anyone give me any clues on
this one?

Thanks
Andrew

RC across each diode is best. The theory is interesting and sometimes
useful, but here is an easy procedure that usually works:
1) observe ringing with oscilloscope and measure ring frequency f0
2) add capacitance C across the switching element (diode in this case)
until f = f0/2. i.e., ring freq is half of f0. This capacitance will
be in the range of the stray and parasitic capacitances of the circuit
and the diode, between 10 pF and 1000 pF usually.
3) Calculate R = 1/(2*pi*f) as a rough estimate of required damping
resistance, and choose a miniature trimpot adjustable above and below
R. Adjust pot for critical damping.

Take care to use very short leads for everything, including scope
ground. In switching power supplies, resistor dissipation may be an
issue. You can estimate power dissipation from:

P = F * C * V^2 / 2 where F is the switching freq and V is the
peak to peak voltage change across the switching element for each
switch cycle. If P is too large for practical reasons, compromise by
reducing C and readjusting R for best results. You can still achieve
critical damping, but the amplitude of the overshoot pulse will be
greater.
Paul Mathews

 

[[email protected]]

RC across each diode is best. The theory is interesting and sometimes
useful, but here is an easy procedure that usually works:
1) observe ringing with oscilloscope and measure ring frequency f0
2) add capacitance C across the switching element (diode in this case)
until f = f0/2. i.e., ring freq is half of f0. This capacitance will
be in the range of the stray and parasitic capacitances of the circuit
and the diode, between 10 pF and 1000 pF usually.
3) Calculate R = 1/(2*pi*f) as a rough estimate of required damping
resistance, and choose a miniature trimpot adjustable above and below
R. Adjust pot for critical damping.

Take care to use very short leads for everything, including scope
ground. In switching power supplies, resistor dissipation may be an
issue. You can estimate power dissipation from:

P = F * C * V^2 / 2 where F is the switching freq and V is the
peak to peak voltage change across the switching element for each
switch cycle. If P is too large for practical reasons, compromise by
reducing C and readjusting R for best results. You can still achieve
critical damping, but the amplitude of the overshoot pulse will be
greater.
Paul Mathews

Excellent, thanks for your reply its exactly what i needed.

Andrew