Bridge rectifiers and a diode's reverse recovery time?
hrh1818
calculate the value a diodes reverse recovery time specification must
meet to avoid shoot through. But what happens when the input to the
bridge rectifier is the output of a SMPS and is a square wave with a
frequency of 25 KHz and a 220 volt peak to peak amplitude? The
Schottky rectifiers I am familiar with have to low of a reverse
voltage rating for this application. Can SPICE be used to calculate
the average power dissipation in a diode during the revers recovery
period? Are designers forced to use 1/2 wave rectifiers in this
application?
Heres hoping the members of the sci.electronics.group can come through
once again,
Howard
Tim Williams
news:aed332b7-e123-4bc6...@l30g2000yqb.googlegroups.com...
> If the input to a bridge rectifier is a sinusoidal voltage, one can
Ed: ^
> calculate the value of a diode's reverse recovery time specification
Ed: ^^ ^
> that must be met to avoid shoot through. But what happens when the input
Ed:^^^ ^^ ^
> to the bridge rectifier is the output of a SMPS and is a square wave with
> a frequency of 25 KHz and a 220 volt peak to peak amplitude?
Ed: I have corrected your quote above to remove ambiguities. Please confirm
if I've read it correctly, otherwise I'm not quite sure what you've said.
Is it a forward converter? What is the typical duty cycle? Why such a low
frequency?
> The Schottky rectifiers I am familiar with have too low of a reverse
> voltage rating for this application.
They do make SiC schottkies up to 800V, but they're way more expensive than
needed at a paltry 25kHz.
Tim
--
Deep Friar: a very philosophical monk.
Website: http://webpages.charter.net/dawill/tmoranwms
Phil Allison
"Tim Williams"
"hrh1818"
>
> But what happens when the input to the
> bridge rectifier is the output of a SMPS and is a square wave with a
> frequency of 25 KHz and a 220 volt peak to peak
> amplitude?
> Is it a forward converter? What is the typical duty cycle?
** Generally speaking - what is the typical duty cycle of a square wave ??
Hmmmmmm.....
> Why such a low frequency?
** Opinion - not fact.
... Phil
Shaun
"hrh1818" <hr...@att.net> wrote in message
news:aed332b7-e123-4bc6...@l30g2000yqb.googlegroups.com...
How much current are you putting through them.
The UF4007 is ultra fast and will handle 1 amp
the UF5408 will handle three amps and both diodes are rated to 1000 volts.
Shaun
hrh1818
> "hrh1818" <hr...@att.net> wrote in message
>
> news:aed332b7-e123-4bc6...@l30g2000yqb.googlegroups.com...> If the input to a bridge rectifier is a sinusoidal voltage, one can
>
> Ed: ^> calculate the value of a diode's reverse recovery time specification
>
> Ed: ^^ ^
>
> > that must be met to avoid shoot through. But what happens when the input
> Ed:^^^ ^^ ^
> > to the bridge rectifier is the output of a SMPS and is a square wave with
> > a frequency of 25 KHz and a 220 volt peak to peak amplitude?
>
> Ed: I have corrected your quote above to remove ambiguities. Please confirm
> if I've read it correctly, otherwise I'm not quite sure what you've said.
Sorry I do not understand the meaning of you editing marks. Hence I am
unable
to you have read what I wrote correctly.
>
> Is it a forward converter? What is the typical duty cycle? Why such a low
> frequency?
No this is a push pull isolated output boost converter. The duty cycle
is
approximately 50%. A 25 kHz frequency is typical for SMPS.
Phil Allison
"hrh1818"
>
> Are designers forced to use 1/2 wave rectifiers in this
> application?
** Never seen a bridge made from high speed power diodes.
Nearest you get is " fast recovery " silicon diode pairs, in common cathode
and common anode types.
Using one of each makes a high speed, high current bridge rectifier.
.... Phil
Robert Baer
Use of half wave geometry will not necessarily guarantee less loss.
Robert Baer
Wimpie
Hello Howard,
Reverse recovery time in Si rectifiers (or charge when you integrate
the reverse current peak versus time) depends on actual current before
reversing voltage, dI/dt , maximum reverse current and temperature.
High junction temperature, and high dI/dt, increases reverse recovery
time. Several manufacturers specify reverse recovery time at different
temperature and dI/dt. They also can cheat you by taking the maximum
reverse recovery current far higher then the forward current. This
results in less recovery time in the spec, but doesn't change the
recovery charge.
When your SMPS output has low output impedance at reversing polarity
(for example a full bridge circuit, or a push-pull fed from a solid
supply), you will get high reverse current that will result in high
switching loss in the switches (probably more loss then in the
rectifiers).
A transformer's leakage inductance or just a series inductance can
limit dI/dt, hence the reverse current peak. This inductance can also
be used to create zero voltage switching.
Regarding simulation. The standard diode model isn't very good. Try to
simulate the reverse recovery setup as used by the manufacturer and
tweak with the parameters (transit time) to have a good match between
datasheet and simulation.
A push pull circuit can be used with a full bridge rectifier or center
tapped transformer with dual diode rectifier (with very good
efficiency). I used series inductance to reduce dI/dt. When diode
speed is really an issue, you might consider expensive SiC (Silicon
Carbide Schottky) rectifiers. These diodes have low reverse recovery,
but still you have to charge and discharge the diode capacitance. This
may also result in high current peaks in a hard switched application
(hence losses in your switches).
Best regards,
Wim
PA3DJS
www.tetech.nl
The email address is ok, but remove abc first
hrh1818
I agree. After I wrote "Are designers forced to use 1/2 wave
rectifiers in this
application?" I have since learned I was all wet when I asked this
question.
Howard
mook johnson
"hrh1818" <hr...@att.net> wrote in message
news:aed332b7-e123-4bc6...@l30g2000yqb.googlegroups.com...
You're waisting time even considering doing such a thing. 60Hz bridge
rectifiers usually don't even specify a reverse recovery time because at
60Hz its usually not an issue.
Even at 25kHz you'll want some soft recovery diodes (trr <60nS) of the
appropriate voltage and current rating.
Don't go rediculously oversized on the diode current rating either because
the junction capacitance goes up with current rating. High Junction
capacitance will cause di/dt radiated noise, ringing, and a high current
spike in the switches at the transitions.
legg
wrote:
>If the input to a bridge rectifier is a sinusoidal voltage one can
>calculate the value a diodes reverse recovery time specification must
>meet to avoid shoot through. But what happens when the input to the
>bridge rectifier is the output of a SMPS and is a square wave with a
>frequency of 25 KHz and a 220 volt peak to peak amplitude? The
>Schottky rectifiers I am familiar with have to low of a reverse
>voltage rating for this application. Can SPICE be used to calculate
>the average power dissipation in a diode during the revers recovery
>period? Are designers forced to use 1/2 wave rectifiers in this
>application?
>
Reverse recovery is not a loss mechanism associated with the diode
itself. The reverse voltage across the diode does not rise
significantly until after the peak reverse current is reached, when
the voltage assumes a roughly capacitive charging relationship that
depends again on diode characteristic "snap".
During the early period of recovery, full circuit voltage, load
current and recovery current are carried by series elements;
switch/source, resistive and inductive components or strays.
"Snap" is a term that you will find in few specifications but is
basically the relationship between the slopes of rising and falling
reverse currents - the falling reverse current slope being a function
of doping chemistry and profile.
The total charge is dependant on the diode chemistry and doping
profile, physical size ( 1/^ ), current density (^), temperature (^)
and externally enforced di/dt (^) during the early switching period.
This is non-linear, and not easily spiced, even when a range of
relevant diode response characteristics have been recorded. Usually
the work is simplified by choosing one known circuit situation that is
relevant - worst case or otherwise critical to specific performance
requirements at environmental extremes, and by splitting the switching
period into two-part approximations that are more easily linearized.
In any event, demonstrated performance cannot be ignored or avoided,
though a comparison of printed spec characteristics can be used (with
a grain of salt) to forcast the relative performance of two or mpre
devices in a practical application.
RL
legg
wrote:
Actually, substitution of larger devices of similar technology (in an
otherwise unchanged circuit) can offer switching loss improvements in
some power applications through the reduction in current density,
prior to turn-off. Their use may also allow the achievement of lower
junction temperatures that can also be beneficial in this regard.
Specs are often misleading, in that they do not often directly compare
performance under derated conditions.
RL