Ultrafast recovery diodes

[oparr]

Currently designing something for which an ultrafast recovery diode is
recommended but on checking the specs and package types for devices so
classified at Digikey, I don't like what I see....Current ratings start at
1A which is much more than I need and SMD package types are typically large
and bulky SMB or SMA. Forward and reverse recovery times are typically 25 -
50ns.

Looking at some switching diodes not claiming to be fast, superfast or
ultrafast, I see current ratings and package types right up my street.
Reverse recovery times are as low as 4ns but no mention is made of forward
recovery times. Two questions...What are all the relevant attributes of a
ultrafast recovery diode? Why wouldn't a switching diode with a 4ns reverse
recovery time not be more than suitable as an ultrafast recovery diode?

 

[Jan Wagner]

Looking at some switching diodes not claiming to be fast, superfast or
ultrafast, I see current ratings and package types right up my street.
Reverse recovery times are as low as 4ns but no mention is made of forward
recovery times. Two questions...What are all the relevant attributes of a
ultrafast recovery diode? Why wouldn't a switching diode with a 4ns reverse
recovery time not be more than suitable as an ultrafast recovery diode?

4ns sounds like a schottky diode or small-signal diode. These are
all fast and don't have to be explicitly hyped as "ultrafast".

AFAIK there are no special required attributes or values for
"ultrafast", "fast", "standard" etc. It depends on the manufacturer
or reseller and is mostly marketing talk. E.g. the BYW98 (50ns trr)
diode is called "fast recovery" by ST, but all resellers seem to
think it's "ultrafast".

Diodes may also be named "slower" than what they are. For example
the 1n4148 (max 4ns trr) is usually either called "small-signal" or
"fast", but haven't seen "ultrafast" anywhere yet.

And then there's also "soft recovery", but again, what can be
considered an objective criteria to qualify as "soft recovery"?

So, summary: marketing BS. But okay, I'm not in the
semiconductor&process business, so maybe someone else can give a
better answer or correct me here...

- Jan

 

[Rich Grise]

4ns sounds like a schottky diode or small-signal diode. These are
all fast and don't have to be explicitly hyped as "ultrafast".

AFAIK there are no special required attributes or values for
"ultrafast", "fast", "standard" etc. It depends on the manufacturer
or reseller and is mostly marketing talk. E.g. the BYW98 (50ns trr)
diode is called "fast recovery" by ST, but all resellers seem to
think it's "ultrafast".

Diodes may also be named "slower" than what they are. For example
the 1n4148 (max 4ns trr) is usually either called "small-signal" or
"fast", but haven't seen "ultrafast" anywhere yet.

Apparently you've never googled for one yet then. They are used in
SPSs when Schottkys don't have enough reverse breakdown. They're
like ordinary silicon diodes, but not with the Schottky doping but
with doping and other characteristics that minimize reverse carrier
recovery time, making them "ultrafast", like, switching in times
comparable to a Schottky or signal diode, but at power diode
power levels.
http://www.google.com/search?q=ultrafast+(diode+OR+rectifier)

Cheers!
Rich

 

[Winfield Hill]

Rich Grise wrote...

 

[John Popelish]

Don't forget ultrafast's brother, or <cough> sister, ultrasoft.
http://www.google.com/search?q=ultrasoft+(diode+OR+rectifier)

These would make a nice, wide range varactor. 470 pF to 47 pF with
less than a 10 volt bias change.
http://www.fairchildsemi.com/ds/IS/ISL9R8120S3S.pdf

 

[[email protected]]

Apparently you've never googled for one yet then.

Google? What's that?

switching in times comparable to a Schottky or signal diode, but at power diode
power levels.

Yes, that's about right. Trr is specified at higher current levels
(typically 500ma as opposed to 30ma). Seems kind of dumb for
Intenational Rectifier to specify an ultrafast diode (with 1A current
rating) where current levels never exceed 50ma.