[SI-LIST] Parasitic resonance in DC-DC converters
- From: "KR, Chandrashekhar" <Chandrashekhar.KR@xxxxxxx>
- To: si-list@xxxxxxxxxxxxx
- Date: Fri, 5 Sep 2003 23:06:19 +0530
Hi all,
This may be slightly out of the SI realm, but should be interesting to most
people out here.
I am trying to kill some ringing in my synchronous buck converter (3.3V to
1.1V, load current 15A). The ringing is, as usual, at the junction of the
main MOSFET, the synchronous MOSFET and the output inductor. The basic aim
in this case is to reduce the peak overshoot as much as possible and damp
out the oscillations to a minimum. I have to resort to a snubber (I did the
best possible layout I could, and can't change it), and fortunately I do
have provision for a RC snubber on the board.
On a previous revision of the board, I could kill the ringing with RC
snubbers across either switches (the snubber across the main switch took
care of the turn-off ringing, whereas that across the synchronous switch
reduced the turn-on ringing). The method chosen to arrive at the snubber
values was one that is common in the literature - adding capacitance across
the suspect node till the ringing frequency is halved, so that the parasitic
capacitance of the tank circuit is known; and then adding appropriate
resistance (ideally equal to the characteristic impedance of the tank
circuit) to damp the oscillations.
When the board underwent a revision, I made provision for a RC snubber right
across each MOSFET. In addition, I added a little extra copper in the power
path, a few more vias, and that's about it.
On the current revision, though, the same technique doesn't work:
- The ringing frequency (no snubber added) is the same as that on the
previous board (62.5MHz)
- Adding capacitance across either FET does alter the ringing frequency; the
FET o/p capacitance is roughly 2.5nF, and adding a 4.7nF capacitor across
it does cause the frequency to reduce to around 30MHz, but in addition there
is a new frequency component of around 72MHz riding on it, though it's
amplitude is quite low.
- Upon adding a resistor (I've tried 1 ohm to 10 ohms) in series with the
capacitor, the 30MHz is no longer seen, and the 72MHz ringing becomes
prominent... almost as bad as the 62.5MHz ringing that I tried to kill.
- For other reasons, I also happened to replace the MOSFET with a different
device (a lower gate charge device, but capacitance characteristics are not
way different). With this, the resonance occurs at around 83MHz, and is
twice as nastier in terms of amplitude; nothing that I try, on the same
lines as above, seems to alter the frequency or the amplitude of the ringing
at all. The MOSFET datasheets do not mention any inductance characteristics
of the device; assuming they are not very different, I guess the difference
in behaviour is because of different turn-on/turn-off characteristics.
Looking forward to any inputs that might be useful. Further, if somebody has
worked on modelling such circuits (including parasitics) to a reasonable
degree of accuracy, good enough for practical purposes but not too involved,
I'd like some clues on how I could proceed with modelling this.
Regards,
Chandrashekhar K
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List archives are viewable at:
//www.freelists.org/archives/si-list
or at our remote archives:
http://groups.yahoo.com/group/si-list/messages
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
Other related posts:
- » [SI-LIST] Parasitic resonance in DC-DC converters
