Re: [SI-LIST] : 3.3V Design

Joe Cahill (jjcahill@us.ibm.com)
Tue, 4 Aug 1998 13:25:23 -0400

Jim,

Think in terms of % of supply voltage and the crucial differences will become
more apparent. Noise margin and voltage swing represented as % of supply will
show you more than if you represent them as absolute voltage levels.

3.3v interfaces are more balanced than 5v TTL, ie the thresholds are closer to
the middle of the swing, so you get improved relative noise margins.

Typically 3.3v interfaces are pure cmos so potentially you can have balanced
drivers and balanced receivers and therefore balanced noise margins.

There are lots of well-behaved 3.3v MSI parts for driving nets the proper way,
and these are not available in 5v interfaces. Look around and take advantage of
them if you can.

The clamping voltages for overshoots above and below the rails are a larger
percentage of your supply, so potential ringing can become a larger percentage
of the total signal swing.

Typically 5v TTL drivers didn't use all of the upper level voltage range for
the signal, and rarely did a rising transition overshoot and forward bias the
clamp diodes, this is a much more common occurence in 3.3v interfaces.

Be careful if you have 5v interfaces and 3.3v interfaces in close physical,
electrical or timing proximity. If the 5v signals influence the 3.3v signals
you could have troubles because of the lower absolute value of the 3.3v noise
margins.

Joe

CEC Analysis and I/O Design
Phone: 507.253.0762 Fax: 507.253.4966
jjcahill@us.ibm.com