On Tue, 14 Jul 2009 15:45:12 -0700, Jim Thompson
Jim Thompson wrote:
Jim Thompson wrote:
Jim wrote:
On a sunny day (Tue, 14 Jul 2009 13:10:57 +0100) it happened "OldGit"
Hello
This is from the datasheet, and I don't really understand it:
"Icc quiescent supply current 80uA.
deltaIcc additional quiescent supply current per input pin Vi=Vcc - 2.1V,
other inputs at Vcc or GND. Vcc=4.5 to 5.5V
An or Bn inputs 180uA
OE input 675uA
DIR input 405uA"
I get the bit about Icc at 80uA, and I get the bit (not quoted) about
dynamic power. But what's the additional Icc per input pin about? And why is
it at Vcc - 2.1V?
Sorry if it's a dumbo question.
Ta for any hints.
Have not read that datasheet, but likely when halfway 'on'
both of some MOSFETS (pull up + pull down) are partly on, so then it draws some power.
I would think that it's the minimum high level input voltage that you
might expect when interfacing to TTL (and the resultant current due to
the voltage being 'halfway on', as noted above).
Jan is right about "half way on". The input inverter (even that of a
Schmitt trigger such as the 74HC14) will conduct substantial quiescent
current when the input pin resides in the middle. The middle is defined
by the IC designers, for HC logic it is truly in the middle, meaning at
VCC/2.
The spec for "middle" is actually 30%-70% :-(
Yeah, you's guys are always +/-8dB while us guys limit ourselves to
+/-3dB
In my custom designs I often make such input devices either quite long
or use current sources and sinks to limit "overlap" current.
When it comes to very speedy devices the chip designer is probably
between a rock and a hard spot.
Wouldn't "speedy" imply very little time spent in the overlap region?
Only slow stuff, or stuff that can float around, requires such
limiting.
Sure, but when the input signal isn't always under your control the
overlap region remains a concern. I get that question over and over
again: "Why is this draining the battery so fast" ... measure, measure,
measure ... "Well, those two inputs are not anywhere close to logic
levels" ... "But it's a Schmitt" ... "Doesn't matter, the input inverter
still seeps" ... "Oh".
You need to learn to speak harshly to your customers when they're
doing dumb-ass things ;-)
No, you speak softly and carry a big bill.
Of course, in my case, I can contrive my own frontend to prevent/solve
such issues.
But, "speedy" and long devices (and other protection means) don't tend
to go together. C == bad
All my "floaty" inputs are slow.