From: "Tom Faloon"
Subject: Re: Why use resistors to tie a logic input to ground?
Date: Thu, 17 Oct 2002 16:47:48 +0100
NNTP-Posting-Date: Thu, 17 Oct 2002 15:49:03 +0000 (UTC)
X-Newsreader: Microsoft Outlook Express 5.00.2615.200
Tim Shoppa wrote in message
> firstname.lastname@example.org (Manvinder Bhullar) wrote in message
> > Hi All,
> > I'm looking at this circuit from Anchor Chips that shows unused inputs
> > of a 74LCX14 tied to ground with 10K resistors.
> > Can someone please advise what the rationale is behind this? I've
> > always tied unused inputs directly to ground. What role do the
> > resistors play?
> If you tie them directly to ground or Vcc, *and* you have substantial
> ground bounce (most of a volt), then you can violate the "Absolute
> Maximum Ratings" for input voltage and current. For typical TTL-like
> logic these say that inputs cannot go below -0.5V or above Vcc+0.5V
> unless you ensure the current is limited.
> If you put a resistor in series, then you do not violate the "Absolute
> Maximum Ratings" because you stay under the input diode current rating.
> Typical ground bounce in a small, well bussed system will not be above
> half a volt and you can hook the input directly to Vcc or Ground. Even
> if you have severe ground bounce problems you can typically get away
> with hooking genuine TTL inputs directly to Vcc or Ground because the
> ground bounce is a transient effect - but you *are* violating the specs.
The question was about unused inputs. I don't see how ground bounce can
damage unused inputs if they are tied directly to 0V, and tied, as they
usually are, reasonably close to the chip's own ground pin. In this
situation both input and chip ground are at the same potential, and held at
the same potential by a PCB track. I can't think of a better way of
protecting the input, regardless of how much ground bounce there is! Using
a current limiting resistor is meaningless, since the voltage across it is
zero, or very nearly so, and no input current can flow anyway.
I would apply exactly the same argument to pull up resistors in most modern
MOS type devices.
I agree there could be a potential problem if there was horrendous ground
bounce, AND the input was connected to 'ground' at a point physically
distant from it's own chip ground. This situation would be unusual. Ground
bounce 'abounds' but unused inputs are usually grounded close by.
I do agree with using a pull down, (or pull up) resistor where possible,
because it allows access to the input for PCB mods, and allows the input
level of active chips to be forced, during debug or fault finding. This has
been discussed in other posts.