From: John Larkin
Subject: Re: isolated ckts for LM339 quad comp
Date: Wed, 30 Oct 2002 13:43:01 -0800
Organization: Posted via Supernews, http://www.supernews.com
X-Newsreader: Forte Agent 1.91/32.564
On Wed, 30 Oct 2002 18:34:24 GMT, Jim Thompson
>On Wed, 30 Oct 2002 10:13:22 -0800,
>John Larkin ,
>In Newsgroup: sci.electronics.design,
>Entitled: "Re: isolated ckts for LM339 quad comp",
>Wrote the following:
>|On Wed, 30 Oct 2002 16:15:52 GMT, "Dan"
>|>I want to use LM339's in an automotive app. Can I have the inputs (+ and -)
>|>on a separate 5vdc ckt which is isolated from the normal 12vdc which powers
>|>the chip (VDD, VSS) and supplies the +12vdc pullups for the output? BTW, I
>|>am using TVS's and decoupling caps to suppress transients and the +5 ckt is
>|>on the output side of a 12v->5v dc-dc converter. See my previous post: "can
>|>someone recommend a flyback zener diode value..." and thanks to all who
>|******DO*****NOT***** (sorry for shouting) *******EVER******* let any
>|of the LM339 inputs go below the chip ground pin, even by a few
>|hundred millivolts. It will make the entire chip go haywire.
>|Nasty gadget, overall.
>That's a wee bit of an extreme recommendation.
>Each input can safely go almost a vBE below ground before there *may*
>be excessive current drawn... protect by using a finite source
>As the negative input gets near a vBE below ground the output phase
>will flip. That can be a problem in digital loops.
>I used to use Ge diodes to ground to avoid such problems.
> ...Jim Thompson
I've seen all four comparator outputs go nuts if a single input went a
few hundred mV below ground. The National datasheet, in a classicly
microscopic footnote, says -300 mV. It's a footnote to the Absolute
Maximums section, I suppose so that not too many people would be
annoyed by reading it by accident. LM324 is similar but worse.
"Only the paranoid survive."
- Andy Grove