From: Winfield Hill
Subject: Re: LMC6001AIN
Date: 14 Sep 2002 12:59:17 -0700
Organization: Rowland Institute
X-Newsreader: Direct Read News 2.97
> The LMC6001 uses CMOS input devices. This should have been clear on
> the data sheet but when I looked at the DS the only reference to CMOS
> was deep inside in the section on Latchup.
> Regards... Jerry
> "vishwa" wrote
> Could anybody tell me about the input structure of LMC6001AIN? Is it
> a bipolar/FET/DiFET/MOSFET stage? I tried to find the info on the web
> (and also on the datsheet!), but could not come up with the answer :-(
Vishwa, the input is almost certainly a p-channel MOSFET differential
long-tail pair, powered by a pMOS current source. We can deduce this
from the input current and common-mode voltage operating range, which
inlcudes ground, but excludes within 1.9V of the plus rail. The input
MOSFET drains would be combined with a classic NMOS current mirror to
form a single-ended current output for the gate of the next stage.
Such an input stage appears on page 3 of TI's TLC272 opamp datasheet,
see http://www-s.ti.com/sc/ds/tlc272.pdf However the rest of NSC's
opamp circuitry differs substantially from TI's circuit. E.g., see
LMC6024 datasheet page 9, http://www.national.com/ds/LM/LMC6024.pdf
The LMC6001's input voltage-noise density is high, like a typical CMOS
opamp's, 22nV/Hz^1/2, comparable to 25nV for the TLC272. In contrast,
a JFET opamp would have lower input noise, but higher input current.
The LMC6001 has only 0.01pA of input current, clearly a MOSFET gate.
A common cheap TL082 JFET opamp has 30pA input current and 18nV noise.
As for details of NSC's low-leakage input protection circuit, all the
IC manufacturers consider that to be highly proprietary information,
and don't provide any detail. Sorry.