From: Winfield Hill
Subject: Re: LM317 or 7808 is better performance for 9V output ?
Date: 10 Dec 2002 02:44:10 -0800
Organization: Rowland Institute
References: <3DEDABCA.906FB592@fanwap.com> <3DEDCD23.48E4B199@earthlink.net>
X-Newsreader: Direct Read News 4.20
John Woodgate wrote...
> Winfield Hill wrote:
>> I was afraid someone would mention that. Strictly in the
>> interest of exploring philosophical possibilities of course.
> I've seen it recommended in several places. What's wrong with it?
> Obviously, there is a performance degradation, but is there anything
780x IC are made by several manufacturers. Flipping through
the datasheets examining the ground-lead specs and curves, I
see the tempco ranges from +0.5mA/100C to -1mA/100C, the load
variation is spec'd at 0.5mA max, and the curent variation with
input-voltage is about 1.5mA (when low overhead voltages are
included). So the uncertainty in ground current can be as bad
as 2 to 3mA, on a quiescent current ranging from 4 to 8mA, which
compares to changes of about 5uA on a quiescent current of 50 to
100uA for the LM317.
So for the 780x we'd want to use a divider current about 600 x
higher than for a 317 to get the same performance sensitivity.
The suggested LM317 divider current is 5mA, so for a 780x we'd
be looking at something in the vicinity of 3 amps! Gurraapp!
Clearly one would compromise on a divider current of say 50mA
for the 780x, even though this means the portion of the output
voltage due to the divider could change by up to 5%. Our O.P.
wanted a 9.0V output. If he started with a 7805 this 5% would
be reduced to a 2.2% variation (0.2 volts), which might satisfy
him fine. He'd do better if he found a 7808 to make his supply.
So yes, a performance degradation, but perhaps not disastrous.
As for me, I'd stick with a LM317 or LM350.