From: Mike Monett
X-Mailer: Mozilla 2.02 (Win16; I)
Subject: Re: LM317 or 7808 is better performance for 9V output ?
References: <3DEDABCA.906FB592@fanwap.com> <firstname.lastname@example.org> <email@example.com> <3DF5DF73.1FB2@Spam.Bots>
Date: Wed, 11 Dec 2002 06:48:59 -0800
NNTP-Posting-Date: Wed, 11 Dec 2002 06:49:01 EST
Organization: Bell Sympatico
Winfield Hill wrote:
> Mike wrote...
>> Bob Wilson wrote:
>>> Noise especially. 3-terminal regulators will do nothing to clean
>>> up noise.
>> The regulators can contribute a great deal of noise, which can be
>> a problem in pll's and low-level amplifiers. I recently analyzed
>> four different methods of reducing the noise and posted the
>> results to
> Anyone using a LM317 who's concerned with noise will certainly add
> the suggested filter cap from the ADJ pin to ground; You left this
> important part out of your drawings and analysis. The same advice
> goes for Bob's comments. Adding this cap lowers the 10Hz to 10kHz
> noise from the 317's spec'd 30ppm of Vout, to 30ppm of 1.2V =
> 36uV. By comparison a 7812 is spec'd with a massive 75uV of noise.
> So there, all you LM317 nay-sayers!
> Mike, further to your various ripple-reducing schemes, you have
> failed to include my favorite highly-effective filter, which is a
> simple FET or BJT 2nd-order filter on the INPUT of the regulator.
> Add it just after the rectifier energy-storage filter capacitor,
> like this:
> . ,--||----------, _____
> . | | | |
> . ---+--------|---- D S --+--+---| |-----+--- out
> . | | G | |_____| |
> . | | | === | ===
> . '--/\/\--+--/\/\--+ | gnd |
> . | gnd gnd
> . ===
> . |
> . gnd
> It can be much more effective to eliminate the problem at its
> source, rather than to struggle with it after the regulator.
> Thanks, - Win
Thanks for taking the time to look at the article. It was written
for situations where you have to work with the noise and ripple on
an existing supply, and can do little to improve the source. An
example is upgrading the local oscillator in a receiver, such as "A
Versatile DDS/PLL Synthesizer for UHF and Microwave Projects", by
Another example is Tom Bruhns comment in a post on Nov 21:
"(Man, am I tired of looking at phase noise sidebands [in a
digitizer system, even] today! And I hate working in an
environment where I'm given noisy supplies with no headroom to
I didn't omit the filter cap from the ADJ pin to ground on the
LM317. The only place the LM317 appears is in Figure 1, which was
copied from Wenzel's article. He forgot it, not me!
Figure 1 shows his original cancellation circuit. It required
considerable modification, and didn't perform very well. I was
forced to look for alternatives, and felt the information might be
useful to others.
I'm sorry I didn't include your favorite prefilter circuit.
Certainly this would help reduce ripple, but it would not affect the
regulator noise. Again, the article was aimed at applications where
you have little or no control over the noise and ripple on the
incoming supply voltage, and need to find ways to reduce it.