Regulator noise

[Michael]

I'm wanting to put together an xmtr and rcvr pair using some Linx modules
(ES series 916mhz) and it's a very simple straightforward process to get it
done but....

I'd like to learn some more about the power supply requirements and the
issue of regulator noise.

I see alot of suggestions for using a LDO regulator like a Micrel part and
it seems that it's not just the fact that's low drop out but also has a good
"noise figure". (The spec sheet says the module needs a supply with noise
and hash below 20 mv).

I'd like to use one of my LM317 parts if possible (because I have a ton of
them).

I know any circuit generates noise but where are the concerns with power
supply regulators? What is going on in the regulator and what is the freq
spectrum of particular concern?

Sure, I know all about fighting 60hz ripple...but I'd like to hear about the
higher freq components that are in the mix.

I suppose the fact that the 317 requires a larger input voltage (compared to
a LDO) is a factor in noise generation as well?

THANKS for help.

 

[Andrew Holme]

Michael wrote:

I'm wanting to put together an xmtr and rcvr pair using some Linx
modules (ES series 916mhz) and it's a very simple straightforward
process to get it done but....

I'd like to learn some more about the power supply requirements and
the issue of regulator noise.

I see alot of suggestions for using a LDO regulator like a Micrel
part and it seems that it's not just the fact that's low drop out but
also has a good "noise figure". (The spec sheet says the module needs
a supply with noise and hash below 20 mv).

I'd like to use one of my LM317 parts if possible (because I have a
ton of them).

I know any circuit generates noise but where are the concerns with
power supply regulators? What is going on in the regulator and what
is the freq spectrum of particular concern?

Sure, I know all about fighting 60hz ripple...but I'd like to hear
about the higher freq components that are in the mix.

The LM317 is perfectly adequate for your application. The datasheet quotes
internally generated regulator noise as a percentage of output voltage. The
frequency spectrum of concern is low to audio frequency.

Higher frequencies are dealt with using passive power supply filtering and
decoupling e.g. using 100nF decoupling capacitors, and - if necessary - RF
chokes.

 

[Michael]

The LM317 is perfectly adequate for your application. The datasheet
quotes
internally generated regulator noise as a percentage of output voltage.
The
frequency spectrum of concern is low to audio frequency.

Higher frequencies are dealt with using passive power supply filtering and
decoupling e.g. using 100nF decoupling capacitors, and - if necessary - RF
chokes.

Which is something I've always done......OK, thanks.
I just looked at the datasheet on the 317 and see they show it as a
percentage (.003) from 10hz to 10khz.

(15mv for a 5v output?)

If it's audio be choosy and use good design but if it's RF
again, use good design but...
the internal noise isn't a factor?

 

[Andrew Holme]

Michael wrote:

The LM317 is perfectly adequate for your application. The datasheet
quotes
internally generated regulator noise as a percentage of output
voltage. The
frequency spectrum of concern is low to audio frequency.

Higher frequencies are dealt with using passive power supply
filtering and decoupling e.g. using 100nF decoupling capacitors, and
- if necessary - RF chokes.


Which is something I've always done......OK, thanks.
I just looked at the datasheet on the 317 and see they show it as a
percentage (.003) from 10hz to 10khz.

(15mv for a 5v output?)

..003% of 5V is 0.15mV

If it's audio be choosy and use good design but if it's RF
again, use good design but...
the internal noise isn't a factor?

Low frequency noise on the power supply can be a problem in any application.