From: Ray Anderson ([email protected])
Date: Tue Nov 07 2000 - 10:18:11 PST
I'm sure there are numerous ways of selecting parts to create
a power filter to supply voltage to the PLL circuit, but I'll
make a few comments on what I've found to work reasonably well.
You mention that you want to filter out noise "up to a few MHz".
This sounds like a high pass filter. Typically I find that you want a
low-pass response (from DC to some freq). A simple L section or perhaps
pi section filter will work well.
+----R----L-----+------+
| | |
Vsrc C Rload
| | |
+---------------+------+
People will typically just pick some arbritrary parts (L and C) and assume
that it is going to work OK. Not always the case..... Decide on a corner
frequency and then synthesize a filter using standard passive filter
synthesis techniques. A couple of gotcha's are that if you design the filter
for equal terminations on the input and output side you will most likely
experience SEVERE peaking (maybe > 10dB) at or about the corner frequerncy
even though you designed for a nice sane Buttersworth or Chebychev response.
This peaking can actually induce MORE noise voltage at the output of the
filter than you have at the input. Why the peaking ? Because you are driving
the filter from an ultra-low impedance power source instead of one with a
nice defined impedance. This can be dealt with by including a small (on the
order of a couple ohms) de-Qing resistor at the input of the filter. Build a
spice simulation where the load for the filter is a resistor valued at
Vdd/Ipll (probably a couple hundred ohms or so) and the source impedance of
the voltage source is really low (milliohms). Do a freq domain analysis and
play with the series resistor until you get a nice LPF rolloff without
objectionable peaking. Make the resistor too big and you'll run into DC drop
problems.
Be sure your spice model includes parasitics for the capacitor mounting inductance
and the inductor model incorporates the appropriate parasitics such as R and C as
well if you want a really accurate simulation.
Alternatively you could design a filter with unequal terminations, but I'd imagine
the resulting theoretical part values will be unrealizable in a real-world design.
Enjoy.....
-Ray Anderson
Sun Microsystems
>
> I have a question about inductor selection. I am designing a filter circuit
> to filter out noise from a power supply input to a PLL circuit. I want to
> filter out noise up to 1MHz. What should be the self resonance of such an
> inductor. The Q can be low, the application is broad band. Any help would
> be greatly appreciated.
>
> Thanks,
>
> Alex
**** To unsubscribe from si-list or si-list-digest: send e-mail to
[email protected]. In the BODY of message put: UNSUBSCRIBE
si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP.
si-list archives are accessible at http://www.qsl.net/wb6tpu
****
This archive was generated by hypermail 2b29 : Tue May 08 2001 - 14:30:00 PDT