Re: [SI-LIST] : Awkward analog power filtering

About this list Date view Thread view Subject view Author view

From: Ritesh Kapahi (rkapahi@idt.com)
Date: Thu Aug 17 2000 - 17:23:43 PDT


Hi Chris
We tried experimenting with one of Gigabit SERDES by providing six
different power planes and finally found that the device is less noisier
with single power plane rathen than fighting the noise with partial planes.
I will agree with you by not using so many RCs and providing a solid plane.
Bye
Kaps

----- Original Message -----
From: Chris Bobek <cbobek@cadence.com>
To: Si-list <si-list@silab.eng.sun.com>
Sent: Thursday, August 17, 2000 4:23 PM
Subject: [SI-LIST] : Awkward analog power filtering

> Hi All,
>
> We are using a mixed-signal framer BGA for 4 ports of T-1.
> The vendor (who shall remain nameless) provided a reference
> design that shows a very unusual scheme for filtering power to
> their device.
>
> The device has about 20 digital power pins, which are to be
> directly connected to the power plane and decoupled using
> standard 0.1uF caps. The device also has about 20 analog
> power pins. These are divided into:
>
> - transmit power for each of the four ports
> - receive power for each of the four ports
> - misc power for each of the four ports
> - generic analog power
>
> The ref design shows an RC network for each group/port
> combination above, so that's 16 RC power nets! Each RC net is
> a 4.7ohm resistor, 47uF cap, and 0.1uF cap. This becomes an
> issue because the package is a BGA. That means I'll have to
> snake these power traces to each of the pins. I think I can
> make the traces pretty wide, but it won't be as nice as a
> plane.
>
> I proposed a "better" solution to them which is to have a
> single analog power plane under the BGA. Then, we'll have a
> few of the 47uF caps and a bunch of the 0.1's. The analog
> plane would be tied to the digital plane through a ~1ohm
> resistor (that can handle the current). This scheme is a lot
> more popular and I think it achieves a better power supply
> system (less L between pin and caps) and a lot less board
> space. They didn't "ok" this solution because they only
> tested their device per the recommendation. They can't defend
> the recommendation, so that's why I thought I'd ask everyone
> on the list if they've seen something this elaborate before
> (especially for something ~ 1.5MHz).
>
> Thanks,
>
> Chris
>
>
> **** To unsubscribe from si-list or si-list-digest: send e-mail to
> majordomo@silab.eng.sun.com. 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
> ****
>
>

**** To unsubscribe from si-list or si-list-digest: send e-mail to
majordomo@silab.eng.sun.com. 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
****


About this list Date view Thread view Subject view Author view

This archive was generated by hypermail 2b29 : Wed Nov 22 2000 - 10:51:05 PST