Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

pad-to-pad capacitance

54 views
Skip to first unread message

sea moss

unread,
Aug 3, 2022, 2:03:53 PM8/3/22
to
Two part question (my google skills have failed me for both):

1. Does anyone know approximate pad-to-pad capacitance for the common SMD packages (0201, 0402, 0603, 0805)? Ever measured it? I can't use the parallel plate capacitor formula, since the plates are adjacent, not parallel. (assume there is no copper below the pads)

2. Given an SMD footprint as above, how close does a ground plane need to be (on the same layer) to decrease the effective pad-to-pad capacitance? Is there a formula which gives a good approximation?

whit3rd

unread,
Aug 3, 2022, 5:54:42 PM8/3/22
to
On Wednesday, August 3, 2022 at 11:03:53 AM UTC-7, sea moss wrote:
> Two part question (my google skills have failed me for both):
>
> 1. Does anyone know approximate pad-to-pad capacitance for the common SMD packages (0201, 0402, 0603, 0805)? Ever measured it? I can't use the parallel plate capacitor formula, since the plates are adjacent, not parallel. (assume there is no copper below the pads)

But, regardless of the pads, the component's mating metal surface has that
capacitance regardless, doesn't it? You don't need always to make the 'plates' larger
than that component feature.
>
> 2. Given an SMD footprint as above, how close does a ground plane need to be (on the same layer) to decrease the effective pad-to-pad capacitance? Is there a formula which gives a good approximation?

The use of guard rings can make such capacitance negligible, if that's required, but...
trying to simply treat the tiny elements as separable capacitances, isn't a good and useful
model.

John Larkin

unread,
Aug 3, 2022, 6:14:09 PM8/3/22
to
On Wed, 3 Aug 2022 14:54:38 -0700 (PDT), whit3rd <whi...@gmail.com>
wrote:

>On Wednesday, August 3, 2022 at 11:03:53 AM UTC-7, sea moss wrote:
>> Two part question (my google skills have failed me for both):
>>
>> 1. Does anyone know approximate pad-to-pad capacitance for the common SMD packages (0201, 0402, 0603, 0805)? Ever measured it? I can't use the parallel plate capacitor formula, since the plates are adjacent, not parallel. (assume there is no copper below the pads)
>
>But, regardless of the pads, the component's mating metal surface has that
>capacitance regardless, doesn't it? You don't need always to make the 'plates' larger
>than that component feature.

Most surface-mount part capacitance is in the FR4, not in air.

>>
>> 2. Given an SMD footprint as above, how close does a ground plane need to be (on the same layer) to decrease the effective pad-to-pad capacitance? Is there a formula which gives a good approximation?
>
>The use of guard rings can make such capacitance negligible, if that's required, but...
>trying to simply treat the tiny elements as separable capacitances, isn't a good and useful
>model.

It's often important to know the three capacitances: each pad to the
ground plane, and 3-terminal equivalent capacitance pad to pad.

Klaus Vestergaard Kragelund

unread,
Aug 3, 2022, 8:24:55 PM8/3/22
to
turn PCB has a toolkit for a number of different calculations on PCB
structures:

https://saturnpcb.com/saturn-pcb-toolkit/

sea moss

unread,
Aug 3, 2022, 9:00:35 PM8/3/22
to
Unfortunately Saturn PCB does not cover this... I tried there first.

jla...@highlandsniptechnology.com

unread,
Aug 4, 2022, 10:09:10 AM8/4/22
to
It's easy to measure attofarads with cheap instruments that a decent
home lab should normally have.

Jeroen Belleman

unread,
Aug 4, 2022, 12:35:53 PM8/4/22
to
Femtofarads, sure, but attofarads? Please explain?

Jeroen Belleman

John Larkin

unread,
Aug 4, 2022, 1:15:36 PM8/4/22
to
Femtos, just make a voltage divider from a sig gen, Cx, and the usual
15 pF input of a scope. Signal average a bit or FFT at the low end. A
few fF resolution is easy. To get to aF, add a jfet or an opamp.

One could also add an inductor to resonate out most of that 15 pF.

Lots of ways to do this.

John Larkin

unread,
Aug 4, 2022, 1:19:32 PM8/4/22
to
The old HP selective (tuned) voltmeters were most cool for things like
this.

0 new messages