At 10:16 AM 3/4/2005, Russell McMahon wrote:
>>By the way, there are practical cases when one of the capacitor leads 
>>cannot be as short as the other.
>>In such cases, it is better to make the Vdd one as short as possible, 
>>rather than the Vss, or it's the same?
>
>>>Unlikely to matter and unlikely to be especially significant if it did 
>>>matter.
>>>In both cases you have two series C's with a capacitor (imperfect) 
>>>between them. As you are not connecting to the midpoints you can't tell 
>>>what order the components are in.
>>
>>I disagree.. Done right, you create a "tee" filter using the lead 
>>inductances.
>>Done wrong, you add inductance in series with the cap on the gnd side.
>>
>>It's a couple dB difference at <1MHz
>
>I think we are answering different questions.
>I took the question at face value to mean that the CAPACITOR LEADS were 
>different lengths.
>You took the interpretation (and I see it is more likely to be the correct 
>one) that the track lengths to the IC are different. I agree that at that 
>stage you can start to build circuits that are different.


I talk about this on my web page.
Between the die and the cap, you're stuck with a certain minimum lead length.
But, you can re-distribute half of it.

But in all cases, the overall length needs to be minimized, and dumping the 
cap into a ground plane at the VCC pin may or may not accomplish that.  I 
would contend that it can be worse than the method I described.




-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist