Long Vcc traces

Not a good idea. Vcc lines are noisy enough without throwing in a series diode.

If you really wanted to minimise cross-talk between various lumps of logic, it can make sense to break the Vcc trace with a ferrite bead or chip. You need to find out the inductance of the chunk of ferrite and and calculate the resonance frequeny this gives with your by-pass capacitor; ferrite beads are made to be lossy at high frequencies, but they aren't always lossy enough at the frequency where they resonate with the by-pass capacitor, and then you.ve got to put damping resistor in parallel withthe bead.

This is rarely done in logic circuits, but can be essential in some analog circuits.

-- Bll Sloman, Nijmegen

The point of bypass caps is to provide a high frequency path for noise escaping the chip, and a source of ready power for fast switching transients. Part of the *need* for these is due to the inductance/resistance of the circuits leading up to the chips. If you put the bypass caps far away from the chips, they can't do their jobs as effectively. Plus, you'll be tempted to use a single bigger capacitor, which may not have the fast response characteristics you need.

As an analogy, think of the Vcc traces as a garden hose. You can only get so much water out of a garden hose, so if you want to dump a bunch on your friend's head, you have to fill a bucket and dump the bucket. The bucket, obviously, has to be on *your* end of the hose - having a bucket up near the house isn't going to help (my house has such a bucket, so that the well pump doesn't have to run all the time). And if you're having a water fight, everyone needs their own bucket.

So, since the goal is to have something near each chip, put something near each chip.

Note that on some of my designs, I *also* add a larger cap here and there to provide a slower response for the bypass caps. These act more like bulk capacitors (ripple filtering) than bypass caps though.

1) Bad idea for the caps. Even you think you are using a chip at a low frequency, say 10KHz, the chip still swicthes in nanoseconds, so the added inductance of the leads will have an effect. 2) If you're thinking of using a diode for power fluctuations, you have other problems that need adressing.

Actually, it's more to do with overcoming inductance/resistance in long traces. The bypass cap at the device ensures the chip has a localised low inductance/resistance power source. Very important for fast switching devices for example.

Bypass caps are need *at the device*. The further away you put them the less effective they are.

Forget the diode, a flawed idea for many reasons.

You haven't told us what the IC's are and the intended application etc, so it's hard to help further with actual specifics. Usually bypass caps at each device are enough for most logic circuits, but in many cases local regulation at the load (each stub) is used, for example with sensitive analog devices. Inductive filters can be used too, but that's a bit more complex.

Dave.

--
================================================
Check out my Electronics Engineering Video Blog & Podcast:
http://www.alternatezone.com/eevblog/

At first i thought you were talking about transmission lines / microwave splitters, etc... If there are any fast transitions, then one approaches that scenario...