Voltage Peaks and Dips in Ground/Power Plane

Hello Chandu,

1) What are these fluctuations in your ground-plane relative to? Could these be artifacts due to your measurement setup? What are you using for a probe reference?

2) Are the power and ground planes adequately bypassed? That is, does the board sport plenty of bypass capacitors and in all the right places?

Attila

ya well , i use 500MHz DSO from TEKTRONIX and best suited probe for the same.. so i don't think its wrong with setup.

In case of bypass capacitors yes i do use as many as required at right place.

I would like to mention 1 thing for ex. Recently i verified with one board with deals with high frequency modulations around 85MHz. In that board to i observed the same frequency and amplitude of the dips and peaks.

My fir and basic st comment is: if it works, don't fix it...

But some background in case you are still worried:
If there is current in the board or in the power leads, there is no way you can have zero Voltage between every ground point in the system. The level of the differences will depend on the resistance and inductance of the section you are measuring, and the frequency and current levels in the signals. The best you can do is to ensure that, wherever a particular signal is being amplified or detected it is referenced to a local ground whose movement correlates with the signal. I assume this is what the board designer has already done...
(I still find it useful to remind myself that "there is no such thing as a global ground").

Now, if there is a problem with the operation...
If ground movement causes you to have problems with trouble-shooting, it could mean that you need to use high-impedance differential probes connected to the signal and the ground - both must be close to the point in the circuit where the signal is sensed.

If your concern is with the power-harness, and the objective is to prevent your assembly of known-working-boards from experiencing problems, the best first-step is to connect only via the intended grounding points on the boards, and to keep the grounding interconnects between daughter-boards as short and low impedance as possible. In some cases you may require screens between the boards - but this is rather unusual these days, and should be obvious from the application. If having taken these sensible precautions you still have functional problems, you probably need on-the-spot expert advice.

BTW, your description of what you observe looks as if you are looking at a digital system. These can accept moderate levels of noise between signal and sensor (10% of the total signal is not usually regarded as significant). If, however, you look with the sort of probe that comes with cheaper digital scopes you will often seee far more than this. The part caused by pick-up in the probe sensor is not an issue for the circuit; however, signals that are imparted to the system via the extra (not-as-designed) "ground" path can (rarely) cause malfunction.

How are you measuring the noise: does your ground lead look like this:-

OR like this:-

If it's more like the first picture then the noise you are seeing is probably not real. You can confirm this by clipping the croc clip of the ground lead to the end of the probe then using the ground loop to "sniff" around the PCB for emitted signals without touching it.

Look at Howard Johnson's brilliant site for high speed digital design.

He describes how to make a basic 1K resistive probe which is much better for measuring noise in ground planes.