Doug's picture
High Frequency Measurements Web Page
Douglas C. Smith

 Address:  P. O. Box 1457, Los Gatos, CA 95031
 TEL:      800-323-3956/408-356-4186
 FAX:      408-358-3799
 Mobile:   408-858-4528

Technical Tidbit - November 2006
Measuring Signals in the Presence of Severe EMI - Part 1, How Not to Do It
(differential measurement by channel subtraction)

two probes connected to metal plane

Figure 1.
Two Passive Probes Connect to Metal Plane

Abstract: A popular engineering troubleshooting technique is to measure power to ground noise on a board or in a system, often with a differential measurement using two unbalanced probes and channel subtraction in the scope. Such measurements at today's speeds are generally unreliable. Data is presented to illustrate limitations of this type of measurement.

Discussion: Figure 1 shows two unbalanced probes shorted by their ground leads and connected to a metal plane. The metal plane simulates the ground plane of a multilayer printed wiring board. I call shorting the probes like this a "null" experiment since the result should be zero. To the extent it is not zero, the result is the inherent error or noise level of the measurement setup. It makes no sense to try and measure signals smaller than the result of the null experiment. Figure 2 shows a close-up of the probe connections. The probe cables were allowed to assume random positions on the table between the metal plane and the oscilloscope.

Close-up of probe connections from Figure 1

Figure 2. Close-up of Probes From Figure 1

To simulate environmental noise, an ESD simulator was discharged to its own ground lead about 1 1/2 meters from the probes and plate shown in Figure 1.  The result is shown in Figure 3. Note that the peak value recorded on both channels was greater than the 4 Volt limit on the scope screen. Perhaps more important that the absolute magnitude of the signals recorded in Figure 3 is the fact that the two probes are ringing at different frequencies due to the probe cables being separated on the table and taking different paths to the scope. Taking the difference in the scope of the two channels shown in Figure 3 would not make much sense. In fact, as the top trace touches - 4 Volts, the bottom trace is just touching +4 Volts. The difference would be at least 8 volts! This is larger than either signal by itself.

Scope plot with ESD noise present
Figure 3. Scope Plot of Probe Outputs with Nearby ESD

The probe cables were then twisted together, about one twist per six inches along the cables. Figure 4 shows the resulting waveforms.  The two channels are at least ringing at the same frequencies now and look similar. Maybe a subtraction of the two traces makes sense to reduce the common mode interference from the ESD event.

Same plot as Figure 3 with probe cables twisted
Figure 4. Scope Plot of Probe Outputs with Nearby ESD With Probe Cables Twisted

Upon closer inspection of Figure 4 though, significant differences are present. Look at the two traces around one division from the left side of the screen, at the first positive peak of the waveforms. The common mode rejection achieved by subtracting the two channels in the scope is compromised and will still yield an apparent differential voltage that is significant in that region of the scope traces. In general, one is lucky to achieve 10 dB of common mode rejection using separate probes and channel subtraction in the scope above a few tens of megahertz.

A better method is needed. One could try to float the ground leads and connect them together, but that will increase the apparent common mode signal requiring more attenuation of the common mode signal. Also the large loop composed of the ground leads and probe tips will pick up significant noise.

The December 2006 Technical Tidbit will propose a solution to the problem of making differential measurements in the presence of significant interference.

Summary: The use of two separate probes coupled with channel subtraction generally does not work well for differential measurements above a few tens of megahertz in the presence of significant interference. Null experiments should always be done to prove the results of a measurement around strong interference such as ESD. The null experiment results presented show the limitations of using two separate probes to make differential measurements.

Other articles on this website related to this topic are:
  1. July 1999: The Shorted Scope Probe Problem
  2. June 2000 supplement, Ground Lead, Friend or Foe?
  3. September 2001, Improving FET Probe Immunity to Unwanted Noise Pickup
  4. March 2003, Minimizing Errors in Oscilloscope Measurements
If you like the information in this article and others on this website, much more information is available in my courses. Click here to see a listing of upcoming courses on design, measurement, and troubleshooting of chips, circuits, and systems.

Click here for a description of my latest seminar to be available in the Fall titled:

EMC Lab Techniques for Designers
(How to find EMC problems and have some confidence your system will pass EMC testing while it is still in your lab).

D-104 mike
Check out my podcast containing mp3 format short tutorials, tech news and more! Just click on the microphone to see the listing of free audio programs. Content is added every week  on technical topics so check back frequently.
Top of page

Questions or suggestions? Contact me at
Copyright © 2006 Douglas C. Smith