Thanks to all the vendors who participated in my recent seminar in San Jose, CA. Starting at 5:00 pm after the first day's lecture everyone enjoyed some great live demonstrations from Sigrity, TDA Systems, Ansoft, Cadence, Innoveda, Mentor Graphics, and Wavecrest. My public classes include after-class demonstrations from many of these vendors. The next ones are scheduled for Dallas on April 18th and Boston on May 9th. The demo session is a great time in a casual setting to speak with applications engineers who really know the ins and outs of their products.
I have a busy schedule of private classes coming up in the next few months. Attention Texas Instruments employees: If you are interested in attending my seminar at TI in Dallas on May 13-14, contact Ted Moody at 972-917-7467.
I hope to have the opportunity to meet many of you on my travels this spring.
Benefits of Resistive Probe
Silence Dogood writes,
Please explain why I should consider using a
resistive-input probe. My FET probe seems perfect.
It has a very low input capacitance, high
bandwidth, and hardly requires a ground wire at
Thanks for your interest in High-Speed Digital Design. Here are ten good reasons to consider using a resistive-input probe.
If you don't know what a resistive probe is, check out: Probing High-Speed Digital Designs, Electronic Design March 17, 1997.
Please note that I'm not suggesting the resistive probe is always best. What I recommend is that every lab have at least one resistive probe available. That way, if you suspect your FET probe is limiting the clarity of your view, or excessively loading your signals, you can try the resistive-input type to see if it is any better.
- Low input capacitance. A good resistive-input probe has a C[in] of about 0.1 to 0.2 pF, as compared to a typical FET probe with a C[in] of 0.5 to 1 pF. The lowered input capacitance reduces the amount of timing skew introduced when you touch the resistive probe onto an active signal. This benefit is counterbalanced by the resistive DC loading which decreases the signal amplitude of a 50-ohm signal on your board to only 95 percent of its true level.
- High resonant frequency. The lower the input capacitance of the probe, the higher the frequency of the first probe resonance. This resonance occurs at the frequency f=1/(2*pi*square_root(L*C)), where C is the input capacitance of the probe, L is the inductance of the ground connection, and f appears in units of Hertz. Assuming C=1 pF (typical for an active FET probe) and L=10 nH (a very short ground connection) you get f=1.6 GHz. If any components of your signal hit this frequency the response goes whacko. With a resistive-input probe the capacitance is about five times less, so the resonant frequency more than doubles.
- ESD hardness. Resistive-input probes are not susceptible to ESD failure (although your scope might be so you should always wear a wrist strap anyway). Repeated exposure to ESD can degrade an FET probe. I've seen more than one partially- functional FET probe that delivers distorted and confusing results. It's frustrating to waste part of your day making measurements with something that is not working up-to-spec.
- Tolerance of weak ground connection. Your probe shows you a true signal as long as the impedance of its ground connection is a lot less than its input impedance. When you use a weak ground connection, part of your signal voltage is lost across the ground connection, and only the remaining part excites the probe. Because resistive-input probes have a higher input impedance in the GHz range than FET probes, they are more tolerant of weak ground connections (see "Mysterious Ground").
- Low price. There's no active parts involved, so scope manufacturers are embarrassed to charge more for resistive probes than FET probes, even through the resistive ones often work better in very high- frequency digital applications.
- High bandwidth. The highest-bandwidth probe I've seen (Tektronix, 20-GHz) is a resistive-input probe.
- Simplicity. There's something satisfying about using a simple, easily understood technology, and there's less to go wrong, too.
- Easy to make. You can make a resistive-input probe yourself. See the instructions in my book, "High-Speed Digital Design", or for a higher- bandwidth version of the same thing (and only slightly more difficult to make) see Doug Smith's write-up at www.emcesd.com.
- Persistence in a laboratory environment. A resistive probe can sit on your bench for a week and nobody will steal it. The half-life of a FET probe, by comparison, is about 15 minutes. This is one of the advantages of using unusual or rare equipment. I know one engineer who pasted a big "Peavey" logo on the front of his favorite high-performance oscilloscope. He claims nobody touched it for months. Peavey, if you don't recognize it, is a famous maker of rock-and-roll amplifiers.
- Tradition. Electrical engineers have had a long history of using resistive attenuation to increase the input impedance of voltage-measurement devices. This tradition dates all the way back to Maxwell (see also "Electrical Measurement", F.A. Laws, McGraw-Hill, 1917).
Dr. Howard Johnson