Page 169 - Troubleshooting Analog Circuits
P. 169

I56                                                  13. Letters to Bob


                      10.  Probes work into a certain, specified scope-input capacitance. You can’t always take
                         a probe that came with one scope and use it at high frequencies on another scope.

                      11.  A simple test technique is waving your hand over a circuit to feel for the hot spot.
                         If something has gone into a latch-up but is not smoking, you can frequently find it
                         this way.

                      12.  Edmund Scientific (Barrington, NJ) sells thermally sensitive liquid-crystal sheets,
                         which you can lay over a circuit to find moderate hot spots. This material works well
                         when you have a known-good PC board to compare with the circuit under test.
                      13.  Drafting departments sometimes erroneously think that they own the schematic and
                         that its only purpose is to serve as a wiring diagram for the PC-board layout. Long
                         after PC-board layout, the production-test, sustaining-engineering, and service de-
                         partments will still need the schematic. Drafting tends to lose notes that I place on the
                         schematic, such as filter poles and zeros, temperature coefficients, normal AC and
                         DC voltages, waveforms, and thermal information. I save myself a lot of calls by
                         putting this information in front of the techs from the beginning.
                      14.  You can make an extremely low-distortion (and slightly microphonic) sine-wave
                         oscillator from a light bulb and an op amp. I got the circuit from Linear Technology’s
                         (Milpitas, CA) AN 5 application note. I built a 3-frequency (400-, lOOO-, and 2800-
                         Hz) oscillator in a small metal Bud box. It had a THD lower than -80  dB.
                      15.  If a circuit’s DC values change when you breathe on it, you may have dirty circuit
                         boards.

                      16.  When testing high-gain, low-signal-level circuits, repeat the measurements with the
                         lights off. You may be surprised to learn that many components are photosensitive
                         and have infrared transparent bodies. One of my colleagues had a photosensitive
                         metal-can op amp that leaked light in around the leads.
                      17.  Protection diodes can rectify high-frequency noise and oscillations.
                      18.  Micro Technical Industries (Laguna Hills, CA) makes a handy thermal probe with
                         which you can individually heat components. The probe has tips to fit various compo-
                         nents, such as small and large resistors, metal-can op amps, and DIPS of various sizes.
                      19.  Some sample-and-hold circuits are sensitive to slew rates on the digital inputs.
                     20.  Even Schmitt triggers can exhibit metastability.
                     2 1.  As paraphrased from an Analog Devices (Norwood, MA) application note, “You
                         may be able to trust your mother, but you should never trust your ground.”
                     22.  Wrapped-wire circuits work pretty well if you can distribute power and ground prop-
                         erly. I use large-diameter bus wire in a rectangular grid for high-frequency logic if I
                         don’t have a wrapped-wire board with internal power and ground distribution.
                     23.  Sometimes powering your test circuit with batteries breaks ground loops and elimi-
                         nates power-line noise.

                     24.  A handy thing to have is a 60-Hz, passive, twin-T notch filter in a small Pomona box
                         with dual banana plugs for input and output.
                     25.  Another handy thing to have is a 20-dB high-impedance amplifier in another Pomona
                         box. The circuit in Figure 13.1 works at audio frequencies.
                     26.  The CMRR of an op amp is not a constant function of the common-mode voltage.
                         This inconsistency often dominates nonlinearity in noninverting circuits.
   164   165   166   167   168   169   170   171   172   173   174