Design Con 2015

Guru of Grounding

's profile
image
Principal Engineer

Bill Whitlock has designed analog electronics since 1972, specializing in high dynamic range professional audio equipment and systems. He is president and chief engineer of Jensen Transformers, Inc., Chatsworth, CA, www.jensen-transformers.com, and also does consulting work as time permits. His landmark paper on balanced interfaces appears in the June 1995 Audio Engineering Society Journal, now the most popular ever printed. Other writings include chapters for several books, Jensen white papers, and numerous magazine articles. He presents seminars at trade shows, professional meetings, and universities, including MIT in 2007. NSCA students voted him technical instructor of the year in 2009 and 2010. Bill is also active in standards work for AES, UL, and IEC. His four patents include the InGeniusĀ® balanced line receiver IC by THAT Corp. He's a Life Fellow of the AES and a Life Senior Member of the IEEE.


Guru of Grounding

's contributions
  • 04.22.2014
  • The G word: How to get your audio off the ground
  • Some years ago, the late Neil Muncy and I were invited lecturers at a Syn-Aud-Con "Hum and Buzz" event in Indiana. I did my myth-busting explanations of balanced interfaces and Neil explained the "Pin 1 problem" (a term he coined). Steve Macatee of Rane was in the audience and, after hearing how a designer can create such a problem, stood up and confessed that Rane had the problem designed-in to much of their product line. Thereafter, Rane quietly (pun intended) did new board layouts for those products. Of course, marketing told customers a story about "new technology" that required alterations to those PCBs. And, of course, lots of customer issues disappeared. I'm now picking up where Neil Muncy left off and include the pin 1 problem in my "How to Design a Robust Product for the Real World" tutorial ... most recently presented at the AES convention, Oct 2014 in LA. My latest area of interest is a similar grounding error that produces what I call the "power-line prima-donna syndrome" ... audio devices (such as powered loudspeakers) that make noise with no connections other than AC power ... if the AC power has the slightest bit of audio frequency noise on it (and it all does, 3-5% THD is typical for 60 Hz power). - Bill Whitlock, Whitlock Consulting, AES Life Fellow - IEEE Life Senior. BTW Bruno, I admire your work and writing style!
  • 10.13.2013
  • PCB design course & checklist
  • I'd add one item to your excellent list (it applies to analog or digital and across a wide frequency range): Be extra careful with I/O cable shield connections. There's a tendency among folks who layout the artwork to attach these at "convenient" points on the "ground network" of a board. Remember that shield connections are often the source of wideband, AC-power-related currents at the system level ... and the currents can reach tens or hundreds of mA. Always give these "alien currents" a separate and independent path to get where they want to go (either the chassis safety-ground connection or another I/O connector). In audio gear this issue is widely known as "the pin 1 problem" (pin 1 is the shield contact in audio XLR connectors) and can make equipment output(s) become noisy simply by connecting the shield contact of an I/O connector. Even in digital gear, where the "frequencies of interest" are far removed from 60 Hz, it can cause common-mode voltage range of many digital receiver ICs to be exceeded ... leading, of course, to "mysterious" field problems that are never seen "in the lab". Bill Whitlock, chief engineer, Jensen Transformers, Inc., AES Life Fellow - IEEE Life Senior
  • 10.06.2013
  • Analog Fundamentals: Amplifiers
  • I'm glad to see that delta-Rs is explicitly shown in Figure 2. Its effect is so often forgotten when instrumentation amplifiers are applied in the real world. The "pro audio" industry loves to tout impressive CMRR figures but, with rare exception, do the test with a perfectly balanced source. Common balanced outputs in audio gear can easily have 10 ohms or more of "imbalance" or delta-Rs. This can easily reduce a touted 90 dB CMRR figure to 50 dB or less. And they forget that resistors needed to DC bias the IA inputs seriously lower common-mode input impedances, which exacerbates the delta-Rs degradation. Anyway, a great introductory piece! - Bill Whitlock, chief engineer, Jensen Transformers, AES Life Fellow - IEEE Life Senior
  • 10.03.2013
  • Design Notes: Matched Resistor Networks for Precision Amplifier Applications
  • Referring to Fig 2, which some may see as a "balanced input stage" connect an XLR audio input to a differential-input ADC, I'd like to caution that the respectable 80+ dB of CMRR will rarely, if ever, occur in a real-world connection. The slightest imbalance in the common-mode source impedances of a real-world audio source will degrade the CMRR to very disappointing numbers. The only way to reduce this effect is to raise the CM input impedances of the diff-amp ... into the meg-ohm region. This is the reason CMRR of a good input transformer will remain 100+ dB in spite of source-impedance imbalances of hundreds of ohms. A bootstrap configuration accomplishes this in the InGenius(r) input stage made by THAT Corp. --- Bill Whitlock, Jensen Transformers, Inc.
  • 09.30.2013
  • Sounds.....in.....(3d)...space...
  • I hate to be a bit of a wet-towel here, but one of the great weaknesses of every multi-channel format I can think of is (drumroll) ... mono compatibility (or stereo compatibility at the very least). I'm sick to death of multi-channel mixes being "converted" to stereo (the format built into almost every TV, for example) and having the center channel all but disappear. So you turn up the volume so you can hear an voice at center stage and then along comes the theme music to blast you out of your seat! It's "in the mix" alright ... and really messed up. BTW, only one of the many "stereo enhancement" technologies offers true mono compatibility ... Spatializer(r) - once built into hundreds of Panasonic and other brand products. Recording and playing back "3D" sound via speakers for one "sweet spot" is relatively trivial ... but making it really work at multiple locations is essentially impossible. In the end, nothing comes close to reality! - Bill Whitlock
  • 09.16.2013
  • Texas Instruments inductance to digital converter (LDC): Necessity breeds invention
  • Actually, in a rather gross sense, metal composition can be determined. You may recall so-called "tuning wands". It was a plastic rod about the size of a thin pencil with a ferrite slug in one end and a brass slug in the other. If you were "aligning" say, a TV set's IF strip (when lots of coils were used rather than the SAW filters of today), you'd try sticking one and then the other end into the coil, which had an adjustable ferrite slug to adjust its inductance. Inserting the ferrite end into the coil would increase its inductance and the brass end would decrease it (acting like a shorted turn). So you could tell if adjusting that coil would make the "tuning" change you wanted. I suspect this IC and a coil could easily tell the difference between a steel "target" and a brass, aluminum, or copper one. In any case, this is a well-known "old-school" technique ... but it seems practical aspects of physics are rarely taught in engineering schools these days ...
  • 09.13.2013
  • Can You Hear the Difference?
  • What ever happened to the Federal Trade Commission anyway? I never saw this kind of BS back in the 50s or 60s. The only item I've heard of being taken off the market in recent decades was the "Rid-X" an ultrasonic generator claimed to rid your entire house of rats, roaches, and other pests. Marketing now seems to be a totally unregulated "wild west".
  • 09.16.2013
  • Texas Instruments inductance to digital converter (LDC): Necessity breeds invention
  • Granted, it will be nearly a no-brainer to use and so will put the technology in more stuff ... both good things. I just object to it being sold as a "new sensor technology". It's really an integration of the support electronics for the old technology of inductive sensing. I remember the old Signetics/Intersil LVDT support chips that had a nice analog output (infinite resolution), which was great because all that was required was an external comparator (analog) to make a precision limit detector. After all, everything doesn't need to be digital. In fact, it would be great if this new IC had an analog output, too!