In the first installment of Audio by Design, we identified the pin-1 problem that unfortunately resides in much audio equipment and I discussed why you don't want it. The pin-1 problem crops up when the shield of a signal interconnection cable is allowed to enter the piece of gear (via the panel connector and internal wiring) and ultimately be connected to the internal signal ground. When that happens, any noise flowing through the shield will flow through the audio circuitry. Various grounding and shielding schemes have been developed to circumvent the problem, but it is best tackled at the source.
The remedy for the pin-1 problem, in addition to proper circuit design and layout, is to tie all shields from interconnecting cables to the equipment chassis at point of entry. The equipment chassis should also be connected to a common ground. This happens (to greater or lesser degree) through normal three-prong power plugs with the ground connections still intact, or by rackmounting without isolation washers. Gear with "wallwart" external power supplies will usually need an external connection between the chassis and a good system ground.LOOKING FOR PROBLEMS
How can you find out if your audio equipment has the dreaded pin 1-problem?
First, do a visual inspection of the gear. Where does the shield on the input and output connectors go? Sometimes you can follow the wiring or trace from the connectors to the circuit board. Be wary of PC-mount connectors. They may be cost-effective solutions for efficient manufacturing, but by their very nature, they bring along pin-1 problems inside the box. Also, pay special attention to multipin connectors and audio connectors on computers. If you don't see the shields connected to the chassis at point of entry, then suspect a pin-1 problem.
If you've tried every good grounding and shielding known to television engineers and nothing seems to work, look for pin-1 problems in the gear itself. Ask manufacturers how they are addressing the pin-1 problem in equipment design and if anyone else has discovered a problem.TESTING FOR PROBLEMS
Pin-1 problems don't show themselves on the test bench with normal measurements for frequency response or total harmonic distortion-plus-noise. That's probably one reason they went undetected for so long.
For a piece of gear already installed and wired, a quick test can be made by removing all inputs and outputs from the device under test (DUT). Connect a wire from the shield of the cable connector to the shield pin on the corresponding connector on the DUT. Monitor the output with a meter or by a listening on a powered loudspeaker (more details on metering and monitoring below). If you detect hum or buzz, then it's a good bet that the piece of gear has a pin-1 problem.
John Windt developed a tester to detect pin-1 problems and presented his ideas at an AES workshop in 1994. His paper, "An Easily Implemented Procedure for Identifying Potential Electromagnetic Compatibility Problems in New Equipment and Existing Systems: The Hummer Test" was subsequently published in the Journal of the AES in June 1995.
"The Hummer" is the name Windt gave the tester that is used to inject a 60-Hz signal (about 50 mA) into the ground system of a device under test, between the shield pin on an input or output connector (pin-1 on an XLR) and the equipment chassis. An output on the DUT is monitored for any change in the noise level. An increase in the noise level with the Hummer signal applied means that the noise injected on the shield is getting into the signal output--a sure sign of a pin-1 problem.
Building a Hummer is a good do-it-yourself project. Start with a 110 V to a lower voltage (6 V to 24 V) transformer (like a wallwart). The current rating on the transformer should be 50 mA or more.
|Fig. 1 and 2 show two different circuits for a Hummer. Both of these are based on the concepts that Windt developed.|
| Fig. 1 |
| Fig. 2|
Fig. 1 and 2 show two different circuits for a Hummer. Both of these are based on the concepts that Windt developed.
Fig. 1 uses a normal electric 100 mA light bulb. Fig. 2 shows the wiring of an LED and diode in parallel (in one leg of the low-voltage transformer) and a current limiting resistor in the other leg. In Fig. 2, LED indicates current flow, the diode protects the LED, and the resistor limits the current when the two leads are shorted.
These versions deliver basically a 60 Hz sine wave current. Bill Whitlock, president of Jensen Transformers, noted that other variations use a diode in series in order to deliver a more harmonic-rich half-wave rectified current, which can be more audible.
Use an appropriate housing and connectors (alligator clips, banana plugs, spade lugs, or whatever type of connector that will allow easy connection to the equipment.)
Before hooking the Hummer to any piece of audio equipment, test the tester (like you would for any test equipment). You may not want to start by testing that $150,000 audio console. Try something simpler instead. It should go without saying, but I'll say it anyway--use common sense.USING THE HUMMER
First, disconnect all inputs and outputs to the device under test. Then, connect a meter and/or self-powered monitor loudspeaker to one of the audio outputs that you wish to measure. This would be pin-2 and pin-3 on an XLR connector; tip and ring on a balanced 1/4-inch phone connector; tip on an unbalanced 1/4-inch phone or RCA connector; and the appropriate signal connections on a multipin or terminal strip.
It's a good idea to listen to the output even if you are metering it. The monitor loudspeaker doesn't have to be fancy, but it will have to reproduce 60 Hz. The meter can be a voltmeter, oscilloscope or any kind of audio level meter, as long as it has the range to measure lower noise levels.
Power up the device under test. Monitor the output. Note the noise and its level. Is it white (hissing) noise? Normally, that's what would be expected.
Now you're ready for the Hummer. Plug in the 110 V side of the wallwart to a normal AC outlet. (We're talking about U.S. voltages here.)
Then connect one lead out of the Hummer to pin-1 of an input XLR connector, sleeve of a phone plug or RCA connector, or the shield connection of a multipin connector or terminal block. Connect the other lead of the Hummer output to the chassis of the equipment.
Listen and meter. Does the noise level increase? Can you hear a 60-Hz hum? If the answer is yes, the device under test has a pin-1 problem. If there is no increase in the noise level, then the DUT doesn't have a pin-1 problem, at least at audio frequencies.
Try different combinations of inputs and outputs. On a large console this will be a lot of work. Document your findings. Don't be surprised if different permutations of inputs and outputs result in different levels of noise.
Try the test by connecting the Hummer to pin-1 of both input and outputs, instead of one side to chassis.
Next time: What to do when you find a pin-1 problem.