This is standard telephone company stuff, however, it would seem that several radio engineers have forgotten this. I was reading on one forum where an AM station was using 1000 feet of 12 gauge romex to send audio from the studio to the transmitter out back. The owner was complaining that the audio sounded bad.
Longer wire runs need to be terminated with the characteristic impedance of the cable being used, normally 110 ohms or so for typical audio wire. This is because impedance mismatches can cause return loss just like in an RF circuit. Exactly what the effect of the mismatched impedance depends on the length and frequencies involved. On shorter cable runs of less than 100 feet or so, this usually is not an issue.
The result of return loss is part of the audio energy gets reflected back to is origin (a standing wave), where it mixes with newer audio. This can cause out of phase issues and usually the result is high tinny sounding audio with distortion in the mid range frequencies. In other words, it ain’t pretty. This can really become an issue with digital audio because of the higher bandwidth requirements for high sample rates. It has always struck me as odd that AES/EBU audio uses XLR type connectors. An XLR connector does not maintain the characteristic 110 ohm impedance of most digital cable and itself can cause pretty significant return loss. But anyway…
There are a number of options for proper termination:
1. Transformers are often used to match the impedances of circuits. A transformer converts alternating current at one voltage to the same waveform at another voltage. The power input to the transformer and output from the transformer is the same (except for losses). The side with the lower voltage is at low impedance, because this has the lower number of turns, and the side with the higher voltage is at a higher impedance as it has more turns in its coil. Western Electric 111C audio transformers were often used in equalized TELCO circuits sending audio over long distances on copper pairs.
2. Resistive network impedance matches such as H or T or L pads are the simplest to implement. They limit the power deliberately, and are used to transfer low-power signals, such as unamplified audio or radio frequency signals. Almost all digital circuits use resistive impedance matching which is usually built into the structure of the switching element.
3. Active balanced converters using opamps with high input impedances (10 Kohm bridging resistance) that first greatly reduce the voltage, then amplify it are often used an audio circuits. They have the advantage of active gain control and are often used in conjunction with gain reduction and limiting circuits.
The above diagram shows an active unbalanced to balanced audio converter. The advantages of such a circuit are active gain controls can be added to set levels. With additional feedback circuit elements, it can also be used for automatic gain control, gain reduction, limiting and so forth.
For most inter and intra studio wiring, professional audio equipment is designed for 0 dBm 600 ohm balanced audio (AKA line level audio). Audio cable such as Belden 8451 or like multi-pair cables terminated on punch blocks or connectors works well. Cable impedances and matching are generally not a design consideration. Long cable runs, longer than 150 feet or so, do need to be terminated in a high quality audio installation.
9 thoughts on “Proper termination of long audio wire runs”
The Western Electric “Repeat Coils” were some of the best transformers ever constructed. The UTC “Linear Standard” and Langevin were also top shelf neighbors of Western Electric. It is really a pity that none of these high quality products are being manufactured today. “Active Balanced” was a cost reduction move in the ’80’s that I never thought much of. Transformer coupled audio (analog) is the most listenable to me and stations running transformer coupled low-level audio to high-level plate-modulated tube transmitters still sound the best. The new stuff just doesn’t seem to sound right in my mind. Another good friend of mine just prior to installing a 3DX50 Destiny at my old employed station had to run the RCA BTA-10U for a couple of days while their Nautel was being fixed, and guess what? The station staff remarked how good the station suddenly sounded! So, PDM versus old conventional plate modulation with triodes seemed to win the A-B comparison.
Those 111c units weight about 10 pounds each. They are worth quite a bit of money on Ebay these days. Shame of it is, about 10 years ago I was cleaning out a storage unit for a radio station that we were moving and I threw about five of these things away. I just didn’t have room to keep them anywhere.
Regarding old transmitters, the old GE BTA25 was the best sounding AM transmitter I have ever encountered. Unfortunately it was full of PCB’s and had to be removed during one of the refinance phases. It was also not the most efficient unit, electrically speaking. Anyway, the PDM units these days don’t sound bad, they just don’t sound all that good.
I have a bunch of the WE coils from several moves that I assisted with and a few from the WZEE property (studios and tower) that I recently purchased. The BT-25 at WCKY (1530) also was one of my favorite dial spots while I was a BC Band DXer growing up. The true, rich, transmission quality of that transmitter and some others, I believe will never be duplicated. The RCA BTA50-F1 was another, as well as the Westinghouse 50HG series.
Equivalents, in many ways superior in performance, to those excellent WE transformers are available off-the-shelf from Jensen. See data sheets at
http://www.jensen-transformers.com/datashts/11ssp8ma.pdf, http://www.jensen-transformers.com/datashts/11ssp7mp.pdf, and
I disagree about the length of a cable that requires termination. It’s easy to calculate that the wavelength in a cable (propagation velocity of about 0.8) is about 40,000 feet at 20 kHz. Only the very slightest effects of reflections will be observable at 5% of this length – and many engineers use 10% as a rule of thumb. Doing the math indicates we should worry about termination only when cables are some 2,000 to 4,000 feet long or longer.
Bill Whitlock, president & chief engineer
Jensen Transformers, Inc.
Bill, In the field, I have found that terminating long wire runs does make a significant difference in the sound quality. One instance was at a sports stadium, the run was almost 1000 feet of category 5 cable being used for balanced, line level audio. The end product was high and thin sounding. I used a pair of 110 coils, one on either side of the wire run. The end result was much better.
On your 1000′ line, it wasn’t just the termination that improved frequency response. More importantly, the transformer lowered the source impedance presented to the line and decreased the load seen by the driving stage. The line capacitance can easily put an op-amp into current limiting. I have had good luck driving a few thousand feet of phone pairs with a 600:150 transformer as the source, and a 600:600 on the other end. Flat 20 to 20k and no noise.
I have a few of the 111c Can someone tell me how to test them with a meter?
If I test between pin 1-2 I get a reading of .0175 meter set at 2k ohms.
Same with pin 3-4 5-6 7-8 is there any further testing to make sure they are ok?
Check pin 1,3,5,7 to pin 6 and case, it should be open. If you have an audio generator/distortion meter, sweep from 25 Hz to 20 KHz; the results should be relatively flat and low (.5 to 2%) distortion.
On short runs I’m willing to use differential amplifier outputs and inputs; right or wrong this has become a way of life. However, depending on many factors affecting long runs I like to use a good quality transformer, and possibly a 3 dB pad on both ends of the line side of each transformer. On differential circuits the common mode may be fine at normal audio frequencies, but at higher frequencies Common Mode Rejection can quickly fall apart allowing RF and other wide band noise to enter a system.