Passive intermodulation distortion in high-end audio?

A week or so ago I looked up “Intermodulation Distortion” to refresh my memory on exactly what it was. A statement in the Wikipedia article caught my eye: “Passive intermodulation (PIM), however, occurs in passive devices (which may include cables, antennas etc.) that are subjected to two or more high power tones”. The reference for this statement was to a paper by P.L Lui called “Passive intermodulation interference in communication systems”. That paper describes some of the mechanisms (and remediations) for PIM. Although Liu’s focus seemed to be for radio and “high power” devices, it seems that it would theoretically be possible that PIM exists in high-end audio. In the last few days, I searched further and found another paper on the Amphenol web site which confirmed Liu’s paper. The more I dug, the more I began to wonder if PIM could be affecting high end audio, especially since music is a highly intermodulated signal, its quality diminished due to distortion of any kind.

I now understand the causes of PIM to be roughly as follows: dirt and other impurities on metal surfaces (presumably in connectors); hysteresis due to ferromagnetic materials in/or near the signal path (the Amphenol paper specifies PIM increases even for gold plated nickel); micro-discharge between microcracks and across voids in metal (presumably in connectors); large current densities at contact between metal surfaces (connectors); and thermal/vibration variations which may increase or decrease the contact between metal surfaces (in connectors). The basic theme seems to be non-linearities introduced by the passive connectors and cables which in turn diminish signal quality (by introducing PIM). Alternatively, the PIM products leak into other components causing problems there.

If PIM (caused by cables) is a source of distortion in audio, then it’s unclear if there is any true remediation other than trial and error. For example, the amp manufacturer measures IM in a lab from input to output. Cable manufactures build their cables to the highest quality their price point allows But PIM occurs due to the interaction of the cable connector with the amp receptacle. The interaction of plug/receptacle in a specific instance could cause PIM due to micron size differences (both well within specs) in either the plug or receptacle. If ever there were an opportunity for finger pointing, this would be a good one. Both audio cable manufacturers and components seem to be primarily concerned with RF. [Speaking of RF, Ted may be dealing more with PIM products due to the megabit signal paths in the DAC and significant PIM caused sidebands.] The point is, I’ve seen no references to PIM, either in marketing materials, the PS Audio forums, or in any audio literature I’ve read so far. Is it something that everyone knows about but nobody talks about?

So I was wondering:

  • has anyone ever tested the IM distortion of a system by putting IM test-tones in a preamp and measuring IM distortion of the power amplifier output – first with a good cable between pre/power amp, and then with a cable known to sound worse?. Presumably changes in IM distortion, all other components being equal, would implicate the cable (or plug/receptacle combination) that, according to the research, are likely the major contributor to PIM.

  • Additionally, the idea of thermal variation causing PIM may be worth investigating. Has anyone ever tested how IM distortion changes as an amp breaks in? Has anyone compared the ambient room temperature where the device is breaking in to the length of the break-in time for good sound?

  • Could a fingerprint on an RCA connector introduce sufficient non-linearities to cause PIM?

Although I originally rolled my eyes at the thought of one cable “sounding better” than another, it seems there is sufficient research to confirm it is possible. RF noise from nearby electronic devices or any of a number of other causes could contribute more distortion than PIM so perhaps it can be ignored. The whole idea of PIM existing in audio is speculation on my part. However, I’m wondering “if it walks like a duck and sounds like a duck, is it a duck?”


P.L Lui: