Intermodulation Distortion?

[Daver]

What is usually the biggest cause of this? That kind of sub harmonic "ghost" note that occur on some amps and pedals for that matter. I'm not talking about speaker induced distortion here. Is it mostly a function of the power supply? Thanks.
Dave

[Steve Conner]

Intermodulation products are generated any time you play two notes that aren't harmonically related through a device that distorts the signal. IOW, any chord made of intervals other than octaves will have audible IM products added to it. Possibly the most audible intermod product is the difference frequency.

The effect of this in a musical context depends on the interval. Of intervals other than octaves, fifths have the nicest sounding intermod products and minor seconds (ie one semitone) the worst: the difference frequency gives a nasty throbbing sound.

So I say the biggest cause of that ghost note effect is the chord voicings you choose. It's probably no accident that heavy metal evolved towards two-finger power chords that sound an interval of a fifth, as the gain of amps increased.

If you hear ghost notes when playing a single note line, it's because hum from the AC line is getting in there somehow, and introducing another "note" of 60 or 120Hz, which intermodulates your guitar signal. Neither of these frequencies is an actual musical note, so it sounds bad no matter what note you are playing.

A push-pull amp can modulate your guitar signal at 120Hz without the 120Hz being audible when the amp is idle.

The cure for this hum modulation would be to replace any bad capacitors, or use bigger capacitors or add a choke. However, some people seem to think that a little 120Hz intermod improves the tone.

[Daver]

Steve,
Yes, I'm talking about playing two notes together or bending one as the other stays at the same pitch. It seems some amps do this more than others. So I guess I'm trying to figure out how to at least minimize the effect. Sometimes it bothers me and sometimes it doesn't. I drives a couple of my friends bat shit. It's worse, of course, when you're listening for it. Thanks!
Dave

[Steve Conner]

Oh, well I guess you could use less gain. If you cut the gain to half what it was, IMD should get 4 times less.

Or try messing with the tone controls. Maybe turning the bass down would help, if it's the low frequency dissonances that bother folks.

Or if you're feeling real adventurous, you could always try changing the music so you don't use so many dissonances. Having said that, I have heard some metal guitarists play dissonant intervals with high gain and make it sound awesome. They tend to just use it as a special effect in solos and the like, though.

[Daver]

Hmmmm, sounds like you're telling me there is no way to get around it. It just seems I'm more aware of it the last year or so. The worst offender, not surprisingly, is the Vox style amp I'm building now. Other amps I've built or own have varying degrees of it. I've got a couple of pedals that do it too. Maybe it's just me! Thanks Steve.
Dave

[bob p]

A push-pull amp can modulate your guitar signal at 120Hz without the 120Hz being audible when the amp is idle.

The cure for this hum modulation would be to replace any bad capacitors, or use bigger capacitors or add a choke.

As it turns out, I have an old RCA PA amp on the back burner that has exactly this problem. It has a huge 60 Hz signal that enters the system at the output stage cathodes (presumably coming from the adjacent pin heaters). Normally that kind of AC signal would get shunted to ground by the cathode bypass cap, but because the old cathode bypass cap has opened, there's no AC attenuation.

The result is that the AC voltage at the cathode shows up as a HUGE AC voltage on the plate, and it overwhelms the grid signal.

This is one of those problems that really made me glad that I bought an oscilloscope. With the scope, its easy to see exactly what's happening: there is a HUGE 60 Hz AC signal that doesn't belong on the plates. It doesn't show up on any of the preceeding stages, and in the power stage its only present on the cathodes and plates, and not the grids. The smaller amplitude 1KHz test signal rides on top of the big 60 Hz output. The test signal is so small riding on top of the 60 Hz signal that the test signal looks like noise!

By varying the frequency of the signal generator, its very easy to see the very ugly intermodulation products of the two signals. There are many frequencies where the two signals don't effect each other much, and it just looks like one waveform rides on top of the other as its frequency changes. When you hit specific frequencies though, the interaction of the two signals can be profound, and all hell breaks loose -- it can produce some strikingly aberrant waveforms. This is one of those examples where a picture is worth a thousand words, and motion pictures are worth a million words.

Sorry that I don't have pictures to offer -- I just thought I'd pass this along in case its at all helpful.

[Arthur B.]

Hmmmm, sounds like you're telling me there is no way to get around it. It just seems I'm more aware of it the last year or so.

That's basically the idea. The usual way of dealing with it is to limit the signal's amplitude and/or frequency.