Clipping Diodes

[carboncomp]

Hey guys,

Been reading some posts on audio forums that are confusing me regards clipping diodes, mainly the notion that one brand of germanium diode having audible differences to another brand of the same diode when used for clipping. for example in this circuit below. 



I was under the impression that they simply clipped the peeks of the AC signal and therefor had no way of imbibing the AC signal that didn't even travel though the diodes?

Ergo, with a function generator, oscilloscope and big bag of mixed brand didoes you could easily match up sets that would clip the wave form at the same point regardless of brand resulting in clipping sets that sound uniform.

Or am I missing something?

Thought I would ask here on the EEV forum, where "its because those brands have more MOJO" is a less exceptional answer. 

[EPAIII]

Diodes do not suddenly start conducting when the Voltage is raised. They will have a small range where conduction starts slowly and then quickly increases to a much larger value. The exact curve and the final Voltage value for full conduction may vary for different brands or for different diodes of the same brand. It will probably also change with temperature and perhaps other factors.

You want a simple curve tracing circuit where the current is plotted against the applied Voltage. Simple devices to do this were sold as scope attachments. They had a current limited AC transformer and a current sensing element (resistor?).

http://en.wikipedia.org/wiki/Semiconductor_curve_tracer

http://www.techlib.com/electronics/curvetrace.html

Many more, just search for "transistor curve tracer". These do not need to be expensive devices. And you can easily match transistors and diodes with them if you have a good eye. But you do need a scope to use them.

PS: In my experience the simple transistor curve tracers are the best way of testing diodes and transistors. They show many problems that would not be apparent with simple transistor testers or with Ohm meter checks.

[T3sl4co1l]

As far as quantities that vary, germanium diodes have low diode drop, but their series resistance varies (I'm guessing a 1N34 is necessarily a high resistance point-contact type, but others may be planar with much more ideal diode characteristics).  Reverse leakage may be relevant as well.  A good substitute might be a schottky diode (also has low Vf) with a series resistor.

As for "mojo", subtle differences in curves may produce noticeable effects.  The harmonic distortion and IMD, combined with a guitar's various waveforms, can do interesting things to the sound.  As far as some particular diode being some amazing hyped component, that's just silly; the differences are subtle at best, something a casual listener would never discriminate.

Tim

[G0HZU]

I was under the impression that they simply clipped the peeks of the AC signal and therefor had no way of imbibing the AC signal that didn't even travel though the diodes?

Ergo, with a function generator, oscilloscope and big bag of mixed brand didoes you could easily match up sets that would clip the wave form at the same point regardless of brand resulting in clipping sets that sound uniform.

Or am I missing something?

I'm not an audio geek but that circuit is a kind of limiter.

Depending of the 'hardness' of the limiting  you will find that smaller signals will tend to get suppressed more relative to bigger signals if the limiter has a very hard (as in sudden brick wall) action as the big signal enters the limiting region.

So if you had one big signal entering the hard limiter and also one much smaller one at a nearby frequency (eg at -20dB amplitude relative to the big one) then you would see the big signal would clip/limit to the limit of the diodes but the small signal could end up being -30dB relative to the big one at the limiter output.

So the smaller signal gets suppressed in the limiter relative to the big signal (if that makes sense)

So I guess the degree of hardness to the limiting will affect the relative levels of big and small signals after the limiting. So to a typical audiophile this opens up all kinds of interpretations as to what make s a good or bad limiter I guess?

[T3sl4co1l]

Assuming your "big signal" is, for example, some bass drums, which produce a lot of high amplitude, low frequency content, yes: when that signal is large, the diodes remain conducting and little else gets through.  However, the other signals still pass during the time when the large signal is crossing through zero.  This modulates the other signals, breaking them into sidebands.  This is called intermodulation distortion (IMD), and is particularly unpleasant on non-harmonic signals.  Music driven into clipping sounds terrible: harsh, overdriven, crunchy and farty.  A relatively simple signal, like a sine wave, has no other frequencies to interfere with, so the results aren't as disastrous, and can be beneficial (hence, guitar distortion effects).

A different type of circuit, which has the limiting effect without IMD, is called a compressor (or its inverse, an expander).

Tim