Audio frequency harmonics poser.

[G6ONEDave]

Given that generally with audio, valves produce even harmonics and that semiconductors produce odd harmonics, what happens in the following scenario?

Guitar amplifier has a solid state input that goes to some design of a silicon effects synthesizer chip, then the audio signal goes to a double triode (ECC83 or equiv) and finally goes to a solid state audio amplifier output stage. So what happens with the harmonics that get to the loudspeaker? It's probably something that I should know the answer to but the more I consider it the more I get confused.

I know that there are members on here that have much more knowledge about electronics amongst other subjects than me, so thought that I would ask.

Thanks in advance.

Dave

[Ted Kendall]

Well, if the thing has been properly designed the valve distortion should dominate. I don't suppose you want hard clipping in a guitar amplifier, so it should have been arranged that the valve stage bends before anything else.

[mhennessy]

The distribution of harmonics has much more to do with circuit configurations than device type.

For example, a long tail pair produces odd harmonics when the standing currents in each device are not equal. That's why it's common to use a current mirror to load the collectors of a LTP. With that done, what's left is a low level of third harmonic.

All active devices - whether valves or solid state - have a non-linear relationship between input voltage and output current. This leads to even harmonics because the shape of the waveform is being changed, but not symmetrically.

Whereas many distortion mechanisms are symmetrical. Clipping and crossover distortion being the most obvious examples. These create odd harmonics.

When using solid-state, it's much easier and economic to use extra devices to minimise the "raw" even-order distortion, leaving behind (much lower levels of) symmetrical types. I guess that's where the generalisation comes from.

In your example, you'd have to consider the contribution from each part of the chain. Ignoring the synthesiser chip because it's impossible to know what it's doing, you could argue that the distortion contribution from the solid-start parts will be negligible compared to the distortion from the valve stage.

Assuming nothing is clipping, of course. But this is a guitar amp, and it's traditional to drive guitar amps so hard that their output stages clip - that's where they get their tone from, right?

That used to be the case. But even a modest guitar amp driven into clipping is far too loud for most scenarios. So now, they tend to be run at lower volumes, and the magical "tone" comes from digital processing, or perhaps a valve stage working at line level. The power amp is intended to run at nowhere near its maximum output level.

That's good from many points of view. For example, excessive on-stage volume is a real problem when doing live sound in small venues. With modern approaches, we only have the drummer to contend with! Also, solid-state output stages are much cheaper than valved equivalents. And lighter! And either work or don't work - they don't degrade over time as the valves get tired.

Cheers,

Mark

[knobtwiddler]

Mark beat me to it re: circuit type. There is much twaddle out there pertaining to what type of distortion is generated by specific active devices, when the truth is that implementation counts for vastly more than the type of device (assuming that the active device is functioning correctly!).

>Guitar amplifier has a solid state input

That's an open-ended description. It could mean a singleton JFET arrangement, which will predominantly generate a large amount of 2nd harmonic (a similar amount to basic grid input valve arrangement). Alternatively, 'solid state' could mean a modern opamp IC with FET inputs (OPA2134 is a good example - now discontinued...aargh!) that has extremely low distortion, to the point where, correctly implemented, it might be under -100dBV - i.e. inaudible.

>that goes to some design of a silicon effects synthesizer chip

Modern DSP is capable of generating any harmonic you want - assuming it's programmed to!

Out of interest: back in the 70s, H+H designed a solid-state circuit for their amps, in response to guitarists complaining that they didn't clip as nicely as the valve ones. Mentioned here: https://www.ssguitar.com/index.php?topic=1429.0

(first link I found - a bit more sniffing might unearth a schematic)

>then the audio signal goes to a double triode (ECC83 or equiv)

Again, very open ended description. An ECCC83 stage can be designed to be highly linear (will need decent voltage), or deliberately designed not to be. As Mark wrote, in an LTP configuration, the 2nd harmonic will be cancelled out. A common technique to add large amounts of 2nd harmonic is to fit a cathode-follower, run on low voltage. It's known as a 'starved plate' design by musicians and has stacks of 2nd harmonic. It's what's used in cheap valve DI boxes and the like.

>and finally goes to a solid state audio amplifier output stage.

This could introduce 2nd, 3rd - as well as higher order harmonic distortion, depending on topology, bias and implementation.

It's rare for SS stages to use an output transformer, but worth noting that their distortion is predominantly 3rd. A cheap transformer will also introduce a fair amount of odd harmonics such as 5th, 7th etc. You really get what you pay for in terms of distortion performance with transformers. Some firms give you the option for more steel / less nickel, in case saturation is what you want.

[knobtwiddler]

Addendum: it's worth noting that when guitar amps moved over to SS, they typically obviated the output transformer. When you compare SS to valve, you aren't just comparing active device types, as the transformer saturation is a big part of the valve amp sound. The transformer should be designed in tandem with O/P stage, to afford safe clipping - i.e. a speaker can be damaged by a relatively low powered DC-coupled SS amp driven into clipping, whereas the saturation characteristic + ac coupling nature of a transformer offers a safety net if it partners the amp well.

[G6ONEDave]

Thanks guys, it probably was a bad question on my part. The answers are quite interesting though and give more for me to consider. I suppose the only way to get a definitive answer is to do a full audio test on each specific amplifier using specialist analysing equipment. It seems to be a case of one size does not fit all as I initially hoped would be the case.

Dave

[kalee20]

Transistors produce both odd and even harmonics.

And for valves - most of the distortion is in the output stage, which if push-pull can't generate even harmonic distortion so it would be odd harmonics only.

Operated well below clipping level, valves do generate mainly even-order harmonics, which are relatively small.