Smothing signal (similar way to vacuum tubes, analog)

[k brown]

Yes, Nelson Pass caused quite a revolution with his 'soft clipping' for SS amps.

[HughBanton]

Sorry to be two weeks late to this party (as usual). Interesting subject.

I confess I have many times been convinced by the superiority of tube amps for guitar. Someone playing full-on blues solos with a vintage AC30 or a Fender Twin just seems so much .. well, LOUDER, apart from anything else. I fondly remember experiencing (ho ho) Jimi's wall of Marshalls way back then. Boy, only 300watts (nothing by today's standards) but practically knocked you out of the room.

So what is it? It seems relatively easy enough to measure and emulate the different overdrive and non-linearity details of valve v solid-state, but what about transformers, and the reaction between a transformer and the connected loudspeaker, wherein the cone movement & any physical overshoot there is going to definitely induce currents back through the transformer, into the amp feedback, then off you go round and round. And then add the fact that the guitar itself is also 'listening' to all of this, adding to the feedback loop back through the amp input some mS later.

So I wonder: audio transformers saturate, and they also ring, particularly when driven hard, and when connected to a loudspeaker. Is this it? I don't recall any algorithm that focuses on these aspects, rather than on the more obvious measurable static distortions of valves or transistors.

How to emulate transformer (mis)behaviour. Anyone?

H

[k brown]

Especially as regards guitar amps, it's the combination of all of it. Especially since they're usually driven hard. One of the main things that makes them seem 'louder' is the increased second harmonic distortion well below clipping; at the same measured spl, the same sound with more 2nd sounds subjectively louder, mostly because it sounds 'fuller'. Tube amps can be designed to reduce this effect, but guitar amps have quite simple circuits, and the effect is a desired one.

[lalalandsynth]

"because what if it sounded smooth when you started but then sounds gritty after you add the graphics?"

How do the graphics affect the sound ?

[HughBanton]

An observation ..

As is well known, loudspeakers readily act in reverse, as microphones. (Most effective for mic'ing bass drums - I've seen it done). I established a few minutes ago that if you clap your hands in front of an open-circuit 12" speaker you can get a good 1 volt swing from the coil - wow. You'd imagine that during this 'Volume @ 11 blues solo' under discussion then speaker-cone overshoots, cabinet resonances etc. are going to induce at least this sort of extra signal, maybe many times larger?

A tube amp output transformer turns-ratio is going to be something like 30:1 I believe, so heading back into the output valves, via the transformer, you'll see 30v upwards added to what they were supposed to be doing. Clamped quickly by the amp feedback, in theory, but if the transformer's already wilting during this guitar solo it'll be too late to stop further transformer core-saturation and ringing. And to some degree this is a self-sustaining process .. valves > transformer > speaker > transformer > valves etc..

None of this will happen with a solid-state amp; there's no transformer to saturate or ring, and it's output impedance is invariably near-zero.

Strangely I don't recall seeing such a hypothesis before .. probably means it's nonsense

Or is it ??

H

[Spogg]

It makes perfect sense to me.

However I think that the output impedance of the transformer would all but eliminate the microphone effect. You said open circuit voltage would be 1 volt or so, but when the voice coil sees a few ohms I would expect virtually nothing, but I don’t know of course. It would be interesting to test.

Many years ago, when I worked in the lab at BSR, I made a system which monitored the current though a loudspeaker. The voltage across a 1 ohm resistor was used to measure the difference between the incoming dry signal and the speaker current (all solid state stuff of course). Then feedback corrected the signal so the speaker current matched the incoming signal. That effectively removed any self-generated signal, resonance effects and inductance related distortion and frequency response issues. I provided an A/B switch and the difference was remarkable. So much so that my boss said BSR should patent it. They never did but Phillips did! This was a time of great innovation when op-amps had just become available.

The system only worked when a single speaker was involved. If you had a cross-over and more than 1 speaker the phase shift would cause oscillation.

[k brown]

Great insights!

Slightly unrelated, but this all reminded me of an eye-opening test a friend of mine and I did back in the 70s. We both owned Phase Linear 700B amps and a pair of Large Advent loudspeakers. Occasionally I'd bring my amp and speakers over to his place and we'd bridge the amps and connect them to what were then called 'double Advents': two stacked vertically tweeter-to-tweeter, on each side. The sound was indescribable, but we kept blowing tweeters. We found out why, when out of curiosity, we connected a light bulb to the amps' speaker terminals, and boy did that baby glow! I think a lot of people aren't aware of just how much current a high-watt ss amp can generate.

[HughBanton]

by Spogg » Fri Jul 17, 2020 3:00 pm

However I think that the output impedance of the transformer would all but eliminate the microphone effect. You said open circuit voltage would be 1 volt or so, but when the voice coil sees a few ohms I would expect virtually nothing, but I don’t know of course. It would be interesting to test.

Speaker & transformer secondary are matched impedance, so you get good transfer. (I must have learnt that somewhere, many many moons ago .. ooh eck). I don't have a proper audio transformer to hand, but a check just now using a 12v mains transformer instead (similar ratio) gives a convincing 20v to 50v swing on the primary, so it looks like this might actually happen in an amplifier. Sort of.

H

[trogluddite]

Whatever the effects of transformer saturation etc., the "microphone effect" in combination with the amplifier output impedence will affect the damping factor of an amp/speaker combination. Moving any magnetic motor will always generate a current which produces a physical force which resists the movement (back-EMF), and for a speaker, the amount of this force will determine how much the speaker cone resists overshooting due to its own inertia - the "damping factor".

The near-zero output impedence of tranny amps allows maximum back-EMF to be generated, so maximises the damping factor. A valve/transformer output stage usually has a lower damping factor due to its higher impedance, so speaker cone excursions will be less well controlled. Since it's dependent on the motion of a physical system and magnetic induction, I imagine that the phase relationships with the input signal would be pretty complex. How important it is to the "valve amp sound", I've no idea - arguments about it get too technical or too subjective to convince me either way.

And how to reproduce it?! Speaker and cabinet effects are often handled using convolution, but that doesn't simulate level-dependent effects, which I imagine would be crucial to this. Maybe the right kind of phase-shifts in a feedback loop (all-pass filters?) might approximate it? I certainly agree with k brown that "it's the combination of it all" - it would take more than just chaining "gain stages" and a "speaker sim".

[HughBanton]

Thanks Trog, the 'damping factor' link is particularly enlightning.

Re 'ringing', I've been using damped-sine-waves for some time, particularly in my own organ-centric world for simulating wind supplies, which 'bounce' when you play big chords.

Here's the basic method in action, I'm sure it has a proper technical name! It occurs to me that it would be ideal both for simulating transformer ringing, at maybe 2-3kHz upwards (he guesses ...?) and also definitely for speaker overshoot, at around 60-100Hz for a guitar 12". Someone might like to experiment.

ringing.fsm

(8.63 KiB) Downloaded 1154 times

Another thing about valves .. they can be highly microphonic. So a combo amp like a Vox AC30, fitted with cheap Russian valves, is going to exhibit a fair degree of internal mechanical feedback to add to the chaos.

The plot thickens ..

H