Small Signal Magnetic Saturation

[Warren]

I want to explore the effect of transmitting a guitar
signal through a small saturated magnetic core. But I
don't want to do this at high power levels, using
saturation of the output audio transformer (in tube
amps), as is usually done. I'm aware that some
Wah pedals do exhibit some distortion using a small
inductor, but I want to consider other approaches.

I want to do is to do this in the "small", at a
signal power level. Something like passing a signal
through a "core memory donut", or iron washer etc. The
idea is to find something more readily available than a
core memory ring.

My fear is also that a washer will not be effective
enough, may also require higher power levels than
desired.

The link below describes how to avoid saturation of
toroid cores:

http://lists.contesting.com/_topband/2004-01/msg00104.html

The example he cits suggests a max current around 2.4 Amps
for the example he cited. So this rules out toroids in my
mind, and probably even smaller ones.

One thought I had was to use some steel wire, bend it into
a very small ring and crazy glue the ends. Then try to
put some windings on it to turn it into a saturable signal
transformer. I'd have to guess at the windings in this
case since I'd have no clue on the magnetic properties of
such.

I'm looking for suggestions or better ideas. Or is it
simply better to purchase a small inductor? If so,
what inductor range is suitable for op-amp signal
levels?

Warren

 

[Tim Williams]

Ferrite EMI suppression rings are generally high permeability and saturate
very easily, just a few amp-turns. You could get away with an inductor of
perhaps 1000 turns to distort signal levels on the order of mA.

I think you'll find that 1., 1000 turns will give a hell of a lot of
inductance, and 2., "small signal" properties are, well, small; distortion
is inherently low under this condition.

Uncommon materials are less linear. Square permalloy is wonderful stuff.
You can find cores at Nebraska Surplus Sales, namely 50B12-1D at:
http://www.surplussales.com/Inductors/FerToro/FerToro-3.html
Beware, permalloy is to magnetic cores what transistors are to vacuum
tubes -- it switches a lot faster, so you'll see a "harsher" cutoff when
it saturates.

Tim

 

[Warren]

George Herold expounded in [email protected]:

You want to add distortion to your signal? You would like for the
larger amplitude signals to have less gain? (Smoosh down the tops of
the Sine waves)

If the above is right, then it seems like some 'soft' diode limiting
might work. (I know nothing about guitar special effects.)

George H.

Diodes are used in everything from "Overdrive" (soft distortion),
to "Distortion" (harsher) to Metal distortion (extreme). Those
are trivial to reproduce.

What I'm looking to do is to emulate the best attributes of magnetic
distortion that happens in certain brands of tube amps by
saturating the audio output transformer. This leads to (or
supplements) a more pleasing distortion sound.

FETs have interesting characteristics as well, but it is the
magnetic attributes that I am after.

Warren

 

[Warren]

Tim Williams expounded in

Ferrite EMI suppression rings are generally high permeability and
saturate very easily, just a few amp-turns. You could get away with
an inductor of perhaps 1000 turns to distort signal levels on the
order of mA.

I think you'll find that 1., 1000 turns will give a hell of a lot of
inductance, and 2., "small signal" properties are, well, small;
distortion is inherently low under this condition.

I could approach 100 mA with a 2n2222 perhaps. Alternatively,
I could do more with a LM386. But more than that
would be more than I want to get into at this point in time.

Uncommon materials are less linear. Square permalloy is wonderful
stuff. You can find cores at Nebraska Surplus Sales, namely 50B12-1D
at: http://www.surplussales.com/Inductors/FerToro/FerToro-3.html
Beware, permalloy is to magnetic cores what transistors are to vacuum
tubes -- it switches a lot faster, so you'll see a "harsher" cutoff
when it saturates.

Tim

Interesting... Harsher is ok if it "sounds good", so some
experimentation might be in order. But I want to explore the
"normal" first, as this is probably what I'm after. I liked
your SS/tube analogy. That hits home. ;-)

Warren

 

[Tim Williams]

At the risk of displaying my ignorance yet again. ("What, Can't you
keep that thing under wraps Herold?") What are the 'attributes of
magnetic distortion' that you are trying to emulate? I figured that
as the transfomer saturated the signal amplitude would level off.
(Less gain at large amplitudes) But perhaps there is more to it?
Does it also depend on the frequency?

Saturation in tube amp output transformers is a low frequency, high
amplitude phenomenon. The effect is a sort of farting sound, where the
whole amplifier gain kind of gets shunted to zero once the transformer
saturates. It stays "off" until the signal reverses and the transformer
moves out of saturation.

The frequency depends on amplitude, so if say 10V saturates at 50Hz, only
2V will saturate at 20Hz, or you'll need a whopping 200V to even get close
in midband ~1kHz. Of course a typical 10W, 8 ohm tube amp won't make more
than 9Vrms, or maybe 13V peak, square wave / in clipping, so there is a
limited range of frequencies which will saturate. (Guitar amps are
probably wound with dinky transformers that saturate even easier...)

The suddenness of saturation will control harmonic production. You can
imagine the physics of farting: expanding gas opens a path through a
restriction; restriction opens, pressure drops, restriction closes. It's
a relaxation oscillator, where gas comes out in bursts, hence the
pulsating sound. If the pulses are very sharp, it will have a popping,
clicking or buzzing quality to it; if they are more rounded (cheeks
together..), you get a deeper, more resonating sound. In the same way,
the sharp saturation of permalloy will have a reedy sound, while the
softer saturation of iron or ferrite will result in less harmonics and, I
suppose, for obvious reasons, a more "tubey" sound.

For the circuit, obviously you'll want the inductor in parallel with the
signal, i.e. from signal to ground, same as the original transformer does.
You'll need a source with moderate current capacity, and some series
resistance (so saturation doesn't overload it), which an LM386 with 10 ohm
series resistor would be good at. It must be AC coupled, which means you
won't see saturation below some frequency, because that frequency is too
attenuated to cause saturation; for guitar, 20Hz is probably fine. You
want the system overdamped, so check that R, L and C come out okay. Or
maybe you do want it a little underdamped, so there's an LF resonance that
adds boominess.

For cheapasses, the "expensive" inductor could be replaced with an
integrator and clamp circuit. Integration counts volt-seconds, just like
an inductor does; when the count gets too high (or too negative), the
signal gets clamped. The actual signal path is not integrated, so it
still has the ordinary frequency response, but it does get shunted when
"saturation" occurs. In this case, harmonic content would be controlled
by comparator gain -- if it crosses the threshold and instantly the signal
goes to zero, you'll have loads of harmonics. If it goes up and lazily
starts tapering off a little at a time, you'll get a rounded curve. This
has the advantages of being able to control all aspects -- inductance,
saturation flux and sharpness.

Tim

 

[Warren]

George Herold expounded in @o14g2000yqe.googlegroups.com:

At the risk of displaying my ignorance yet again. ("What, Can't you
keep that thing under wraps Herold?") What are the 'attributes of
magnetic distortion' that you are trying to emulate?

In short, the technical term is the "tone". ;-)

I figured that
as the transfomer saturated the signal amplitude would level off.
(Less gain at large amplitudes) But perhaps there is more to it?
Does it also depend on the frequency?

George H.

Since there is an amplifier involved, there are aspects
like class A (and hence a DC bias) etc. that factor into
this (for some smaller amps). For other amps, there is
the class B or AB amps (like Vox AC30 is contraversially
reported to be when driven hard).

In my google travels (forget where) I saw that adding
DC bias eventually makes the 3rd harmonic higher,
and drops the fundamental in the process. With even more
bias you can see even more of that and the 5th rising up.
This is one controllable aspect that I want to experiment
with. I have seen this kind of behavior from diode
circuits too (for one diode only), but I suspect that the
magnetic process is better (cleaner) sounding. It is
this tough to measure quality that I want to explore.

I also suspect that the BH curve has other side effects that
are difficult to reproduce in other ways. Finally there
may be interactions with the amplifier which give it its
characteristic sound. For example, the final sees more
resistance when the core saturates, which must change
the response of the amp. There might even be some
signal folding going on, depending on amp etc. So I
need to investigate some of that.

So really, I am just looking for a jump off point for
some magnetic saturation experiments. Whether they
lead to something useful or not, is another matter.

Warren

 

[Warren]

Tim Williams expounded in

..

Saturation in tube amp output transformers is a low frequency, high
amplitude phenomenon. The effect is a sort of farting sound, where
the whole amplifier gain kind of gets shunted to zero once the
transformer saturates. It stays "off" until the signal reverses and
the transformer moves out of saturation.
Tim

I wonder about some amplifier designs in this regard. I have
read (or perhaps imagined) that in some finals there could be
signal folding going on. Once the transfomer saturates, any
extra signal will only see resistance. If this changes the
nature of the load seen, I could see the possibility
of it's gain actually decreasing even if only slightly, for
the isntance during saturation. This too, would introduce new
harmonic content. But I need to be careful to evaluate things
from the speaker end and not the primary. ;-)

What I do know for sure, is that preamp induced distortion alone
is not nearly as satisfying tone wise as combining it with
final stage distortion.

So there is some "tone" aspect to final amplifier and/or magnetics.
Companies like Murcury Magnetics seem to be successful at enhancing
otherwise tame little amps like the Valve Junior.

Warren

 

[Clifford Heath]

Tim,

Saturation in tube amp output transformers is a low frequency, high
amplitude phenomenon. The effect is a sort of farting sound, where the
whole amplifier gain kind of gets shunted to zero once the transformer
saturates. It stays "off" until the signal reverses and the transformer
moves out of saturation.

Thanks for your wonderful description. I've often wondered what it
was that made "that toob sound", but you nailed it.

Clifford Heath.

 

[legg]

Saturation in tube amp output transformers is a low frequency, high
amplitude phenomenon. The effect is a sort of farting sound, where the
whole amplifier gain kind of gets shunted to zero once the transformer
saturates. It stays "off" until the signal reverses and the transformer
moves out of saturation.

Intentional tube amp overloading is not normally an effect of core
saturation, it's one of tube GM shift at reduced plate voltage - a
form of amplitude compression. There are other forms of overloading in
the tube element with more objectionable effects, but these are rarely
induced intentionally; they don't sound nice and the tube screen
elements can be damaged in an uncontrolled manner.

RL

 

[Warren]

legg expounded in

Intentional tube amp overloading is not normally an effect of core
saturation, it's one of tube GM shift at reduced plate voltage - a
form of amplitude compression.
RL

It really depends upon the amp (design). Some have over designed
audio transformers, while others can be saturated. But I agree that
there are other effects, which are also part of the big picture.

Warren

 

[Warren]

George Herold expounded in

Thanks for the explanation Tim. Perhaps one could just use a smaller
transformer (made out of the same material as the output tranny) in
the signal chain. If you can get the same current-turns-area as the
output tranny then it should saturate in the same way. I can imagine
a 'distortion knob' that puts more current through the small
transformer and then throws the extra signal away before the output.
(That way you'd have more distortion without having to rail things
out.) If you wound your own transformer you could add a third set of
turns for DC bias.

George H.

I'm not against winding my own. In fact I can get some surplus cores
locally, which would seem ideal for that. The difficulty is determining
in advance what it will take to saturate them.

I'll have to review their stock to see what sizes there are to choose
from. Perhaps I can get a crude test done with a signal generator, if I
can get it "small" enough. Otherwise, I'll have to breadboard a LM386 for
the purpose.

Warren

 

[Warren]

Michael Kellett expounded in

Why not get yourself some little audio transformers - Farnel list
quite a few, some with power ratings as low as 10mW which should be
easy to overdrive.

10 mW eh? Hmmmm....

Some of the nicer tiny ones are a bit pricey so you
need to look yourself and pick some.

You could even make a tiny guitar amp using fets and transformers.

Michael Kellett

That would be a fun little project. I feel another
project coming on..

Warren

 

[Warren]

George Herold expounded in

Well, I should start by saying that I know very little about magnetic
materials. But if you can find out what the core is made of, and can
get a B-H curve of the material. (I'm not sure this is called a B-H
curve.) Then as long as you don't let all the different magnetic
'units' confuse you it should be fairly easy to get an idea of how
many turns at what current (given the Area) you will need to saturate
it. I'm sure there are many others here who can help with that.

George H.

About all I will get "spec wise" from the surplus store will be
size and price. ;-)

But I will acquire some and try some experiments. This kind of
thing is what I need to do over the Christmas holiday season,
after I've recovered from the "stuffing" of myself.

Warren

 

[Warren]

George Herold expounded in

....

I think iron tends to saturate at a field of about 1 Tesla (10^4
Gauss). I don't know the permability of your core, but a WAG is maybe
100 to 1000. Which means you need an 'air core' field of 10 to 100
Gauss. Now that is a pretty easy field make if you've got and amp or
so of current. Heck, I made a pair of 8 inch diameter Helmholz coils
that give 100 Gauss at 3 amps.

Have you taken any physics? Do you know how to calculate the B field
for a simple solinoid?

George H.

Yes- it was one of my favourite subjects but that was over 30
years ago. ;-) I'd have to do some "review".

Ideally, it should be accomplishable in a small effect pedal
powered by a 9V battery. 8->

Warren