output transformer question

[frank]

I am goofing with a small transformer and I try to saturate the core. Could this be magnetic core saturation?

[brett]

Hi
The "wobble" in the near-zero part is probably crossover distortion.  It is a result of misbiasing the output stage, or unevenness between the two "half" amplifiers.  Although it looks fairly "wobbly", it is surprising how bad it can get before it becomes audible.  Can you hear it?  I suspect that it wouldn't sound too good.

Saturation occurs at maximum and minimum input voltage (ie when the current and magnetic flux are greatest), and will appear as "clipped" output voltage.  If you can control the amp gain, just increase it until the peaks start to flatten out.  It might take a surprising amount of input, as transformers are usually under-rated (ie a 350mW transformer might take 1W before it saturates).  Also, transformers saturate more at low frequencies, so a good frequency to do your testing is about 400 Hz, not the oft-used 1 kHz..

By the way, frank, I reckon that there is a lot of unexplored distortion territory in saturating little transformers.  Lots of desirable vintage amps had small, low quality output transformers.  So keep at it and let us know the results.
cheers

[frank]

No,
this just an signal generator that goes 15Vpp in a unknown and small transformer, (if someone is interested i will put a photo), that is plugged in a small radio 8Ohms speaker (2,5in d.).  The oscilloscope is plugged at the output of the OT and the spkr..  This bump appear if I augment the voltage of the generator or if I lower the frequency of the apparatus.  then it is a LPF and this bump appears.  I don't think that it is a frequency effect, but I am not sure.  If I test it with an oscilloscope that have persistence (more of that phosphor screen I suppose), I also see a "vibrating halo" on the screen of the osc.. Could that "vibration" could be due to hysteresis and the bump be caused by saturation of the small iron ?  There are mysterious things happening in small transformers, or I am naive ?

Brett, I appreciate your responce. 

The photo has been taken at about 100Hz, so yes, it is in the lower frequencies. If I augment the frequency the bump disappears.  At VERY LOW frqcy, if there is no more audible sound in std sin wave, if I switch in square wave mode,  THEN IT BUZZ LIKE HELL.
I would like to take the specs of that (OT?).
For the sound, imagine a 2.5cm spkr on a hard table.

[frank]

Hen then it would mean that the OT have a reaction in the powerfull, BASSY and DISTORTED amps characteristics that would be fed in. before the sound  begin to be "alive".

[George Giblet]

I think the wobble is crossover distortion or overload protection in the oscillator (or more precisely, the output stage amplifier of it).  At low frequencies the inductive reactance is low, so for a given drive voltage, the output current from the oscillator is high.

If you can:
- increase the frequency to make the wobble disappear then increase the output voltage to make the wobble return, or similarly,
- decrease the output level to make it disappear then decrease the frequuency to make it return,

then that would explain it.  The first action decreases the output current, removing the problem and the second increases the current again making the wobble return.

[brett]

Hi
at saturation, you'll be increasing the input and not getting an increase in output.  Saturation definately happens at the peaks and troughs of the sine wave.  I suggest using an emitter-follower buffer after the signal generator to provide a low impedance source to the transformer.  Just in case the signal generator isn't up to the task and is producing the distortion.

Transformers aren't particularly "tricky" - they follow the basic laws of electronics.
cheers

[R.G.]

The bump in the middle is one result of leakage inductance. As you load up current one direction in the primary, you also unavoidably load up current in the leakage current. Then when you try to reverse the current at the zero crossing, the leakage inductance tries to keep it going. This was one of the the "fatal flaws" in hifi OTs that led to the introduction of the McIntosh output circuit with its quad primaries.  Saturating with big and/or low frequency signals makes it worse by putting more current into the leakage inductance because the primary inductance can't hold back the input current.

[frank]

Mr Rg,
you cleared up all my false toughs.
My rhyzomatic thingking lead me to absurdities that put my imagination in flames.  And I was thinking also that leakage inductance was negligible so I could not see it on a scope.  I went to 30Vpp on the primary, and I did not see sing of saturation then.  It was way cooler than my twin rev. OT  . So I guess you have to push a lot of current in there to have saturation.  Can this bump be corrected with negative feedback?

[frank]

Mr Rg,
you cleared up all my false toughs.
My rhyzomatic thingking lead me to absurdities that put my imagination in flames.  And I was thinking also that leakage inductance was negligible so I could not see it on a scope.  I went to 30Vpp on the primary, and I did not see sing of saturation then.  It was way cooler than my twin rev. OT  . So I guess you have to push a lot of current in there to have saturation.  Can this bump be corrected with negative feedback?

This forum is handy when you don't have a teacher "on hand".
Thanks alot RG., Brett and George.

Do someone knows what semi I could use between the generator and the OT to put more juice in there.
I might burn the windings before having saturation (No?).

[R.G.]

Saturation is a result of putting too big a voltage x time integral into the primary.

The voltage where a transformer saturates increases linearly with frequency. So to saturate at 1khz, you have to apply ten times the AC voltage that you do at 100hz because each half-cycle is 1/10 th as much time. So saturation is definitely a bass-response phenomena.

If you are feeding a large bass signal in and also a smaller treble signal, the bass signal can push the core to near saturation so the treble peaks push it in, but this happens near the end of a half-wave. If you feed DC voltage to a transformer, time becomes infinite, and the core will saturate - except for any magnetic field path which is air gap. This is why flat lamination cores have smaller inductance but more ability to tolerate DC than toroids - toroids don't have a functioning air gap. It's why matching the DC currents in your tube amp by using matched tubes or biasing individually is a good idea.

It's probably not practical for you to saturate the core at frequencies above 100Hz or so. Too much voltage needed. And you can never get saturation clipping evenly at a range of frequencies - I spent a lot of time trying to figure out how to do that a ways back.

The leakage inductance bump can be reduced by good transformer design, but it's hard to eliminate it with feedback because the leakage inductance is effectively in series with the OT primary and to an extent "hides" the transformer from the output tubes. The original article on the McIntosh transformer/amplifier made a big deal of this.

Using saturation for an audio effect is more difficult and subtle that people think. I hear people talking about the sound of an amp depending on OT saturation. It gives me an inner chuckle.

[raulgrell]

Using saturation for an audio effect is more difficult and subtle that people think. I hear people talking about the sound of an amp depending on OT saturation. It gives me an inner chuckle.

It doesn't depend on it, but it is certainly (positively) influenced by it. I have experimented with them in the past, I used a small transformer from a condenser mic with a virtual load (I think that's the right term) in a distortion unit, and the effect is quite pleasing. It colours the sound quite nicely, IMO.

[R.G.]

It doesn't depend on it, but it is certainly (positively) influenced by it. I have experimented with them in the past, I used a small transformer from a condenser mic with a virtual load (I think that's the right term) in a distortion unit, and the effect is quite pleasing. It colours the sound quite nicely, IMO.

Oh, I don't doubt that it colors the sound nicely. A transformer all by itself is a second order filter even if you ignore the nonlinearities.

What I doubt is that saturation of the core has anything to do with what you like. Saturation is effectively restricted to the lowest possible notes because it is very, very hard to drive a transformer to saturation at high frequencies. So it simply can't be a broad based tone shaper. Filter, yep, sure. Saturation can't "positively influence" anything if it does not happen.

Having a red car that goes fast does not mean that the car goes fast *because* it's red, even if there are many, many fast, red cars.

[Gus]

raulgrell

Chapter 5.3 of the big red book might have hint.

Read about source R and distortion with the different alloys.

[frank]

« Saturation is a result of putting too big a voltage x time integral into the primary. »

It begins well.

« The voltage where a transformer saturates increases linearly with frequency. So to saturate at 1khz, you have to apply ten times the AC voltage that you do at 100hz because each half-cycle is 1/10 th as much time. So saturation is definitely a bass-response phenomena. »

That is what I see on the persistence of the screen when I do a manual fr. swept.

« If you are feeding a large bass signal in and also a smaller treble signal, the bass signal can push the core to near saturation so the treble peaks push it in, »

I have only one big signal generator, the others that I have are small signal.  Could I pump the small signal with a power semi?

« but this happens near the end of a half-wave. »

What angle of a std sin wave do you mean ?

« If you feed DC voltage to a transformer, time becomes infinite, and the core will saturate - except for any magnetic field path which is air gap. »

If I understand, DC is infinite in time and can induce a magnetic flux in the core to saturation (not necessarily).  Foucault currents comes in the iron but not in the air gaps.  Lost power = f(w^2, flux density^2) in the iron when AC magnetisation.  Sheets of magnet iron is for diminution of Foucault and hysteresis therefore better transmission of power to the secondary.  The top and bottom of the hysteresis curves may move (am I wrong?), depending on saturation characteristic of the material.  Difference between ferrite toroids and laminated cores could be due to air gaps, because H therefore current, must be hier in air to have same H.  Saturation is a characteristic of the material, not the current.  Higher current and you'll hit the saturation (but not in air).

« This is why flat lamination cores have smaller inductance but more ability to tolerate DC than toroids - toroids don't have a functioning air gap. »

More air = less inductance

« It's why matching the DC currents in your tube amp by using matched tubes or biasing individually is a good idea. »

Yes, I had big trouble in my 135W twin because of that.

"It's probably not practical for you to saturate the core at frequencies above 100Hz or so. Too much voltage needed."

Yes, it is difficult.  Not possible in ROG designs as to have an effect.  And 100Hz is the turning point at 15Vpp

« And you can never get saturation clipping evenly at a range of frequencies "

It is by far, not a constant behaviour. 

« I spent a lot of time trying to figure out how to do that a ways back. »

It is profitable for everyone to do so.

« The leakage inductance bump can be reduced by good transformer design, »

Better interleavings = less leakage inductance = smaller bump

« but it's hard to eliminate it with feedback because the leakage inductance is effectively in series with the OT primary and to an extent "hides" the transformer from the output tubes. »

Is this a Zen koan on which I can meditate more to have Illumination?

« The original article on the McIntosh transformer/amplifier made a big deal of this. »

Is it written in Pali language or it is accessible?

« Using saturation for an audio effect is more difficult and subtle that people think. I hear people talking about the sound of an amp depending on OT saturation. »

Yes you hear it. Its not an illusion.  Some think that it is easy but I don't agree that an OT is a simple thing.  There are thousands of pages written on the topic and it could be a lifetime pursuit if someone keeps on digging nonstop.

« It gives me an inner chuckle »

You are a three phase AC GENERATOR -R.G.- don't waste all your power in Foucault loops.

[brett]

Hi

So it simply can't be a broad based tone shaper. Filter, yep, sure.

Indeed you are a wise man, RG.  For some time I was thinking that the tone of the Deacy amp and other popular small amps had a lot to do with lousy OT causing distortion.  But maybe they are just causing a mids "hump", a la tubescreamer, which is distorted by being seriously overdriven.

More air = less inductance

The air is an insulator, totally preventing DC current flow.  Maybe that's what you meant.  The amount of air isn't important.  Paper or plastic work, too.

To get saturation with a normal sized signal (say 1 to 5V p-p, 400 Hz), may I suggest an OT that has a cross-sectional area of the core of about 6 to 8 mm square.  Those little 1K:1K centre-tapped transformers at Radio Shack are a good size (15mm x 15mm overall).  As a rough rule of thumb, a transformer with a core of 10mm square can handle about 8W before saturating at 400 Hz.  Power handling rises with the square of stack area. (or approx the square of the square of stack width)
cheers

[R.G.]

« but this happens near the end of a half-wave. »
What angle of a std sin wave do you mean ?
It's not angle specific. What saturates the core is the primary accumulating a certain amount of volt-time product. If you start with a high frequency sine wave at some peak voltage, the core can't be saturated at all, as a full half-cycle does not include sufficient integral of volt-time. As you lower the frequency or increase the voltage, the core gets to saturation at the end of a half-wave first, as the end of a half-wave is where the volt*time integral is maximum. As you further lower the frequency or increase the voltage, the point of saturation creeps back into the half wave linearly.

« If you feed DC voltage to a transformer, time becomes infinite, and the core will saturate - except for any magnetic field path which is air gap. »
If I understand, DC is infinite in time and can induce a magnetic flux in the core to saturation (not necessarily). 
Yes.
Foucault currents
?? ?? ?? ?? Foucault currents? That must be a term for some other factor that's not used here in the USA. Eddy currents maybe?

comes in the iron but not in the air gaps.  Lost power = f(w^2, flux density^2) in the iron when AC magnetisation.  Sheets of magnet iron is for diminution of Foucault and hysteresis therefore better transmission of power to the secondary. 
You do describe it as eddy currents, so I'll assume that it is. Eddy currents act like a nonlinear resistance in parallel with the primary inductance and cause heating of the core material. They really have no effect on saturation.

The top and bottom of the hysteresis curves may move (am I wrong?), depending on saturation characteristic of the material. 
You're right, I think. The flux density B and the MMF H are functions of the material on a per-unit basis. Total flux requires multiplying by core cross section and total MMf per length of core path.

Difference between ferrite toroids and laminated cores could be due to air gaps,
That's yet a third animal.
Stacked laminations like standard E-I parts have a gap where the edges of the laminations meet. Wound iron toroids have essentially no air gap; it's a very wide, large area gap between the layers of the wound iron spiral donut. Ferrite toroids are a third thing. Ferrite is a ferrous metal compound with rare earth minerals to produce dust-fine particles of material. The tiny dimensions of the particles make eddy (Foucault?) currents much smaller than laminations can. But there are gaps between each dust particle in the final product too. Powdered iron is much the same. The permeability of ferrite and powdered iron is much smaller than iron laminations. They are generally used when frequency is too high to endure the power lost to eddy currents, which rises as frequency rises.

because H therefore current, must be hier in air to have same H.   
I'm not sure what you mean there. H is proportional to magnetomotive force per unit length, or ampere-turns per meter. Air is much more difficult to get the same ampere turns through, so it acts like a high magnetic field "resistance".

Saturation is a characteristic of the material, not the current.
But this is dead correct - materials saturate at a flux *density* regardless of their size.

Higher current and you'll hit the saturation (but not in air).
Current, as wrapped up in ampere-turns per meter has almost nothing to do with it. It's flux density that is the key to saturation. Literally, flux density stops increasing with increasing volt-time drive because all the available magnetic domains in the material are already aligned. When that happens, the permeability of the material reverts to that of free space at higher flux densities. The increase in H for higher volt-time drives reflects that fact that you can drive the material as hard as you like, but the flux density will not increase. That's saturation.

« This is why flat lamination cores have smaller inductance but more ability to tolerate DC than toroids - toroids don't have a functioning air gap. »
More air = less inductance
While that is true, that's not what I meant. It's not about the inductance of the iron. It's about the flux density of the magnetic field. For each ferromagnetic material, there is a maximum flux density it can support and still retain the high permeability that makes it act ferromagnetic. For iron this is something like 10k-20kGauss depending on the mix of materials and the history. As long as you work under that limit, the iron has a permeability that is high. When you force the flux density over the max for the material, there is no longer any high permeability and the material reverts to the permeability of free space. This is in fact how flux-gate magnetometers work; the have a core of ferromagnetic material is that is run into and out of saturation. When it's let fall below saturation, the permeability gets high and the earth's M-field prefers to be inside the core. When it is then saturated, the earth's M-field sees no reason to divert through it, and the field jumps back out of the core. The sense windings pick up a voltage from the M-field jumping into and out of the core.

It's not about inductance, although when you saturate a core, its incremental inductance falls too, like the permeability. This is how magnetic amplifiers work.

« I spent a lot of time trying to figure out how to do that a ways back. »
It is profitable for everyone to do so.
I agree. Very few people ever take the trouble to reason out how M-field and ferromagnetics work.

« The leakage inductance bump can be reduced by good transformer design, »
Better interleavings = less leakage inductance = smaller bump
Yes.

« but it's hard to eliminate it with feedback because the leakage inductance is effectively in series with the OT primary and to an extent "hides" the transformer from the output tubes. »
Is this a Zen koan on which I can meditate more to have Illumination?
You can if you want, I guess. But it's technically accurate as stated.

« The original article on the McIntosh transformer/amplifier made a big deal of this. »
Is it written in Pali language or it is accessible?
I've never found it on the web. I found it in a 1950's issue of a UK hifi magazine. I doubt it's translated to Pali.
I'll sure be glad when we get that Star Trek universal translator figured out. About the best I've been able to do is to memorize the Chinese pictographs that mean "Exit" in English. 

« Using saturation for an audio effect is more difficult and subtle that people think. I hear people talking about the sound of an amp depending on OT saturation. »
Yes you hear it. Its not an illusion.
What the people describe is almost always something that saturation cannot do. Saturation can be heard OK, but it's not what gets described in any case I've ever heard.

« It gives me an inner chuckle »
You are a three phase AC GENERATOR -R.G.- don't waste all your power in Foucault loops.
Good advice!

[frank]

More air = less inductance

The air is an insulator, totally preventing DC current flow.  Maybe that's what you meant.  The amount of air isn't important.  Paper or plastic work, too.

I think R.G. was talking about mutual inductance.  Wether you have a conductive material or not, you can still induce current in the output coil.  Permeabilities is valid for magnet steel, ferrite and air (or plastic), etc..  Augmenting the "conductive content" seams to increase the permeability ratio compared to emptiness. That is what I mean.

To get saturation with a normal sized signal (say 1 to 5V p-p, 400 Hz), may I suggest an OT that has a cross-sectional area of the core of about 6 to 8 mm square.  Those little 1K:1K centre-tapped transformers at Radio Shack are a good size (15mm x 15mm overall).  As a rough rule of thumb, a transformer with a core of 10mm square can handle about 8W before saturating at 400 Hz.  Power handling rises with the square of stack area. (or approx the square of the square of stack width)

Mine is 25mm^2 at the core.  I begin to saturate at 15Vpp-100Hk as said before.  If I remember well, I took the photo at that settings last night.

Are you speaking of Resistive KOhms:KOhms  (for DC)?  Not the vectorial impedance sum?   Is this the standard to speak of OT windings or these numbers are the real impedance?

Your suggestion is compatible with what I have seen and is welcomed.
Thanks again Brett.  

[frank]

Mr R.G., I will read your text a 2nd and 3rd time before putting more text.  I am in my emotive phase and my brain does not track too well after all that snow.  I will perhaps be hanging in the lounge.  All your texts helps put bits of info that have been acquired but not applicated.  Plus, playing with the scope is very instructive and more easy than spice models.  I am having a lot of fun.

Wiki quote:
"An eddy current (also known as Foucault current) is an electrical phenomenon discovered by French physicist Léon Foucault in 1851."

They call it like this in french literature.  Perhaps France is a very proud nation and its a burden for Quebec's people.

What the people describe is almost always something that saturation cannot do. Saturation can be heard OK, but it's not what gets described in any case I've ever heard.

I meant that we hear more of saturation on the forum than we hear it in the speaker.  lol
As said before imagine a small 2,5in. spkr on a hard table.

See Ya Soon Mr RG

[raulgrell]

raulgrell

Chapter 5.3 of the big red book might have hint.

Read about source R and distortion with the different alloys.

the big red book?

[frank]

Its a secret scripture that is kept away in the Vatican I think.  I do not find it in the bible.