Output Transformer Saturation

[lowell]

I've read that the 5E3 transformer saturates. What is saturation? How do you pick a transformer that WILL saturate w/ a certain power tube complement? Is it the VA rating specifically? My initial thoughts are that a transformer w/ a lower VA rating will saturate earlier. Is this correct? I'm mainly asking because I've built an amp w/ 2 channels. One channel is stock 5E3, however the OT is a 50w Mercury Magnetics unit and it has a solid state rectifier (plate voltages are still 5E3 voltages though). The amp breaks up but is nowhere near that "gonna explode" sound that stock 5E3's get when cranked.

[Diablo]

Saturation is magnetic saturation of the transformer iron. Iron is a magnetism multiplier. When you apply a magnetic field to a piece of iron by wrapping it with a coil with some current flowing through it, the magnetic domains of the iron line up in the same direction as the applied field, and this creates an induced magnetic field that can be around 2000X the strength of the applied field. However, once all the magnetic domains are lined up in the direction of the applied field, the iron is saturated. You can increase the applied field, but you won't increase the induced field (well only slightly). The more iron you have in your transformer, the harder it is to saturate. The type of steel can have an effect too on when it saturates. When the transformer saturates, it acts like a compressor because it can't put out any more volume. It also adds distortion because certain frequencies saturate first. I would compare the weight and size of your 50W OT to the more typical 5E3 OT. I bet your transformer weighs 2-3 times as much.

[tubeswell]

Yep the stock 5E3 PP OT is a little thing rated at about 12-15W tops. The bigger they get - the cleaner they'll be (and the better they will be at carrying bass freqs).

[lowell]

Ok thanks for that. Now, why doesn't the little 5E3 OT burn out when being driven so hard and over spec?

[R.G.]

Diablo's description of the domains aligning inside the iron is a good overview.

Transformers are hard to get a handle on because everything affects everything else to some extent.

Some tidbits:
1. You cannot saturate a transformer from the secondary. You can only saturate a core with what you do from the primary. You may be able to overheat a transformer by overloading it on the secondary, but that's not magnetic saturation. You can overheat a transformer with a propane torch too; it isn't saturation either, and is exactly analogous to what an overheating secondary is doing.
2. Transformers are easiest to saturate at low frequencies. They saturate when they have too much volt-time integral across their primaries. In a transformer which is fed from a voltage source, the (current x turns), and hence magnetomotive force (MMF) in the core increases with time unless either the wiring resistance stops the current from increasing, or if something external to the transformer limits the current. If one of these doesn't happen, the core current ramps up in response to the voltage-time product on the primary. Stripped of all the transformer specific language, you can only saturate an output transformer by driving it with a signal that is too big a voltage at too low a frequency.

It's the product of voltage and time that saturates them. The voltage a transformer can withstand on its primary without saturating decreases linearly with frequency.

3. Correspondingly, the power a transformer can transfer without saturating is proportional to frequency (ignoring core losses for a moment). A transformer's power rating is largely determined by the lowest frequency you feed it. Bass amp output transformers have to be bigger than guitar output transformers because of that lower octave.

[Bruce / Mission Amps]

Yep the stock 5E3 PP OT is a little thing rated at about 12-15W tops. The bigger they get - the cleaner they'll be (and the better they will be at carrying bass freqs).

The smaller 10 watt Vibrolux and BF Princeton amps used a smaller +12 to 14 watt OT but the 5E3 is a little bigger... closer to 18 to 20 watts, closer to the dims and specs of a BF Deluxe Reverb.
The 5E3 amp will not make a clean 18 to 20 watts though due to the low plate and screen to cathode voltage, high idle current and saggy cathode biasing.
A really good one will just barely get to a clean 12 watts. A frickin' GREAT sounding 10-12 watts!
The exact same amp with the same output tranny and higher B+ of +410vdc (think GZ34 rectifier), running with fixed bias will easily hit a clean 18 watts or more.... they just don't sound like a tweed Deluxe anymore.
However, some builders do use the smaller Princeton OT because it also is about 7500-8500 ohms, same as a tweed Deluxe and they still sound quite good with the smaller OT.
In my entire amp career I've only seen two or three tweed Deluxe amps that had bad output transformers, but I have seen way too many BF Princeton amps with bad OTs and PTs.
Mostly from knucklehead guitar players trying to run a pair of 6L6Gcs in the amp to get more power output.

[R.G.]

Good post. In addition to an OTs power rating being dependent on the signal level (i.e. voltage) and lowest frequency (i.e. time) it must carry, it is dependent on core loss (i.e. high frequencies it must carry) and resistive losses (current density in the copper), as well as the power supply (B+ size) that feeds it and the amount of DC it carries (biasing in the output tubes).

It's hard for most people to comprehend, but *everything* is a compromise in transformer design: high frequency, low frequency, bandwidth, pulse fidelity, distortion, primary inductance magnetizing losses, core material, magnetic path length, gaps or lack thereof, coil width, coil height, coil interleaving and/or sectioning, window winding area, number of windings and taps, required voltage insulation, copper thickness, weight, allowable temperature rise, potting material or lack thereof, physical arrangement of the core and coils, everything.

It makes the term "rated power" almost meaningless unless all the other criteria and information about how it's used are also specified.

The smaller 10 watt Vibrolux and BF Princeton amps used a smaller +12 to 14 watt OT but the 5E3 is a little bigger... closer to 18 to 20 watts, closer to the dims and specs of a BF Deluxe Reverb.
The 5E3 amp will not make a clean 18 to 20 watts though due to the low plate and screen to cathode voltage, high idle current and saggy cathode biasing.
A really good one will just barely get to a clean 12 watts. A frickin' GREAT sounding 10-12 watts!
The exact same amp with the same output tranny and higher B+ of +410vdc (think GZ34 rectifier), running with fixed bias will easily hit a clean 18 watts or more.... they just don't sound like a tweed Deluxe anymore.
However, some builders do use the smaller Princeton OT because it also is about 7500-8500 ohms, same as a tweed Deluxe and they still sound quite good with the smaller OT.
In my entire amp career I've only seen two or three tweed Deluxe amps that had bad output transformers, but I have seen way too many BF Princeton amps with bad OTs and PTs.
Mostly from knucklehead guitar players trying to run a pair of 6L6Gcs in the amp to get more power output.

[Enzo]

I'm still thinking about that propane. Y'think I might get better sustain with propane?

[wizard333]

The same is true of pickups...........DCR doesnt measure output............

Unless you know you are comparing two pickups of the same magnet type, not de-gaussed, wound with the same wire gauge. Then it is.

If you fix some of the variables, then talking about transformers in terms of rated output makes some sense. In a vacuum, no; but with some things specified, yes.

[R.G.]

The same is true of pickups...........DCR doesnt measure output............

Correct. DCR measures... resistance. It's only when you know the physical size of the coil and wire gauge that DCR becomes an estimate of turns.

Unless you know you are comparing two pickups of the same magnet type, not de-gaussed, wound with the same wire gauge. Then it is.

DCR can only ever be a guide to output, even in the case of two physically identical magnet structures, and even winding tension can both pack wires tighter (shorter average turn length => lower DCR and vice versa) and can stretch wires (smaller wire cross section => higher DCR) and these two effects compete. Magnet flux density can vary even with the same metallurgical composition and history. And output is determined by
(a) flux crossing turns of conductor and (b) both obvious and parasitic loadings on the impedances of the coil.

If you fix some of the variables, then talking about transformers in terms of rated output makes some sense. In a vacuum, no; but with some things specified, yes.

The "some things" gets pretty extensive. You often have to nail down a bunch of things to get a fair estimate.

[Guitarist]

With those Princetons, perhaps swapping in 5uF caps on the cathodes and the odd .01 plate cap may save an OT by reducing the bass content. They likely also pushed it with some mean little pedals

My Carvin TS100 was designed (I figure) this way to use smaller "50 watt" OTs. I widened the pre-stage V1b caps a bit from their stock .0047 uF resulting in much better tone. It was rather anemic prior. I also added KT77s but I never crank it and there is a fan inside

I generally like 6V6 tone over el84.

[lowell]

Anyone know of a way to make a transformer saturate earlier? Eg, putting an iron can over it, or imposing some external EMF on it?

[Old Tele man]

Anyone know of a way to make a transformer saturate earlier? Eg, putting an iron can over it, or imposing some external EMF on it?

...run a DC-current through it, with the AC-signal current "riding" on top.

[R.G.]

Anyone know of a way to make a transformer saturate earlier? Eg, putting an iron can over it, or imposing some external EMF on it?

Sure. Run a lower frequency through it, or a bigger primary signal. Transformer saturation is inextricably linked to frequency and signal size.

The simplest thing to do is to put DC through a winding, any winding. That shoves the magnetic operating point toward one saturation or the other, depending no the current direction. Problem is, it's only one direction, so one side of the signal saturates, the other doesn't.

The bias current of the output tubes runs through the two half primaries, but it's set up to cancel - the current in one half-primary causes an MMF that cancels that from the other half-primary, so the net MMF from the bias current is - ideally! - zero. It's not possible to make the bias currents reinforce on both sides of the B-H curve at the same time.

You can saturate part of the magnetic path with windings on the outer arms of the E-I core, but this doesn't saturate the iron under the primary and secondary, only makes for a lot of leakage.

Putting an iron can over it does nothing except hold in some of the leakage field.

The frequency dependence is something that most people never tumble to when thinking about transformer saturation. The simplest way to saturate a transformer is to run a lower frency signal into it. Trouble is, if your transformer saturates at, say, 20Wrms output levels at 80Hz, it won't saturate until 40Wrms levels at 160Hz, and 80Wrms levels at 320 Hz.

[R.G.]

...run a DC-current through it, with the AC-signal current "riding" on top.

Yep - but it only saturates the polarity of the signal which reinforces the DC current.