What causes output transformer saturation, Current, Voltage, Frequency, DC?
The reason I ask is when loading the output tube with higher reflected impedance (8ohm speaker on 4ohm secondary tap) the output tube will swing less current and more voltage (load line is less steep).
If current is the primary cause for transformer saturation, then high load may cause less saturation, but this will increase the danger of arcing inside the transformer because of flyback voltage.
The reason I ask is when loading the output tube with higher reflected impedance (8ohm speaker on 4ohm secondary tap) the output tube will swing less current and more voltage (load line is less steep).
If current is the primary cause for transformer saturation, then high load may cause less saturation, but this will increase the danger of arcing inside the transformer because of flyback voltage.
Last edited:
"Standing" DC is an issue. SE O/P "iron" is air gapped in order to cope with the DC.
Deep bass or infrasonic noise can be troublesome. Power handling capability declines, with decreasing frequency. This source of core saturation is of great importance in designs that incorporate a GNFB loop. FWIW, I always employ an infrasonic noise suppressing high pass filter at the I/P of designs that employ GNFB.
Deep bass or infrasonic noise can be troublesome. Power handling capability declines, with decreasing frequency. This source of core saturation is of great importance in designs that incorporate a GNFB loop. FWIW, I always employ an infrasonic noise suppressing high pass filter at the I/P of designs that employ GNFB.
It's just the primary windings who can saturate your transformer, with dc (SE or unballanced PP) and ac (signal). For ac only the lowest frequencies.
Excess Current Through Transformer Coil either DC or AC will cause core saturation.
Voltage itself does not cause any saturation, only the flow of Current Through the coil--Either Primary OR Secondary. As Secondary is Driven by Primary, and you're not supplying any current or voltage from external source except by coupling from Primary--this can be discounted.
Arcing is only an issue when there's NO load on the secondary. Mis-matching a 16 ohm speaker on a 4 ohm tap will not cause an arc.
Voltage itself does not cause any saturation, only the flow of Current Through the coil--Either Primary OR Secondary. As Secondary is Driven by Primary, and you're not supplying any current or voltage from external source except by coupling from Primary--this can be discounted.
Arcing is only an issue when there's NO load on the secondary. Mis-matching a 16 ohm speaker on a 4 ohm tap will not cause an arc.
Last edited:
Great answer Alastair! Exactly what I needed to know.
So mismatching with higher impedance will actually saturate less the output transformer because of less current on it's primary as can bee seen on a load line.
What then are the risks/benefits of running double the impedance?
Looking at the impedance curve and knowing that a pentode behaves as a current source, I can see only benefits.
More power at output is one of them.
So mismatching with higher impedance will actually saturate less the output transformer because of less current on it's primary as can bee seen on a load line.
What then are the risks/benefits of running double the impedance?
Looking at the impedance curve and knowing that a pentode behaves as a current source, I can see only benefits.
More power at output is one of them.
Last edited:
To my knowledge this is true with SS amps but with Tubes it is the opposite.
Valve Amps: Valve verses Solid-state amps
A constant current source like a Pentode tube will deliver more power into higher loads, that's why it follows speaker impedance and results in a boost of power in the resonance frequencies of the impedance.
SS on the other hand is a constant voltage source which will deliver less current a higher load, resulting in less power to the speaker.
Is this wrong?
If so, please correct me (and Lenard Audio Institute).
Last edited:
The first step in modelling a real transformer is to have a large magnetizing inductance (in parallel) across the primary of an 'ideal transformer'. The primary current into the transformer model is now the sum of the current into the magnetizing inductance plus the current into the 'ideal transformer' (which depends on the load on the secondary).
It is the current into the magnetizing inductance which causes saturation. The current into the load (via the 'ideal transformer' part of the model) has no effect on saturation.
Therefore it is the voltage across the primary and its frequency (especially down to DC) that controls the current into the magnetizing inductance, and hence saturation.
It is the current into the magnetizing inductance which causes saturation. The current into the load (via the 'ideal transformer' part of the model) has no effect on saturation.
Therefore it is the voltage across the primary and its frequency (especially down to DC) that controls the current into the magnetizing inductance, and hence saturation.
Last edited:
not the last word - builds in assumptions which obscures the principles - you need to read more broadly, any one source may even be "correct" but slanted to a particular purpose that misleads as to general principles if taken alone
tubes or SS can be used as "current sources" in common cathode/emitter connection taking output from plate/collector or as "voltage sources" in cathode/emitter followers
triodes are a special case because of the built in plate V feedback that gives the mu limit, lowers plate resistance
pentodes with the screen grid AC grounded interrupt the internal feedback from the plate V, give very large mu - enough that plate output is commonly considered a current source
the current and voltage handling limits of SS can match to today's dynamic driver Z without transformers
tubes develop their power at a much higher impedance, have lower current and often need step down transformers on the output to effectively drive lower Z speakers
tubes an be used with feedback, its the output transformer upper frequency performance that is limiting - you can't apply lots of feedback around output transformers - doesn't depend on whether tubes or SS are used
triodes as mentioned have lower plate Z because of local internal feedback, some classic multistage tube amps use feedback from the primary, "before" the output transformer
Last edited:
Another way to look at this, without recourse to modelling, is to recall that if current in the primary magnetizes the core in one direction, the corresponding current in the secondary magnetizes the core in the opposite direction. If primary and secondary ampere-turns are perfectly balanced there is no net magnetization. We can therefore have very large primary and secondary currents without saturation.
In practice the ampere-turns never exactly balance. It is unbalanced ampere-turns which cause saturation.
In practice the ampere-turns never exactly balance. It is unbalanced ampere-turns which cause saturation.
My assumptions are in exact proportion to the principles I know.
The goal is to use my assumptions in reality and receive a nice predictable result, and avoiding damage or death along the way.
I am open to learn.
🙂
Can you please put this in relation to primary impedance that the tube sees and resulting power vs core saturation?
jcx,
Okay, I need some confirmation here.
If a Pentode behaves like a current source, can I assume that it will produce more power over a bigger load?
If a Triode behaves more like a voltage source, can I assume that it will produce less power over a bigger load, like SS?
The goal is to use my assumptions in reality and receive a nice predictable result, and avoiding damage or death along the way.
I am open to learn.
🙂
Malcolm, this is way above my head,Malcolm Irving said:
Can you please put this in relation to primary impedance that the tube sees and resulting power vs core saturation?
jcx,
Okay, I need some confirmation here.
If a Pentode behaves like a current source, can I assume that it will produce more power over a bigger load?
If a Triode behaves more like a voltage source, can I assume that it will produce less power over a bigger load, like SS?
Last edited:
If you are running an 8 ohm speaker on a 4 ohm tap, then (as you say) the load line on the power tubes will be less steep. You will get more voltage swing on the OT and less current swing. According to my understanding of transformer theory, that will increase the possibility of saturation. I think you will also probably get less power output (if the amp was designed for a 4 ohm load).
However, transformer saturation is only ever a problem at very low frequencies - so I'm not sure if the effect would be noticeable in practice.
However, transformer saturation is only ever a problem at very low frequencies - so I'm not sure if the effect would be noticeable in practice.
Pentodes behave like current sources when they are operating 'above the knee' of the plate characteristics. But if you increase the load too much, and reduce the slope of the load line, you have more chance of excursions 'below the knee'. That would increase distortion and would increase screen grid current.
Are we talking about hi-fi here or instrument amp? Excessive screen current can kill power tubes which are continuously overdriven (like in a guitar amp).
Are we talking about hi-fi here or instrument amp? Excessive screen current can kill power tubes which are continuously overdriven (like in a guitar amp).
Last edited:
The flux density in the core of a transformer varies with v/f. A transformer will saturate at a given primary voltage at a given frequency, more or less regardless of the impedance. Since power varies with v^2/R, doubling the load will give you half the power at a given frequency at saturation.
I have seen evidence to the contrary. Operating a tube amp without a load and allowing the amp to clip will force the voltage in the OPT to rise well above the B+ voltage (theoretically reaching infinity) often causing an arc inside the OPT or elsewhere in the circuitry.
Operating a tube amp with a load impedance significantly higher than intended (16 ohm SPEAKER on the 4 ohm tap) will ask the tubes to deliver more voltage into the OPT to reach the same power output. The B+ voltage is now the limiting factor, so the amp will reach clipping at a lower power output level than with the correct load.
Operating a class AB tube amp into clipping will cause the amp to reach a point where both output tubes are not conducting for an instant. With no load, an arc can happen since the magnetic energy being released from the core's collapsing field will create a large voltage in the primary. With a speaker load the energy is usually dissipated in the speaker.
If the note causing saturation is near the resonant frequency of the woofer or guitar speaker, the speakers instantaneous impedance can be very high. High enough for an arc to form. The resonant frequency of many guitar speakers is INSIDE the guitar's fundamental frequency range, so this IS a likely possibility.
I have seen a arc across the tube socket in a guitar amp. The user had a 16 ohm cabinet on a 4 ohm amp. A certain note on the low E string would cause an arc to erupt from pin 3 (plate) to pin 2 (grounded heater) on the 6L6GC tube. This caused no damage to the amp, but could have fried all sorts of stuff if the heater circuit wasn't grounded.
I have measured over 2.5 KV on the plates in a normal HiFi amp with a 16 ohm RESISTIVE load on the 4 ohm tap when the amp was driven to clipping. A speaker load will be worse around resonance.
A tube amp will deliver the most power into the proper load. The power tapers off as you get further away from this point in either direction, although the curve can be quite broad in some cases.
As the load impedance is reduced, you are asking the tubes to supply more current. At first the distortion will rise a bit and the power will increase. Eventually you will reach a point where the tube can not deliver the required current, and something bad will happen. It can be excessive distortion, overloaded tube, melted tube, or blown parts.
As the load impedance is increased you ask less current from the tube, but require more voltage to keep the power output constant. At some point the voltage swing approaches the B+ supply voltage and there is no more headroom, the amp clips. If the driver runs out of voltage swing before this point, the amp clips.
If you are designing an amp that requires say a 5K OPT and you want to use a 2.5K OPT, this MAY be possible. Usually the 2.5K OPT will have less turns wound with thicker wire, so it will have less primary inductance, and will saturate easier on bass frequencies.
Going the other way, using a 5K OPT when you need a 2.5K OPT MAY work, but the high frequency response may be compromised due to excessive leakage inductance.
OPT's, especially SE OPT's are an exercise in compromise. There are several factors that must be balanced to create a successful design. A well designed OPT will always work best at it's designed impedance.
That, said, I have often used OPT's at half or twice their rated impedance and IN A WELL DESIGNED CIRCUIT the results can be pretty good.
An OPT will exhibit large phase shifts at their frequency extremes. Mismatching impedances can make these phase shifts worse, which may render GNFB loops unstable. Avoid, or minimize GNFB with cheap or mismatched OPT's and avoid GNFB on a cheap mismatched OPT!
Select tubes and circuits that provide the lowest possible plate impedance (Rp) to the OPT, and / or use local feedback to reduce it. This will reduce OPT inductance related issues at the frequency extremes.
I have a bunch of 6600 ohm P-P OPT's that were designed for budget guitar amps. They work quite well as a HiFi OPT when used at 3300 ohms and driven with a low impedance sweep tube.
It is a combination of ac voltage applied to the primary, and dc current in the primary. For fairly complex reasons ! Each erodes the margin for peak core flux, the limit at which a transformer saturates. HTH!
AC current in the load has no bearing on saturation. That is not why it can be important to match load impedance in an OP tx to design value !
Transformer cores saturate when the number of magnetic lines exceeds the ability of the core to channel them in it's mass. This saturation causes the magnetic lines to flow outside the core in the air where the magnetic permeability is extremely low, resulting in a signal loss.
In AC music signals, the fields are constantly reversed (class AB). If DC is created from improper bias balance or class A, a great area of the core becomes magnetized by the electrons circulating in one direction.
However a constant field doesn't generate any electricity in the secondary.
AC signals can saturate a core if the VxA max is exceeded at any specified frequency.
In AC music signals, the fields are constantly reversed (class AB). If DC is created from improper bias balance or class A, a great area of the core becomes magnetized by the electrons circulating in one direction.
However a constant field doesn't generate any electricity in the secondary.
AC signals can saturate a core if the VxA max is exceeded at any specified frequency.
Last edited:
- Status
- Not open for further replies.