Drive an audio transformer from an AOP. Good or bad idea?

[Matt Syson]

Operational Amplifier of course referring to the initial premise of a block of circuitry (amplifier) that had inverting and non inverting inputs and a single 'output' that given certain characteristics could be used to do linear computation functions such as integration, differentiation, summing and others which were useful to the arms industry (gun aiming) where the overall function relied on external components and for convenience simple plug in active parts (the amplifier itself) so that exchange in the field was a simple afair.
Fun to implement with valves because they are unipolar devices but when transistors became a thing of both NPN and PNP flavours many new possibilities were dreamed up. Many things were understood and theorised WAY back but it was often near impossible using the technology of the day.

 

[Newmarket]

Yes, I thought about legacy being a possible explanation. Apparently in ignition applications, the rationale was to simplify wiring, since both windings ended up grounded at a side.
This cut-out, taken from Wiki, shows that the construction is not an autoformer at all. Pri and sec are clearly distinct.
View attachment 139374

The only thing that makes it an autoformer is the connection. Saves one terminal.
View attachment 139376
Sorry for the OT.

Yes. Terminology can be a bit "loose" here. But nowhere near as loose as "Tremelo Arm" in elec guitar world

 

[abbey road d enfer]

Yes. Terminology can be a bit "loose" here. But nowhere near as loose as "Tremelo Arm" in elec guitar world

Direct consequence of the fact that Leo was not a musician.
Did not prevent him to create 3 miraculous instruments, Tele, Strat and PBass.

 

[Newmarket]

Direct consequence of the fact that Leo was not a musician.
Did not prevent him to create 3 miraculous instruments, Tele, Strat and PBass.

Yeah - although if I had to choose I'd take a Jazz Bass. Or a Musicman bass. Maybe a G&L...

 

[abbey road d enfer]

It is not recommended to use an operational amplifier (AOP) to directly drive an audio transformer because AOPs are designed for voltage amplification and have limited current drive capability, while transformers often require higher current.

Typical old wives tale.
Many opamps have enough current drive to drive a transformer, as long as the transformer is designed for the specific task.
I have 10's of thousand products in the field that use a 5532 ot 5534 with a transformer.
Most of the current required to drive a transformer comes from the reflected impedance at the primary, except at low frequencies.
A 1:1 xfmr reflects a 600 ohms load as about 600 for most of the frequency range.
Problems may happen if the nominal inductance is too low, which increases the current demand at low frequencies.
A good 600:600 ohms transformer should run with a magnetizing current about 1/10th of nominal current.
Many designers use cheaper transformer, with low nominal inductance, resulting in large magnetizing current at low frequencies. That is why they have to resort to output stages capable of much more current drive than necessary.

 

[abbey road d enfer]

Yeah - although if I had to choose I'd take a Jazz Bass. Or a Musicman bass. Maybe a G&L...

I love my '51 reissue PB. Totally utilitarian, like the Telecaster.
I must say I'm not a dedicated bass player. I play bass occasionally in jams and for my own demos.

 

[dai h.]

Direct consequence of the fact that Leo was not a musician.
Did not prevent him to create 3 miraculous instruments, Tele, Strat and PBass.

that was my understanding also, but apparently he did have some experience playing (according to Fender, so the info looks legit) :

https://www.fender.com/articles/beh...n music,radio and electronics took precedence.

(kind of mind-blowing to me that (besides the PBass) he also designed the Jazz and Stingray--all hugely popular standards in the world of electric bass--plus loads of other stuff)

 

[abbey road d enfer]

that was my understanding also, but apparently he did have some experience playing (according to Fender, so the info looks legit) :

Should have writ "not a guitarist". I stand corrected.

 

[Mihobro]

Typical old wives tale.
Many opamps have enough current drive to drive a transformer, as long as the transformer is designed for the specific task.
I have 10's of thousand products in the field that use a 5532 ot 5534 with a transformer.
Most of the current required to drive a transformer comes from the reflected impedance at the primary, except at low frequencies.
A 1:1 xfmr reflects a 600 ohms load as about 600 for most of the frequency range.
Problems may happen if the nominal inductance is too low, which increases the current demand at low frequencies.
A good 600:600 ohms transformer should run with a magnetizing current about 1/10th of nominal current.
Many designers use cheaper transformer, with low nominal inductance, resulting in large magnetizing current at low frequencies. That is why they have to resort to output stages capable of much more current drive than necessary.

I believe back in the day it was Jensen that offered output transformers for specific audio purposes. Inputs designed for discreet, OP amp, and tube(not sure) drive circuits. They were very highly regarded for they're bandwidth and phase linearity. they were also very helpful with technical support, and often had basic examples of the circuits the transformers were designed for.

 

[abbey road d enfer]

I believe back in the day it was Jensen that offered output transformers for specific audio purposes. Inputs designed for discreet, OP amp, and tube(not sure) drive circuits. They were very highly regarded for they're bandwidth and phase linearity. they were also very helpful with technical support, and often had basic examples of the circuits the transformers were designed for.

Thye still have many app notes and typical schematics.
In particular app note AS30 shows an output transformer driven by a 5534.

 

[rh001]

Most if not all audio output transformers give better sound if the driving amp has a generous amount of current drive and that amount almost always requires medium power transistors. Such transistors will not fit in a integrated circuit opamp. An integrated circuit opamp can have a medium power transistor output stage tacked on. Sonically the prob if not using medium power output transistors is heard in the low end, particularly on kick drum hits when output voltage from opamp is more than +4db and the problem becomes more audible the higher the voltage. The undesirable sonic quality is similar to the sound of driving the output transformer into saturation clipping..similar but not the same. Money-wise a discreet opamp kit from Classic Audio Products of Illinois is quite affordable. Finished and ready to use discreet opamps can be rather expensive, $50 to $90.

 

[Khron]

Most if not all audio output transformers give better sound if the driving amp has a generous amount of current drive and that amount almost always requires medium power transistors. Such transistors will not fit in a integrated circuit opamp. An integrated circuit opamp can have a medium power transistor output stage tacked on. Sonically the prob if not using medium power output transistors is heard in the low end, particularly on kick drum hits when output voltage from opamp is more than +4db and the problem becomes more audible the higher the voltage. The undesirable sonic quality is similar to the sound of driving the output transformer into saturation clipping..similar but not the same. Money-wise a discreet opamp kit from Classic Audio Products of Illinois is quite affordable. Finished and ready to use discreet opamps can be rather expensive, $50 to $90.

Does it not actually depend what load impedance you might be attempting to drive, and to what levels? Surely you don't need 1A-capable drivers, for a 100mV RMS signal into 1-2kohm impedance...

Good luck powering a discrete opamp from phantom power, by the way - how many dozens of mA do those idle at?

Context, people, context - it matters!

 

[abbey road d enfer]

Most if not all audio output transformers give better sound if the driving amp has a generous amount of current drive

That means the magnetizing current is not as small as it should be. Generally due to too small magnetic core, or attempt to mitigate HF loss by reducing nominal inductance.

and that amount almost always requires medium power transistors. Such transistors will not fit in a integrated circuit opamp. An integrated circuit opamp can have a medium power transistor output stage tacked on. Sonically the prob if not using medium power output transistors is heard in the low end, particularly on kick drum hits when output voltage from opamp is more than +4db and the problem becomes more audible the higher the voltage. The undesirable sonic quality is similar to the sound of driving the output transformer into saturation clipping..similar but not the same. Money-wise a discreet opamp kit from Classic Audio Products of Illinois is quite affordable. Finished and ready to use discreet opamps can be rather expensive, $50 to $90.

As I mentioned earlier I have dozens of thousand of transformer output stages in the field, driven by a single monolithic opamp.
Don't let band aids made necessary by poor design become gospel.

 

[rh001]

Does it not actually depend what load impedance you might be attempting to drive, and to what levels? Surely you don't need 1A-capable drivers, for a 100mV RMS signal into 1-2kohm impedance...

Good luck powering a discrete opamp from phantom power, by the way - how many dozens of mA do those idle at?

Context, people, context - it matters!

An output used in Pro Audio equipment is/should be designed to drive into a 600 ohm load while still sounding great and drive into a 300 ohm load without too much loss of fidelity. That requires an output transformer of the size generally seen in pro audio gear,like for instance CMOQ3S from Cinemag. Although Dave Geren of Cinemag tells me that the next size up, CMOQ2S is a better seller. The latter is cleaner but I prefer the CMOQ3S because of it's sound, or rather the sound of a transformer with that core size. Medium power output transistors are needed to drive if the load is 600 ohms or lower and I'm not sure what the upper resistance is where one could switch to small signal transistors but yes there is such an upper resistance but that's irrelevant. The load resistance is not a free variable because though higher load resistance means less current flowing through the transformer it won't fire up and going to a smaller core size, as it will fire up, it just won't sound as good for some reason. Pro Audio output transformers are all approx the same size because that's as small as they can be and still be functional while sounding good.

As for the context, if the transformer is intended for microphone output then due to size and weight constraint it might be very small and signal level will be very small and so little current drive is needed and so medium power opamp output transistors are not needed. But then an opamp is not needed either even if one could be used via phantom power by being overly clever. I have never seen that in a pro quality microphone. Not all microphones run on phantom power.

 

[rh001]

That means the magnetizing current is not as small as it should be. Generally due to too small magnetic core, or attempt to mitigate HF loss by reducing nominal inductance.

As I mentioned earlier I have dozens of thousand of transformer output stages in the field, driven by a single monolithic opamp.
Don't let band aids made necessary by poor design become gospel.

I guess Neumann and Neve and Universal Audio were just wrong all those years.

 

[Khron]

The load resistance is not a free variable because though higher load resistance means less current flowing through the transformer it won't fire up and going to a smaller core size, as it will fire up, it just won't sound as good for some reason.

What's the last piece of "pro audio" gear you know of, designed and manufactured in the last 40 years, that's NOT a clone of a 40s-60s design, that has a 600-ohm *input* impedance?

Microphone preamp input impedance is routinely 1-3k nowadays, and line-level input impedance is 10-20k. To the best of my knowledge, anyway...

And what exactly is the "firing up" of a transformer? What's the process, what happens there and how?

 

[abbey road d enfer]

I guess Neumann and Neve and Universal Audio were just wrong all those years.

I don't know which Neumann or UA you are referring to, but regarding Neve, they used a single 24V rail in the context of being able to deliver +24dBu into 600 ohms, which indeed requires a step-up transformer, which increases the current demand by a factor 2 at least.
I've never seen a piece of equipment that relies on +24dBu into 600 ohmms to function properly.
With a 5534 driving a 600:600 xfmr, you can achieve +21dBu, which is enough to overload any kind of equipment.
Most of Neumann products used single 24 rail, which also justifies a similar approach.

One must learn not to make a universal law of a peculiarity.

 

[abbey road d enfer]

An output used in Pro Audio equipment is/should be designed to drive into a 600 ohm load while still sounding great and drive into a 300 ohm load without too much loss of fidelity.

Said who? What's the justification, in a world of bridging connections.