Teach me more about transformers please!

[Babaluma]

I know there are a lot of very clever EE people on here, so I was wondering if you could help me out.

Transformers have been a bit of a buzz word recently, and I have ten in my stereo mastering chain:

Chandler Germanium Compressors - Custom MI 3.11 Transformers on Inputs, Custom Massive C-3795 Transformers on Outputs

Pullet EQ - Sowter 9857f Transformers on the Inputs (still in line even when bypassed), no transformers on Outputs.

Chandler TG2 - Custom Massive C-3763-2 Transformers on Inputs, Carnhill VTB9045 on outputs.

I think the custom ones in the Chandler gear may be made by Cinemag, but I'm not sure. Any clues from the numbers/letters on them?

I also have four little cute ones in my pair of Real Ring Mods!

This is most of what I know about them, based on practical experience and research so far:

1) Some types can magically do crazy shit, like change your house voltage from 110V to 240V (or vice versa), seemingly defying the Second Law Of Thermodynamics.

2) They can change a single signal into a double signal with reverse polarity (like a DI box balancing a guitar signal), which can be sent further than an unbalanced signal, without degradation or picking up extraneous noise.

3) They can have a sound of their own, and the harder you hit them with an audio signal, the more you can hear their sound. It can sometimes sound pleasing, in a warm and fuzzy way.

4) Best of all, they have the amazing power of removing DC offset in an audio signal. I sometimes receive tracks for mastering with really bad DC offset. After passing them though my outboard chain, the DC offset is miraculously gone, and I've been reliably informed that this is due to the transformers (I just don't understand the how or the why).

What I'd like to know is:

1) More general info pertaining to transformers. What the hell are they and what exactly do they do, and why are they needed, and why are they sometimes NOT used etc.

2) The how, what and why of it all (more in depth principles, the physics etc.)

3) How they differ in form and function from Inductors (they are both just coils of wire, right?), tape heads, or Tesla Coils.

4) Are there huge transformers powering our houses, and even huger ones at the power stations?

5) Any of my mistakes corrected!

I love them and use them every day, so would just like to know a little more about them. All links or explanations gratefully received!

[EDIT] Pretty informative link here, but a lot of it is way above my maths ability...

http://en.wikipedia.org/wiki/Transformer

[wavecircle]

I can give you the physics of them.

It is a process called mutual induction. When a current flows in a wire it generates a magnetic field, likewise if a conducting wire is placed near an alternating magnetic field a current will be induced in that wire.

A transformer comprises of an input coil and an output coil. Both of these coils are wrapped around usually a square shaped iron block though there are other mechanisms. (Iron has a very high permeability, this means a magnetic field will carry very happily in iron).

For a step up transformer the number turns on the coils of the output will exceed that of the input. If the ratio of turns is say 1:2, an input voltage of 120v will produce an output of 240v. Like you mentioned this seems to defy thermodynamics but what you are ignoring is the current. Power in an electrical circuit is voltage and current multiplied. In a transformer when the voltage is stepped up the current is stepped down. So in our example of the step up transformer an input current of 2A will have an output of 1A.

The same principle applies to stepping down voltages just the number of turns on the output will be less than the input.

Their uses: National grid, high voltage transmission of electricity means very low currents which equates to very low losses of energy. Once something has been stepped up it will have to be stepped down in a local substation for domestic use.

Most of the appliances in your house that don't use mains will have transformers in them. AC-DC converters will have transformers in them along with rectification circuits.

[Babaluma]

Hey wavecircle, thanks so much for helping me to tackle this! Your explanation is very clear and didn't bring in any hard core algebra, so thank you!

So they operate kind of like building an electro magnet when you were a kid (making a wire coil and putting a 9v battery on it)? Except they are passive devices, right?

Yep, have heard of nickel and steel transformer cores too I think.

Is the term "primary" and "secondary" referring to the input and output coils?

I have heard they can be used for "Isolating", but am not sure what this means. Does it mean the input and output coils are never actually directly connected together, but only influence each other through some kind of magnetism?

Think I got the step up and step down thing down pat now, thanks for the really lucid/clear explanation.

Toroidal power transformers and wall warts are also different types then, I am guessing.

Sorry for more questions and inane rambling, and thanks once again!

[wavecircle]

Yeah, they are a bit like building an electromagnet except instead of a nail you have a square shaped iron block or ring. Transformers are completely passive yes, I am sure some EE will interject here but I am a physics person so let them have it. Primary and secondary is the common term in physics for input and output.

The primary and secondary coils should never touch, it is the magnetic induction in the iron core which links the two coils. The primary is bringing the current, this induces a magnetic field which the core carries. This magnetic field then "links" with the secondary coil, inducing a second current.

As far as I am aware all transformers use the same principle, I am guessing an EE will know more about the differences between toroidal and wallwarts.

[Graham Hinton]

1) Some types can magically do crazy shit, like change your house voltage from 110V to 240V (or vice versa), seemingly defying the Second Law Of Thermodynamics.

No magic, you cannot get more out than you put in, but a transformer has to be considered as a "power gearbox". It is almost the perfect machine, but there are small losses due to the windings heating up (copper losses) and the core not being perfectly coupled (iron losses), it is still normally over 99% efficient.

When you realise that power out is almost the power in it will make sense of what happens to the voltage, current and impedances. If the voltage is doubled then the current is halved so the VI product (=power) is the same.

2) They can change a single signal into a double signal with reverse polarity (like a DI box balancing a guitar signal), which can be sent further than an unbalanced signal, without degradation or picking up extraneous noise.

All the power going into the primary winding(s) is converted to a magnetic field and then back again to the secondary winding(s). This means that there is isolation between each winding and any common reference has to be made externally. This makes the tricks you mention very easy, if you want to reverse the phase just reverse the connections of a secondary.

A DI box is a microphone step up transformer wired backwards. The reason that it drives a longer line is that the output impedance is lowered and the capacitance of the cable has less effect. You don't get anything for nothing, the price is that the signal level is lowered so there is degradation. A transformer will also pick up any external magnetic field unless it magnetically screened, specialist audio ones sometimes have mumetal covers, but it may lose its properties if heated or given a mechanical shock.

3) The can have a sound of their own, and the harder you hit them with an audio signal, the more you can hear their sound. It can sometimes sound pleasing, in a warm and fuzzy way.

All transformers are optimised for a particular use by their size and windings, both number of turns and size of the wire. This puts a limit on the strength of the magnetic field than can be generated and is called saturation when the core cannot take any more. This distorts the waveforms.

Lower frequencies require more iron core, higher frequencies can work in just air so the size and shape effects the frequency response.

4) Best of all, they have the amazing power of removing DC offset in an audio signal. I sometimes receive tracks for mastering with really bad DC offset. After passing them though my outboard chain, the DC offset is miraculously gone.

The coupling through a transformer works because the magnetic field is changing. At DC there is no change.

Most of the audio usage of transformer goes back to early telephone systems before there were amplifiers. That's where the 600ohm audio standard comes from and maximum power transfer was often the main consideration. In modern audio maximum voltage transfer is often more important.

Transformers have some disadvantages too, like weight, noise and cost.

What I'd like to know is:

1) More general info pertaining to transformers. What the hell are they and what exactly do they do, and why are they needed, and why are they sometime NOT used etc.

Well, I would have said the Wikipedia entry...

2) The how, what and why of it all (more in depth principles, the physics etc.)

Well, I would have said the Wikipedia entry...

3) How they differ in form and function from Inductors (they are both just coils of wire, right?) or Tesla Coils.

An inductor has only one winding and its magnetic field operates back on the same winding. A transformer has two or more windings coupled by the magnetic field. A Tesla coil is a transformer specialised for a high step up ratio and high frequency operation.

[Babaluma]

Fantastic wavecircle, thanks once again for clarifying!

The primary and secondary coils should never touch, it is the magnetic induction in the iron core which links the two coils. The primary is bringing the current, this induces a magnetic field which the core carries. This magnetic field then "links" with the secondary coil, inducing a second current.

Ah, OK, thanks! This is my first ah-ha moment in this thread. So somewhat like the magic of tubes where electrons are flying around from one point to another in a vacuum, as some form of heated plasma, but are not strictly physically attached to each other, the primary and secondary windings of a transformer are also "isolated" from one another, yet are nonetheless affecting each other via induction (I need to look that one up, some form of magnetism?)

I just find it all so incredible, because that actually means the audio you put in to a transformer is completely different than the audio you get out of it, and yet it somehow sounds remarkably similar (or even subjectively nicer). Absolutely amazing/crazy/magical!

Any idea on where the DC offset goes to?

[wavecircle]

I think you are referring to Crooke's tubes. That isn't plasma, it's just excitation of gases. Getting into quantum physics here.

Induction isn't a special type of magnetism. It's simply the natural process of generating a magnetic field whenever a charge moves i.e. current. All currents generate magnetic fields.

The audio out of a transformer won't be that different, the waveform will remain the same, just the output will be at a different voltage level to the input. I guess the core could colour the waveform. This is something Graham will know a lot more about than me.

[Babaluma]

Graham Hinton, thanks so much for taking the time to reply. I really apprecite it. One of the reasons I made this thread was due to your response in the DI thread, where I realised my knowledge of transformers was pretty lacking. I was hoping you would reply!

Thinking of them in terms of a "power gearbox" and voltage/current ratio (whilst keeping the same overall power) is certainly helpful.

Also, thinking about their "insertion loss". My all passive EQ drops the signal by about 30-40dB, I am guessing due to the use of transformers and inductors.

So, to simplify, the bigger/beefier the transformer, the lower frequency it can go? I need to spend some more time deliberately saturating the different transformers in the chain in turn, and listening to the results. But I'm guessing some higher quality transformers will not saturate unless fed with silly levels that you would rarely find in audio applications?

Thanks for answering the DC offset removal question. Transformers only work with AC, they only "do what they do" when the signal is moving/changing, and hence do not react/pass DC. Interesting!

And finally, thanks for clearing up the inductor/tesla coil conundrum too! Aren't magnetic tape heads a form of inductor?

I'll go back to Wikipeadia and study the main transformer article again, and inductance/induction/inductor...

[Babaluma]

I think you are referring to Crooke's tubes. That isn't plasma, it's just excitation of gases. Getting into quantum physics here.

Induction isn't a special type of magnetism. It's simply the natural process of generating a magnetic field whenever a charge moves i.e. current. All currents generate magnetic fields.

The audio out of a transformer won't be that different, the waveform will remain the same, just the output will be at a different voltage level to the input. I guess the core could colour the waveform. This is something Graham will know a lot more about than me.

Thanks once again. Induction is what forms a magnetic field when a current flows, I will try to remember that!

[Jarno]

To ground, mostly, but it will also try to magnetize the core.
Since the transformer is merely a coil of wire on each end, for DC it is also a resistive path to ground.

As for the mains transformers in our powerlines. I used to live in a Nijmegen, and there's a company called Smit Transformers, they make them the size of a small house, so yeah, they are everywhere.
From the power generator --> transformer --> high voltage transport --> transformer --> distribution --> transformer --> your house --> transformer in PSU of modular synth

For signal transformers, primary impedance is determined by the thing you drive it with (guitar --> high impedance, preamp tube --> medium impedance, solid state --> low impedance (if you have a beefy output stage)). Secondary impedance is usually 600ohms (line).

There are plenty of other considerations though, like with microphone inputs and such.

[Babaluma]

Yeah, I just find it all fascinating, so thanks for the further info.

I've seen quite a lot of mention in audio circles recently of using 1:1 (600/600) transformers, just for the sound (and changing it by driving them with differing levels). Apparently different ones have different sounds. Good article by Alan Farmelo/Scott Hampton on building a transformer colour box in TapeOp a while back, you can read the full thing here:

http://www.farmelorecording.com/in-the- ... orm-audio/

[Jarno]

I made a sort of instrument preamp with a discrete opamp (API2520) based on a API line input card (API325) using a cheap Edcor 600:600 transformer. It certainly adds a little something to the sound (and subtracts something from it, in the form o a high roll-off). I recently also built a cleaner version of Scott Hampton's JFP preamp from TapeOp.

Nice article, I've saved it to read it later on (at work now ).

[Babaluma]

yeah i think those discrete opamps also have a sound of their own. there are some in my chandler tg2 preamp, with massive transistors with the name/make scratched off.

i was gonna get a dual hampton tube pre at one stage, but scott got unresponsive so i went with the chandler instead. have heard nothing but good things about scott's gear though, and it's certainly competitively priced.

[Babaluma]

aside from the possible perceived sound quality differences, why would someone generally use a 1:1 (or 600:600) transformer, just for the isolation?

[Graham Hinton]

I was hoping you would reply!

If you want me to reply to something specific it would be best to PM me. It was only by chance I saw your question on the main page while having a cup of coffee.

Also, thinking about their "insertion loss". My all passive EQ drops the signal by about 30-40dB, I am guessing due to the use of transformers and inductors.

That is more likely due to the EQ network. Active (valve) EQs normally have an input buffer to drive the network and an output/makeup stage.

I'm guessing some higher quality transformers will not saturate unless fed with silly levels that you would rarely find in audio applications?

It is amazing how much can be passed through a small transformer. The maximum operating level for audio transformers is quoted at 1% distortion, but I've had 160Vac coming out of small mic transformers. The only limitation on voltage is the insulation breakdown.

I'll go back to Wikipeadia and study the main transformer article again, and inductance/induction/inductor...

You might enjoy a documentary on the history of electricity that was on BBC recently: Shock and Awe. It's on YouTube:
Part 1 Part 2 Part 3

For signal transformers, primary impedance is determined by the thing you drive it with (guitar --> high impedance, preamp tube --> medium impedance, solid state --> low impedance (if you have a beefy output stage)). Secondary impedance is usually 600ohms (line).

Transformers don't exactly have an impedance as such, rather they are made to be suitable for matching one. They do have a dc resistance, but that is different, that is what you can measure with a meter. What determines the primary impedance is the secondary load seen through the winding's "gear ratio". While the turns ratio dictates what happens to the ac voltage and current the square of that ratio also dictates the impedance ratio, by Ohm's Law.

On a transformer with a 1:n turns ratio, Zin = Vp/Ip.
If a load, RL, is put on the secondary then Vs = Is x RL.
The primary and secondary voltages and currents are related by the turns ratio, n, so n x Vp = Ip/n x RL
or Vp/Ip = RL/n^2 = Zin.
So if you buy a 600:50k mic transformer all that really means is that n^2 = 50,000/600 = 83.3 and n = 9.128.

[Babaluma]

Thanks again Graham.

My Pullet EQ is totally passive, with a mic style impedance output of 1200 ohms, so I follow it with a mic preamp to get the gain back up to line level for feeding to my ADC.

Will start watching those documentaries now!

Found this great/easy to understand explanation from Scott Hampton at the end of the article I linked to above:

A Simple Transformer Explanation
A signal goes into one side of the transformer (usually the side driven is referred to as the primary), and the voltage that results on the output (secondary) is the input multiplied by the turns ratio (neglecting internal losses). The impedance ratio is the square of the turns ratio. Example: a 1:10 step-up transformer on a mic pre input would have a 1:100 impedance ratio. Meaning the voltage gain is x10, at a cost of increasing the output impedance by x100. A DI box is an example of going the other way, stepping the voltage down, to reduce the driving impedance. The lower the impedance a signal is being driven with, the less susceptible it is to noise. -SH

[Graham Hinton]

Aren't magnetic tape heads a form of inductor?

Well, yes they will have an element of inductance, anything with a winding does. Even a straight wire does, but it is very small and often ignored. Motors, solendoids, relays, loudspeakers all have coils and inductances which often means that they have to have special considerations for driving. They tend to "fight back" because the magnetic fields cause a reverse voltage (called back emf) to be generated. If a relay is driven by a transistor a diode has to be used to short out the voltage generated by the coil which would otherwise zap the transistor.

Magnetic tape heads are principally focussed magnetic field generators or receivers. The construction of record, sync and playback heads are optimised for each function, but like all these things they will work backwards too, just not quite so well.

[daverj]

Some great info here so far.

A little more info about "cores":

The windings of a transformer are wound around a core. The core can be made from various materials and can have various shapes. Those both change the characteristics of the transfer of energy.

The most common shapes of a core are variations on the "E" shape, and the toroid. Most rectangular transformers have cores that are shaped like an "E" with the windings around the center fork. Usually there is an end piece (an "I") or there are two "E" shapes facing each other, either butted together or interleaved.

The primary and secondary windings will sometimes be on top of each other, and sometimes next to each other. Or they might be wound together so the wires are interleaved. Each method can change the efficiency and the saturation points.

Even the shape of the wire itself can change the transformer characteristics. There are some transformers that use wire with a square cross section shape.

Toroids are just a different shape of the core that changes the saturation points, and changes the efficiency at different frequencies. Because of the shape they tend not to vibrate as much, and because of the curved shape they tend not to pick up external magnetic fields as much. Also because of the curved edges they tend to be a bit more efficient, since magnetic fields tend to also be curved.

Different materials are used in cores, and change the saturation at different frequencies. "Soft" iron is the most common material used, but alloys are sometimes added to change the characteristics. Since iron is a conductor it can induce fluctuating magnetic fields (eddy currents) which interfere with the transfer of energy. So usually iron cores are made as very thin pieces of metal that are laminated in layers to make up the core. The lamination insulates the layers from each other so you have essentially hundreds of thin cores each transferring a small amount of the current.

Transformers each have a specific bandwidth with low end and high end frequencies defined by their core shape, material, and winding style. Power transformers are optimized for low frequencies. Audio transformers are optimized for the audio range of frequencies. RF transformers are optimized for very high frequencies. RF transformers are often done with toroid shapes and often using ferrite materials because of the high bandwidth those offer.

The high end audio transformers used in pro audio equipment are engineered to handle the full range of frequencies that is in an audio signal. Smaller cheaper transformers, such as those used in the Real Ring Modulator are typically optimized for telephone frequencies, which are only a subset of the audio range. They typically drop off signals above about 3KHz, while pro audio transformers go 10 or 20 times higher in frequency. That's the reason the transformers used in the Real Ring Mod cost about $3 each while pro audio transformers usually cost $40-$100 each, and are much larger.

[Thonk Support]

Question: Do Transformers come in balanced and unbalanced flavours or are they naturally one or the other?

[daverj]

Transformers by their nature are balanced. The signal going into and out of them is the difference between the two connections on either the primary or secondary.

So a transformer can be used to connect two unbalanced signals, by grounding one wire of each side, or can connect two balanced signals by feeding the positive and negative in/out each end, or can connect a balanced to an unbalanced signal to each other by connecting each end appropriately.

[Thonk Support]

Gotcha, thanks Dave.

So Babaluma, don't some mastering engineers espouse unbalanced signal chains? Or is it more a Transformer-less signal chain perhaps?

[Babaluma]

Fascinating stuff daverj, thanks for all the info, and good question Monobass!

[Babaluma]

Gotcha, thanks Dave.

So Babaluma, don't some mastering engineers espouse unbalanced signal chains? Or is it more a Transformer-less signal chain perhaps?

Both I think.

Some people hate the sound of transformers in their chain (Bernie Grundman, who I think leaves the first and last transformers in the chain, but rips out all the intervening ones, I have no idea why he doesn't just remove them all...)

Some people say you should go unbalanced if possible (i.e. if you can get away with very short cabling), as it means you can therefore remove the transformers, again for a purer sound (Bob Katz talks about this in his book).

Some recording/mixing engineers love the sound transformers give and use them as much as possible (Joel Hamilton).

Me, I haven't tried all the possibilities, and it's difficult to know how much of my mastering chain's "sound" is given by the ten transformers in the chain, and how much is all the other discrete class A and A/B circuit topology and components, but it usually seems to sound better than what went in, at least to my and my clients ears, which is the main thing!

[daverj]

Balanced signals are best for running signals between equipment, especially over long distances, because they eliminate (or greatly reduce) ground currents from being introduced into the signals, and because they greatly reduce picking up stray signals (interference) along the way.

But they do make the input and output circuitry more complex, as well as the connectors and wires.

In the old days balanced almost always meant using transformers. Today's modern precision chips can create excellent balanced (differential) inputs and outputs without requiring transformers. By their nature transformers will "color" the signal somewhat by rolling off frequencies or saturating at certain signal levels and frequencies. With solid state differential drivers and receivers the main concerns are adding noise at low signal levels or clipping at high signal levels.

The concern in some modern studios with transformers is that if you go through a lot of them, the "color" they add is cumulative and can end up changing the signal quite a bit after running through a lot of them.

[Graham Hinton]

Question: Do Transformers come in balanced and unbalanced flavours or are they naturally one or the other?

It's really a function of the way they are connected. A common configuration is to have two primary windings and two secondary windings and these may be connected in series or parallel to give different turns ratios. That is why you often see mic amps with 300ohm/600ohm/1200ohm inputs because it has a transformer with two 600 ohm primary windings.

By connecting two windings in series you also have the centre connection available and this is equivalent to a centre tap. In vintage valve equipment the only way to get a high impedance valve output to drive a low impedance load, like a line or a loudspeaker, was to use a transformer and these were often quite large and had a centre tap in the primaries and secondaries. Two valves drove the transformer, one on each half cycle and for symmetry they usually grounded the centre tap of the output giving equal and opposite waveforms about ground. This has led to the false assumption that balanced means equal and opposite voltages when really it means equal impedances. Cancelling interference does not require complimentary signals, it only requires a differential input and two wires assumed to pick up equal interference, one of which may be signal 0V (not ground 0V). (I say assumed because it won't be equally picked up if the source is very close to the cable pair.)

Now think what happens if you have a grounded centre tap output and the balanced line goes to a patchbay where it might meet an unbalanced input. One of the signals will be grounded which will short out one half of the output winding and rapidly turn it into a fuse. For this reason it is best to disconnect the output ground centre tap and let it "float". Vintage ribbon microphones often had grounded centre taps and this is why you hear scare stories about connecting them.