Transformers

[asknone9]

I'm new here, and I've run into some technicalities that I would like to clear up. What is the functional purpose of a transformer? Why a would a mic like the Blue Mouse have a transformer and transformerless design? What are the benefits/disadvantages of having a transformer? Thanks for the help guys. Don't be afraid to throw whatever technical information you have at me. If I have to research just to understand what you're talking about, I will.

[Matt Syson]

Hi
Transformers are used to 'convert' AC power from one voltage and current level to another voltage and current level.
The main proviso for this is that the POWER remains constant (minus a small amount 'lost' in heating and magnetising the transformer itself, usually a couple of percent or less of total power).
This notion of being able to 'transform' what you have to something perhaps more useful is one of the best attributes.
They can also 'isolate' the secondary connection from the primary circuit which is of significant interest to audio users particularly when equipment is spaced widely and may be on differing mains supplies. (elimination of 'ground loop hum)
There are of course downsides, cost, limit to frequency range (within a design group) distortion and cost. They will not by themselves pass DC power from primary to secondary.
Audio users tend to think of a transformer adding 'colour' due to the distortion added and possible 'saturation' effects (the magnetic core cannot be magnetised any 'harder' beyond a certain power level so the output signal deviates significantly from the input signal).
For this reason you tend to find groups of equipment without transformers being referred to as 'clean' sounding' and with as 'coloured' although the likes of Jensen and Lundahl are some of the 'cleanest' sounding transformers. These are of course not cheap and around 50 Dollars or more apiece equipment costs mount up rapidly compared to a 'direct' connection with the extra components necessary for a 'balanced' interface coming in at less than a Dollar.
Of course this is a bit of a 'cheat' as if you add the extra components to make an electronic stage as relatively 'bulletproof' as a transformer the picture changes significantly.
Rummage around on GS and other places and you will end up with lots of 'opinions'. If you really want to know what is going on in a transformer look to the 'traditional' textbooks for equations and so on.
Matt S

[brianroth]

Transformers have pluses and minuses:

Plus: galvanic isolation:

Galvanic isolation - Wikipedia, the free encyclopedia

Minus:

additional "stuff" in the signal path which adds distortion, limits frequency response.

Bri

[asknone9]

Thanks guys! My understanding is bounds clearer than it was before. I have another question though: at what distances (as far as spaced equipment is concerned) does a transformer start to become necessary and/or useful, when considered for only this purpose of course.

[Matt Syson]

Hi
Difficult to say but 'out of the area', say a couple of hundred metres and the advantages of isolation can make plugging gear in almost trouble free whereas you MAY start to run into ground differential problems with non transformer solutions.
There are however many ways to configure equipment so it is never a 'hard and fast' answer.
Fully electronic in / out CAN be made bullet proof, and better performance than a transformer.
Not many would stand and even be operating quite well with say 300 Volts DC superimposed on the wiring whereas a transformer would most likely be quite happy.
Matt S

[asknone9]

I see, so many studios would probably find the necessity of a transformer for the purpose of increased signal distance irrelevant. Generally. Also, the term "bulletproof" seems to encompass other modifications that can be applied to an electronic stage in order to give it more or less the same effect of a transformer. Am I correct, or is that assumption off?

[Matt Syson]

Hi
I was using 'bulletproof' to describe attributes that make the stage more reliable and immune (as far as is possible) to pickup of unwanted signals outside the audio band.
Many pieces of cheap gear will die if subjected to a decent size 'static' shock or if it is accidentally shorted for a while.
Putting 48 Volt phantom power onto the output of a CD player for example will most likely destroy the output capacitors on the CD player if left for more than a short time.
Matt S

[drtechno]

microphones first had transformers to match impedance
(RCA 77-CI and 77-BI capsule types had two output impedance models MI-4044B and MI4044A. the difference was the output transformer )

Some people, after the invention of phantom power, protected certain types of mics (ribbons) from destruction by using transformers for isolation.

the advantage of the tranformerless mic has over the same mic with the transformer is there is one less component that could degrade the signal ( insertion loss ), but has a great disadvantage as there is no isolation from phantom power if present, and in some cases, that could damage the mic or the pre.

The manufacture will warn you about the transformer less design hazards (if any) in your product manual. You can even contact them directly to ask questions about your product too.

[foldback]

Transformers continually get a bad rap because people don't take the time to understand them completely.

Galvanic isolation is huge. This means two pieces are electrically communicating (AC signals) but not directly electrically connected. This is how ground loops are broken. It's darn close to magic :-)

Impedance matching: This is the feature that allows a 50 ohm mic output to connect properly with a 600 ohm input.

Noise Free Gain: The 1:10 microphone step up transformer in the API 3124 is one of my favorite sounds in the whole universe. Noise free as in no active components required.

Limited Frequency Response? Not really. A precision quad filar transformer with high-nickel alloy laminations has three dB down points around 2 Hz and over 200 kHz depending on how it's manufactured. In between the frequency response is flat like a straight line. Distortion at 20 Hz on a proper transformer component is under .01% even at +30 dBu.

Sound Color: Custom alloys used in the construction of the laminations are capable of producing the much sought after "vintage" sound. Some describe this as a "forward" voice.

The only thing that really sucks about a transformer is the price. The high-nickel lamination material is very expensive. Fortunately, the iron used in the colorful vintage components is still pretty inexpensive.

[memphisindie]

This is why there is no need for a decoupling cap, DC shunt to ground, on transformer gear. Transformers don't pass DC without help.

[asknone9]

Wow thanks a lot guys. I'm 100% understanding this with all of your explanations. I'm curious, though, about the safety practices one should be aware of in order to protect transformerless gear, in case there's something I haven't taken note of. I will take a look at those areas of the manuals I have too.

[martymart2]

I would add to the attributes of transformers: line balancing!

Historically, the old ribbon mics were inherently balanced, but had an unusably low output. A transformer with turns ratio of 1:33 was used to drastically increase the output voltage, maintain the balanced line output and to better match equipment impedances.

The old tube condensor mics were inherently unbalanced plate output designs of very high impedance. The output transformers actually converted the signal to a balanced output, reduced the signal voltage and decreased the output impedance to match the standard 500/600 ohm line.

Early transistor designs still used transformers to allow cheaper unbalanced amplifier topology. Early transistors were expensive.

For the last 30 years, or so, electronically balanced circuits have undercut transformers in cost and size in most budget gear. Balanced lines could now be interconnected with moderate success without the use of expensive and large transformers.

mm

[Matt Syson]

Hi
Protection of transformerless inputs is relatively easy to say 50 or 100 Volts by using non polarised capacitors in the input (with a bit or resistance) and 'clamp' diodes to prevent spikes taking it beyond the circuit power rails.
Filtering to prevent HF getting in.
You can provide protection to mains voltage but the capacitors get big and expensive.
Outputs are more tricky and again can include clamp diodes and possibly fuses as there is not much scope for large current limiting resistors as the output is normally required to have an impedance of 10 - 100 Ohms so it can drive long cables without serious degradation. 600 Ohms was an earlier 'standard' and is of course perfectly useable but that was from days when 15KHz was the main requirement with 30K a 'luxury'.
Again RF filtering should be included to prevent 'interference' coming back into an amplifier stage which would then be demodulated.
If course nothing is totally destruction proof but beyond the basic circuit things start to get expensive and many companies only do the minimum, partly because it starts to erode the 'in band' specifications. To say your preamp can manage 5 Hz to 200KHz looks wonderful on paper. In practice if it picks up 'radio' easily, it will get a reputation for being 'rubbish'. 198KHz is the 'radio 4' frequency in the UK and there is a similar frequency used in mainland Europe for frequency reference (and radio) with very powerful transmitters. Not so much fun if your mic amp is only 2dB 'down' at this frequency as it will be blasted into the following gear even if the mic amp itself is happy.
Matt S

[mixedupsteve]

Originally Posted by martymart2

I would add to the attributes of transformers: line balancing!

Historically, the old ribbon mics were inherently balanced, but had an unusably low output. A transformer with turns ratio of 1:33 was used to drastically increase the output voltage, maintain the balanced line output and to better match equipment impedances.

The old tube condensor mics were inherently unbalanced plate output designs of very high impedance. The output transformers actually converted the signal to a balanced output, reduced the signal voltage and decreased the output impedance to match the standard 500/600 ohm line.

Early transistor designs still used transformers to allow cheaper unbalanced amplifier topology. Early transistors were expensive.

For the last 30 years, or so, electronically balanced circuits have undercut transformers in cost and size in most budget gear. Balanced lines could now be interconnected with moderate success without the use of expensive and large transformers.

mm

Almost all gear is still unbalanced inside the box. The balanced signal is first unbalanced by a transformer or another method and then rebalanced on the way out. This is to allow balanced interconnection as well as all the other functions mentioned.

[foldback]

When it comes to common mode noise reduction a proper transformer wins by 20 to 30 dB over the best transformerless designs.

A proper Mu-Metal enclosure provides 30 dB of magnetic shielding to keep slower moving magnetic disturbances away from the transformer (that's the shiney metal can you may see). In some cases the first package may be placed inside of a second Mu-Metal enclosure for 60 dB of shielding (mission critical apps like nuke subs).

Transformers use Faraday shields to limit bandwidth so that those pesky 198k radio frequencies can't enter the audio circuit.

Modern semi conductor components allow circuit designers to virtually eliminate capacitors from the audio path. A well designed transformer used at the input stage means the audio signal path that follows can be direct coupled (no caps = lower distortion). Bipolars are just two in series for double the loss, get rid of these too.

In the US the 600 ohm transmission system grew out of the telephone industry and helped connect cities across many miles, in those systems transformers were the devices that kept the impedances right and blocked out the common mode noise. The 300-3K frequency response was purposefully designed and not a fault of the transformers.

Some people make their living as repairmen (replacing dried out caps) so they don't like transformers because they don't inherently crap out with age.

Good transformers cost a lot, they are expensive and time consuming to make. A $1.00 opamp is cheap, a $60 transformer is not.

[memphisindie]

That's a heck of a good analysis, but, it may leave people feeling that there's only one way to go and that's not the case, there are well designed and crappily designed circuits on both sides of the isle.
They do sound different.
I like both but for different things.

[asknone9]

Interesting point, foldback. I hadn't heard that assertion yet. And thanks for looking out memphisindie. It's safe to say that I have realized that the most important aspect is how it sounds, which is something I probably should have assumed at this point, but with all this technical info I know now, I will have no reason to be unnecessarily biased when choosing gear, so long as the quality is there. I've picked up a book on basic circuitry, too, so that should give me a little more to chew on.

[Matt Syson]

Hi
Foldback has been a little 'tabloid headline' over some assertions.
Semiconductor balanced input circuits can be way better than transformers AND with lower distortion. OK many are not so good but the 'everything is rosy in the transformer garden' is also a bit of a myth.
They are not 100 percent efficient and they do have their own distortions, both at low level and high.
By the time you have created a semiconductor input stage that is robust, RF immune and stable over the long term it can approach the cost of a mediocre or even expensive transformer, the fact is many manufacturers don't really try because the cost is not worth the benefit. An input 'balance' or CMRR of about 40dB or more is usually sufficient for 90 percent of interfaces, particularly if this is at powerline fequencies as 'hum' is the most frequent complaint.
Since capacitors are virtually a necessity in any gear they cannot be eliminated, the caps on the power rails ARE in circuit at least to some extent although probably not so influential in a totally symmetrical (differential) system.
To hear some commentators bang on about capacitor 'distortion' it is suprising we listen to anything as there are often hundreds used between a microphone capsule and a loudspeaker coil.
To achieve this with transformers would be an interesting challenge.
Matt S

[foldback]

Originally Posted by Matt Syson

Hi
Foldback has been a little 'tabloid headline' over some assertions.
Semiconductor balanced input circuits can be way better than transformers AND with lower distortion. OK many are not so good but the 'everything is rosy in the transformer garden' is also a bit of a myth.
They are not 100 percent efficient and they do have their own distortions, both at low level and high.
By the time you have created a semiconductor input stage that is robust, RF immune and stable over the long term it can approach the cost of a mediocre or even expensive transformer, the fact is many manufacturers don't really try because the cost is not worth the benefit. An input 'balance' or CMRR of about 40dB or more is usually sufficient for 90 percent of interfaces, particularly if this is at powerline fequencies as 'hum' is the most frequent complaint.
Since capacitors are virtually a necessity in any gear they cannot be eliminated, the caps on the power rails ARE in circuit at least to some extent although probably not so influential in a totally symmetrical (differential) system.
To hear some commentators bang on about capacitor 'distortion' it is suprising we listen to anything as there are often hundreds used between a microphone capsule and a loudspeaker coil.
To achieve this with transformers would be an interesting challenge.
Matt S

Matt, I'm not sure what you mean, your text is so thick with weasel words, generalizations and ambiguous blather.

No myth, just facts, my only assertion is this, transformers don't kill audio, bad circuit designers do! There are crap transformers and there are good ones. All transformers are not created equal.

Transformers fell out of vogue because of cost not inherent poor performance. The 1512 by THAT is a popular example of a "semiconductor input stage that is robust, RF immune and stable over the long term". I just checked on Mouser, it's $5.47 in singles (and it's used in a lot of popular mixers because it's good enough, simple and cheap). A single Jensen JT-115K mic input transformer is going to set you back $76.36 according to their web site. It's easy to see why a console filled with Jensens is going to cost a lot more than "transformer-less".

When the lamination material in a transformer is made with high nickel content it costs A LOT more than iron. If you want a good low distortion transformer, high nickel content is required.

Transformers are not inherently frequency response limited in the audio range (neither are vacuum tubes). Some people have built inferior gadgets with these components while others infer that transformers (and tubes) are responsible for the resulting "less than stellar" performance. Too many inline transformers can degrade performance, again, it's all a matter of proper application and circuit design. Rupert Neve is an example of one designer who knows how to design with transformers, in fact they are integral to his designs and the mojo of his products sound.

Transformers have several parts that can be optimized for specific performance criteria. The metal core material used in a transformer has a huge impact on distortion and signal level handling ability. The external metal package and the internal shields are critical components too.

For the last year I've been in the aerospace sector of Southern California (El Segundo). We have noise from big computer data lines, huge power distribution systems, and some top secret sources. I don't know what the heck you're working on, 40 dB of CMRR is total crap. The active mic-inputs on my little M-Audio Buddy are better than that and it's a toy compared to my API 3124. A good mic pre transformer can achieve 115 dB CMRR at 60 Hz and provide 20 dB of voltage gain. The mixer in the Sisters of St. Mary presentation facility had to be retrofitted with transformers because of noise generated when the nearby hospital dumped its computer data across the site. The preamps in their mixer were capable of over 55 dB CMRR and that was not enough!

I don't make my living doing repairs but I've been doing repairs for clients since 1973. All of my clients (from poor musicians to wealthy corporate types) want the best quality, the difference between "most" and the "best" makes a big difference here (I can throw a rock from my office and hit a Disney studio).

"Capacitors are a necessity in some gear" would be a more accurate and true statement. There are modern components available that don't have output DC offset problems, these make coupling caps unnecessary, the resulting signal path can be shorter, distortion is reduced, cloudy sound goes away, bass and high frequencies sound real not exaggerated. Modern opamp components can swing the full power supply rails across 600 ohm low impedance loads without excessive distortion, into a good output transformer that can yield over 30dB of output level and that kind of headroom can come in handy when mixing OTB. This level of circuit design means that when the kick drum and bass guitar both "hit" together the sum amp stage won't burp out a distortion peak, or worse, flat line.

When I want color or vintage sound, I go to the patchbay and insert any number of signal processors, otherwise, I like hearing the sound I'm working on as accurately as possible. When quality transformers are combined with modern active components and the caps are removed it's like taking off ear muffs. I could mix a record with ear muffs on but it sure sounds better without them.

The signal path, from microphone input, to voice-coil output, in my studio does not contain hundred of capacitors, if I included the power supply caps I'd be surprised if there are more than a couple dozen.

Maybe it's time to take a break from GS troll duties and develop real hardware to match the imaginary cartoon products on your web site. If you actually have an Audio Precision system (as you've asserted in some of your other postings) I find it hard to see why you're making these foolish and unsubstantiated generalizations about circuit design.

[foldback]

Originally Posted by memphisindie

That's a heck of a good analysis, but, it may leave people feeling that there's only one way to go and that's not the case, there are well designed and crappily designed circuits on both sides of the isle.
They do sound different.
I like both but for different things.

Thanks for your kind words. I like the appropriate use of transformers, nothing more.

It seems to me that too many times, audio transformers get a derogatory inference because they are an old technology.

Transformers don't have to be bad, that's the only point I was trying to make along with highlighting the often over-looked positives they can offer.

Back to making good music.

[foldback]

Originally Posted by asknone9

Interesting point, foldback. I hadn't heard that assertion yet. And thanks for looking out memphisindie. It's safe to say that I have realized that the most important aspect is how it sounds, which is something I probably should have assumed at this point, but with all this technical info I know now, I will have no reason to be unnecessarily biased when choosing gear, so long as the quality is there. I've picked up a book on basic circuitry, too, so that should give me a little more to chew on.

Not sure what assertion you mean. All the scientific data is available on the Jensen web site along with Audio Precision test results.

If they are educated your ears are the best tool for evaluating things. Uneducated ears can lead to poor purchase decisions tho. Once I bought an Allen and Heath recording desk that sounded quite nice at the demo but after it lived in my studio for 6 months I became quite annoyed with a harsh quality it had. I could not hear that until I'd worked on that desk for many many hours training my ears. Finally I had to sell it because I could not stand the sound. Other people would look at me like I was mad (they could not hear it) but it was driving me insane and it was very real distortion. I've been a lot more cautious since then.

All the best and good music to you.

[FredYeah]

You guys are all wrong...

Transformers - Wikipedia, the free encyclopedia

[memphisindie]

Originally Posted by foldback

Thanks for your kind words. I like the appropriate use of transformers, nothing more.

It seems to me that too many times, audio transformers get a derogatory inference because they are an old technology.

Transformers don't have to be bad, that's the only point I was trying to make along with highlighting the often over-looked positives they can offer.

Back to making good music.

Sometimes and often they are a stellar choice.
There are some very good to great transformer-less designs too, but, you really need to need one to get one. For most music a good to great transformer design will yield excellent, desirable, sexy, results. I would go as far as to say that they can be more musical than transformer-less BUT sometimes you need the clean fast sound of a transformer-less design, and they're both great in the extremely high headroom designs. For me, the highs set them apart and the low mids and lows really spell out a transformer's cool spots.
I hated them in the GI Joe crossover series though.

[Matt Syson]

Hi
Quote
[(I can throw a rock from my office and hit a Disney studio). ]
Unquote.
Fortunately I can't, and this makes me happy.
100 Volts common mode into a 'THAT 1512' without the support of other components won't impress it much.
The OP wanted to know more about transformers, OK much of what I have written are generalisations and of course there are extremes in everything but it is not intended to be a 'master class' on the finer details of transformer design, I am perfectly happy for others who are that interested in transformers to take over.
Personal slurs are unnecessary and I don't see a great deal of real information in much of your writing on this thread.
Matt S

[memphisindie]

Yes, it was definitely uncalled for.

[foldback]

Originally Posted by Matt Syson

Hi
Quote
[(I can throw a rock from my office and hit a Disney studio). ]
Unquote.
Fortunately I can't, and this makes me happy.
100 Volts common mode into a 'THAT 1512' without the support of other components won't impress it much.
The OP wanted to know more about transformers, OK much of what I have written are generalisations and of course there are extremes in everything but it is not intended to be a 'master class' on the finer details of transformer design, I am perfectly happy for others who are that interested in transformers to take over.
Personal slurs are unnecessary and I don't see a great deal of real information in much of your writing on this thread.
Matt S

Matt
I apologize profusely. A good transformer can be a tasteful thing but being a dick is never a good thing. I really mean this, I'm sorry.
Mark

[Marik]

Originally Posted by foldback

Noise Free Gain: The 1:10 microphone step up transformer in the API 3124 is one of my favorite sounds in the whole universe. Noise free as in no active components required.

Well, this one is not as simple as that. Any resistance is a noise. The source resistance together with the pri copper DCR transformed with 1:10 ratio is actually becoming 100 times larger. On top of that we have added copper DCR of the secondary and as a result in many cases we already cannot ignore the added noise. Indeed, there is also a signal gain, but the resistance noise increases in a faster rate.
So to answer the question, no, this gain is not Noise Free.

Best, M

[Matt Syson]

Hi Mark
Thank you.
Generally:
I think we all know that in the world everything COULD be made to fantastic standards of quality but probably unworkable as a business model. The buying public simply won't buy at the price necessary.
I don't like the idea of 'military spec' on the grounds that it is overpriced and they break it in 'no time flat' so could happily be built to a slightly lower spec but more readily available.
GP (Jeep) as used in WW2 had an average 'life' expectancy of about 48 Hours I believe.
This is the GEEK part of the GS board please keep it forward looking as opposed to the 'slag off XXX product because it is fashionable to do so. The Behringers and Mackies probably make better profit margins overall than Neve and other 'high end' gear. I suspect they also manage to 'satisfy' a significant segment of the market.
Matt S

[asknone9]

Well, I haven't been able to check back in a few days but I see that you guys have gotten into some things amongst yourselves. The interesting thing I had mentioned was your reference to transformers being superior by 20-30db in noise reduction in some cases, foldback. I definitely didn't intend for this to be a "masterclass" either, but as I read through your posts, I find that the terms used are becoming more and more familiar to my understanding, so please don't feel you need to limit yourself on my behalf. I do my best to decode what I don't get. Either way, be assured that the one thing I realize is to not take one fact, like the one that mentioned above, and use it religiously to influence my decisions on gear. You guys clearly have spelled out the good and bad of both sides. Again, I appreciate it all. I am, first, a musician, but I am a musician who has a craving for understanding the intricacies and technicalities of sound. In the end, I know, however, that I will be ultimately satisfied with what sounds good for what I am trying to accomplish, and that sound may very well come from less-than-prestigious gear.