Mic input transformer grounding questions

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
MisterCMRR said:
Sorry guys but I got too busy to even check this site last week!

First, in answer to CJ's question about the Jensen 10K61 output transformer, its core is an 80% nickel alloy but the "secret sauce" (which shall remain a Jensen trade secret) is in the custom profile of the heat treatment (annealing) for the core material. It both increases magnetic permeability (inductance, indirectly) and lowers THD at low signal levels (the magnetic hysteresis part of the curve, as opposed to the magnetic saturation part at high signal levels).

Regarding the "anti-transformer" camp, I could go on and on about this but I'll keep it brief.  Transformers, if properly designed and fabricated, are amazing performers.  But, as the late Deane Jensen used to say, "first I'll tell you why you need a transformer, then I'll tell you why you need a Jensen." 
RIP Deane was brilliant and generous with his wisdom. I had some great conversations with him back in the 80s.
With poor transformer designs, I'm often tempted to say "If you can't afford a good one, you should probably do without!"  Jensen isn't the only source of pretty good audio transformers, but I know of no others that do just about everything extremely well. 
I am probably the poster boy for anti-transformer sentiment around here. Of course good transformers are good but expensive. I was born with a sharp pencil in hand and over the decades found many cases where the function can be performed as well (or better) for less cost. 

About the only market I experienced where customers insisted on true transformer outputs was the constant voltage (70-100V) fixed install market. My suspicion was that preference was probably related to transformer coupled speaker lines still working after one leg or the other is shorted, as long as both are not shorted. In that industry service calls are expensive and problematic so they were willing to pay a premium to avoid even a few service calls. They were also resistant to change, even though output transformer-less constant voltage amps could be a bunch cheaper (several companies tried with limited success).
The best aspect of a transformer is, of course, galvanic (electrical) isolation. Since the windings float electrically, the transformer has no idea what "ground" means. That gives it excellent noise rejection when used at inputs and an "I don't care what you connect me to" attitude at outputs (will drive a balanced input or an unbalanced (grounded) one with equal ease and no risk of damaging the output line driver stage.
yes, most forgiving for extreme conditions. Good for live sound where you can encounter dodgy power and noisy grounds.
Now comes the bandwidth issue. The worst shortcoming of most other transformers is high THD at low frequencies and high levels (this is where most of the energy in contemporary music is, in fact). The others do it because physics tells us that high levels of low frequencies is what creates the strongest magnetic flux the core material will ever see ... and it simply takes more core material (a bigger core stack) to do that ... no escaping that. So a transformer barely bigger than a postage stamp (like that in the IL-19, for example) simply will have double-digit THD at 30 Hz and +20 dBu signal peaks.
Imagine constant voltage audio power output transformers. With step down transformers on every speaker. The fixed install market is price sensitive so rarely oversizes the magnetics for extended LF performance. This can get dicey when the customer wants to turn up the bass boost, on top of a loudness contour when feeding a background music system.  A saturated output transformer is not a friendly amplifier load. Not to go too far down this veer, back last century I patented a cheap bass clipper (that didn't suck) for use in this market so customers could turn up the bass to 11 and not saturate output and speaker magnetics when the audio got louder.
Then there's phase distortion. As Deane wrote many years ago, phase shift is not necessarily phase distortion. Phase shift, as long as it changes linearly with frequency, is no different that moving your head a few inches toward a sound source. But phase shift that isn't linear with frequency IS phase distortion, because different frequency components of complex music will have different arrival times ... and change the timbre of the sound. All that being said, some may find it surprising that the most serious phase distortion occurs at low frequencies.  Marshall Leach wrote a paper back in the early '80s about this. He concluded (and I agree) that to keep LF phase distortion low at 30 Hz, the -3 dB frequency of a system (including the speaker and mic) needed to be over a decade lower. It affects the timbre of kick drums for example ... and explains why I personally dislike bass-reflex speaker systems (they behave as 4th-order high pass filters, while sealed boxes behave as 2nd-order high-pass filters - and, in general, the steeper the filter, the worse its phase distortion).  Getting to the point here ... it's why Jensen transformers have what seems to most over-designed low-frequency response. A typical Jensen design will have a low-end -3 dB point around 0.2 to 0.5 Hz and the reason is ultra-low "deviation from linear phase" or DLP as it shows on the data sheets. These parts have an incredibly tight and clean bottom end.
I am not sure I follow... are you talking about something other than minimum phase response from real poles?  I routinely set DC blocking poles octaves below the pass band just to avoid the accumulation of roll offs adding up in series and impacting the audible pass band. 
Regarding HF bandwidth, I'll simply say that more is rarely better. Going back to the Leach paper, at high frequencies. the shape of the roll-off is crucially important to minimize phase distortion. A class of low-pass filters called Bessel have the most linear phase/flat group delay response. Therefore, the top end of a Jensen design is made to conform to a 2nd-order Bessel response. As expected, such filters have near-perfect square-wave response. You can see this on the data sheet's DLP plot vs frequency, staying typically within 2-degrees of linear phase from 20 Hz to 20 kHz.  It's also the reason that the recommended load on mic input transformer secondaries, whether just a resistor or an R-C network be followed precisely.  All that being said, I much prefer listening to a clean, phase-distortion-free bandwidth of 20 kHz rather than a 100 kHz bandwidth. And there's another reason to intentionally restrict HF bandwidth ... it's called "spectral contamination" in a Deane Jensen/Gary Sokolich AES paper of that name. It describes a complex inter-modulation distortion that is made much worse when ultrasonic components are included in the signal applied to any active device (transistor, op-amp, vacuum-tube, etc.). These components are most often harmonic distortion products of previous active devices in the signal chain, DAC clock residue, or even stylus chatter in LP playback.  This would mean that the cleanest possible signal chain would include an intentional low-pass filter just ahead of any active stage.  I could go on further, but I'd like to end my "pro transformer" rant with an observation from 25 years of talking with Jensen customers.  Jensen's "transformers in a box" product, ISO-MAX, are widely used to fix ground loop issues in very high-end audiophile and home theater systems. At least half the customers I talked with say something like "Not only is the buzz gone, but my system has never before sounded so "clean" with "open space" and "air" that lets me identify each instrument ... it's amazing."  I usually just said thanks rather than trying to explain that spectral contamination literally clutters the audio spectrum with non-harmonically-related new spectral lines ... it's very hard to describe but some call it a "veil" that's lifted when the transformer is inserted in the signal chain.  And, for what it's worth, the most effective place to insert one is just ahead of the power amplifier (which makes sense to me since those amps have, in general, more inherent open-loop non-linearity and lower slew rates than low-level amplifiers.  Anyway, I have to stop myself ... but don't give all transformers a bad rap just because a lot of crappy ones are in wide use. Overall, a good transformer can be a tremendous benefit.
Yes, an obvious benefit of transformer input is to scrub RF out of the audio path, so slower active stages following the transformer are less likely to rectify the hf signals they couldn't keep up with. This can also work for output stages where slow drive electronics can experience rectification from RF coming back into the output.
And please accept my humble thanks for the warm welcome I've received here. Maybe I'll get to meet some of you in person someday. Incidentally, I'll be teaching a workshop on power and ground issues for live performance venues (and possibly another class aimed at equipment designers about internal grounding of power supplies, shield connections, decoupling, etc.) at the AES convention in NYC this October.  Kind regards,  Bill Whitlock

What do you think about adding a transformer to create some euphonious nonlinearity? A surprisingly popular concept in some circles. I guess for EFX anything goes just not my preference.

Welcome again, and I hope I didn't come off as argumentative or disrespectful. I've used lots of audio transformers in the markets where customers insisted on them. I stopped using transformers in console mic preamps by the late 70s when bipolar devices got quiet enough for transformer less designs to be competitive for a fraction of the cost. (I'm cheap).

JR
 
JohnRoberts said:
About the only market I experienced where customers insisted on true transformer outputs was the constant voltage (70-100V) fixed install market. ... They were also resistant to change, even though output transformer-less constant voltage amps could be a bunch cheaper (several companies tried with limited success)
Not so much today. Most of the fixed install I supply equipment to today use xfmr-less amps.
And sometimes the speakers themselves are xfmr-less. 70V into 16 ohms is "only" about 300W. Distributed sound is not what it used to be...
 
JohnRoberts said:
.yes, most forgiving for extreme conditions. Good for live sound where you can encounter dodgy power and noisy grounds.
JR
Extreme can be rather common in pro audio. It is worth remembering that at the dawn of pro audio, the most common application was broadcast which meant in may cases the audio system had to survive nearby transmitter levels in the KW region. These days that has been replaced by the prevalence of mobile phones, wifi and Bluetooth. A major culprit is SMPS. Althogh these will meet FCC and CE standards for emissions and conducted interference, they are allowed to squeal as much as they like in the audio band. Laptop external power supplies are a particularly bad culprit.

Cheers

Ian
 
abbey road d enfer said:
Not so much today. Most of the fixed install I supply equipment to today use xfmr-less amps.
And sometimes the speakers themselves are xfmr-less. 70V into 16 ohms is "only" about 300W. Distributed sound is not what it used to be...
It was about 30 years ago when Peavey entered the fixed install market and they were conservative and not very receptive to change (or Peavey).

In our typical sharp pencil approach we tried to use auto-formers in place of the typical constant voltage output transformer, to improve efficiency and reduce cost. The auto formers actually spec'd out a little better but they were not impressed. We were widely rejected until we provided true floating transformer outputs.

In fact the thing that finally opened the door for Peavey was the media matrix computer controlled audio systems for big stadium installs. The DSP based Peavey system could eliminate several equipment racks full of processing gear and the wiring (with labor) to connect everything together and make it work. The price advantage was compelling.

When the alternative to using Peavey was losing the bid for being too expensive they finally held their nose and tolerated us.

JR
 
JohnRoberts said:
In fact the thing that finally opened the door for Peavey was the media matrix computer controlled audio systems for big stadium installs. The DSP based Peavey system could eliminate several equipment racks full of processing gear and the wiring (with labor) to connect everything together and make it work. The price advantage was compelling.
I spoke with Hartley P. at the time, offered him to distribute media matrix in my country; he was not interested, said he wanted to set up his own subsidiary. Since it never happened, I surmise he figured it was too costly and set up his UK operation. I have never seen a single piece of Media matrix installed in my country...

When the alternative to using Peavey was losing the bid for being too expensive they finally held their nose and tolerated us.
I would think xfmr-less CV amps are accepted now for the same reason.  :)
 
ruffrecords said:
Extreme can be rather common in pro audio. It is worth remembering that at the dawn of pro audio, the most common application was broadcast which meant in may cases the audio system had to survive nearby transmitter levels in the KW region. These days that has been replaced by the prevalence of mobile phones, wifi and Bluetooth. A major culprit is SMPS. Althogh these will meet FCC and CE standards for emissions and conducted interference, they are allowed to squeal as much as they like in the audio band. Laptop external power supplies are a particularly bad culprit.

Cheers

Ian
I specifically mentioned live audio, because recording and to some extent broadcast has the luxury of more than a brief sound check to ring out the system. Prudent solid state design can generally manage the majority of recording applications. Of course premium products still include transformers to give the customers warm fuzzy feelings. 

I have heard horror stories about broadcast studios colocated with high power transmitting antennas...  :eek:

JR

 
abbey road d enfer said:
I spoke with Hartley P. at the time,
The last time I sat down with Hartley face to face, he had me escorted out of the building by security guards.  :eek:  That was almost 20 years ago. 
offered him to distribute media matrix in my country; he was not interested, said he wanted to set up his own subsidiary. Since it never happened, I surmise he figured it was too costly and set up his UK operation.
The British distribution from Corby was before the Media Matrix acquisition and AA market entry. IIRC before Corby was set up, distribution was handled from Germany(?), with containers landed and broken out for distribution in Belgium or something like that (not my area of involvement).

After I was kicked off the island I recall hearing about Medimatrix Software types being set up in a development /support lab in the UK. I have no idea what they are doing these days.
I have never seen a single piece of Media matrix installed in my country...
Neither have I  ;) . Reportedly media matrix has been well accepted around the world for very large installs. I suspect in the decades since MM emerged, some credible competition has surely evolved. In fact if MM was patented those patents would have expired by now.
I would think xfmr-less CV amps are accepted now for the same reason.  :)
Transformer-less CV amps are a no brainer from a design engineer perspective, the customers were less open minded. Of course cost always matters and that is/was always a price sensitive market, so logically over time the cheaper solution should gain traction (but a small company could go out of business waiting for the market to catch on). 

JR
 
Awesome information Bill!  :D :D :D

i have a few more questions,

at what frequency is leakage inductance measured in order to determine it's effects?

do you have a good circuit for a B-H tester for DIY'ers with X-Y scopes?

and why are there partial layers in the 4:1 output with the top secret core?  thanks!  cj
 
My question would be how could we evaluate the distortion characteristics of different core materials ,in a meaningfull way , no matter what you do it seems like comparing apples and oranges .
Counting THD isnt a very meaningfull way ,as harmonics tend to become more objectionable  the higher you get , and topology has a huge impact on harmonic spread , a single ended triode ,transformer coupled will produce nearly pure second HD which might be useable ,depending on source upto 5% THD , on the other hand an op amp driven to 5% THD is liable to destroy itself and hurt your ears .
The transformers abillity to provide some sponginess/ limiting/compression during unschedualed peaks which always occur with real life sounds ,even if these peaks only last microseconds ,for me is invaluable .
 
Back
Top