Long tailed trio

[JohnRoberts]

Rather than add another veer to the differential thread, here is a different veer in a different place...

It's a pitty that there are not more fully differential opamps suitable for audio around; particularly for mic preamps and converters, where there is a fully differential signal path from input to output, they are of great use.


Samuel

We've engaged on this point before and I am not a fan of fully differential paths inside gear. For simple paths, it generally doubles the cost while providing little benefit over a single ended path. For more complex circuits the degree of difficulty and cost way outpaces any benefit.

These ICs seem to have a fairly specific area of application..
=====

That said, in the spirit of all differential all the time, here is a circuit I scratched out (but never melted solder over) several years ago. My desire was to make a DC coupled  differential in/ differential out level shifter, so I could float a discrete DC coupled mic preamp front end gain stage up to phantom voltage levels and shift it's output down to the nominal 0V DC for interface with other gear. (I guess most would just use a transformer for this, but I consider transformers as little better than capacitors, on my list of components to avoid. )

This simplified schemo does not show the actual front end preamp gain stage, just the level shift from nominal 50V DC down to 0V DC.

Basically it's a DDOA with 3 device LTP input, with T3 and T8 looking like inverting opamp inputs relative to T7 that acts like a common non-inverting input.  FWIW this looks like it passes the two legs intact, albeit inverted and level shifted. CM signal does cancel.

Caveat- this schemo is simplified and unproved so may not even work,,, I never bothered to develop it further, since it seems an even better idea would be to float an A/D converter up there to the phantom voltage level, then use optical or whatever to level shift the digital output down to the real world.       

JR

 

[Jean Clochet]

Hi John,
thanks for posting this - it's the kind of thing I like playing around with in Spice and on my bench.
Do you mind if I take the idea and have a play?  I'm interested in eliminating the input coupling caps on a mic amp that's on the bench right now,

Cheers.

 

[JohnRoberts]

Please do...this is published here and AFAIK original, at least I didn't copy it from somebody else, I can't say somebody else didn't have had a similar thought, that I don't know about.  I have wasted too many hours doodling out discrete circuit design. In another lifetime I'd try to get a gig designing ICs, that is where the rubber meets the road for serious design (IMO), while they have a richer palette of active devices to work with inside ICs.

Another characteristic of the LTT is reduced transconductance, versus a similar current density LTP, so slew rate may be improved (slightly). If you actually build this you may want to mess with different value degeneration resistors in series with T3 and T8 wrt T7.  I notice that I didn't include any dominant pole stability compensation on my simplified schemo, so that will likewise be a little unconventional.

have fun..

JR

PS: While I've been thinking about this for years, the most developed capacitor less design that I am aware of is Wayne's  http://www.proaudiodesignforum.com/forum/php/viewtopic.php?f=6&t=14  He has actually made one work and started vetting the other problems sure to arise, like from real world microphones, etc.

 

[gemini86]

Really good read! Thanks for posting that.

 

[JohnRoberts]

I consider this one of the few "heavy lifting" tasks for capacitors (others are passive loudspeaker crossovers, and DC blocking in series with mic pre gain pot).

This seems a little like chasing our tail, since most (all?) phantom powered mics will have capacitors inside their internal electronics.

It seems the blocking caps should be unnecessary for amplifying non-phantom mics, so a couple extra poles on the phantom power switch could shunt across the caps when phantom power is removed. I suspect this is not done because the spurious clicks and pops would be more concerning to customers, than any real or perceived degradation from leaving the caps in path.

JR

 

[Jean Clochet]

I consider this one of the few "heavy lifting" tasks for capacitors (others are passive loudspeaker crossovers, and DC blocking in series with mic pre gain pot).

Yes I agree. 
As far as the mic-amp gain cap, I would personally rather have some "whoosh" sound as gain is adjusted and maybe have to replace the pot every few years?  Even though I don't have excellent ears, even I can hear a benefit without a cap in that spot. 

It seems the blocking caps should be unnecessary for amplifying non-phantom mics, so a couple extra poles on the phantom power switch could shunt across the caps when phantom power is removed. I suspect this is not done because the spurious clicks and pops would be more concerning to customers, than any real or perceived degradation from leaving the caps in path.

Yep.  There are a couple of mic amps marketed specifically for non-phantom use - one of them sold by Wes Dooley which, if I'm not mistaken, was a Fred Forssell design?
I'm not sure if it is but, I would think it might be best if this type of pre were to have J-Fet input devices for the low input current.  Although, at mic impedances, noise is worse so you might be looking at a gazillion of them in parallel which has it's own issues with making sure you don't build an oscillator!
Anyway, good stuff, I appreciate the idea and the links 

 

[JohnRoberts]

I consider this one of the few "heavy lifting" tasks for capacitors (others are passive loudspeaker crossovers, and DC blocking in series with mic pre gain pot).

Yes I agree. 
As far as the mic-amp gain cap, I would personally rather have some "whoosh" sound as gain is adjusted and maybe have to replace the pot every few years?  Even though I don't have excellent ears, even I can hear a benefit without a cap in that spot. 

That is the paradox of designing for the mass market... they are far more sensitive to those easy to hear things like switch clicks and  scratchy pots, and judge the sound quality based on those, leading to the use of multiple blocking caps in every path. It is difficult to explain that the path could actually be better, when their ears tell them otherwise.

It seems the blocking caps should be unnecessary for amplifying non-phantom mics, so a couple extra poles on the phantom power switch could shunt across the caps when phantom power is removed. I suspect this is not done because the spurious clicks and pops would be more concerning to customers, than any real or perceived degradation from leaving the caps in path.

Yep.  There are a couple of mic amps marketed specifically for non-phantom use - one of them sold by Wes Dooley which, if I'm not mistaken, was a Fred Forssell design?

I knew Wes.. AES associates or something like that, out in Cal. Don't know FF, while I think he is known for a popular mixer circuit

I'm not sure if it is but, I would think it might be best if this type of pre were to have J-Fet input devices for the low input current.  Although, at mic impedances, noise is worse so you might be looking at a gazillion of them in parallel which has it's own issues with making sure you don't build an oscillator!
Anyway, good stuff, I appreciate the idea and the links 

Actually the lack of DC blocking cap does not change the input device criteria from a noise perspective, while perhaps you are thinking of DC input bias current? In more recent years I have seen some modern very low noise JFET pairs in the sub 1 nV range so perhaps they are candidates for a front end with no/low input bias current, while any input bias will be CM to the mic. Wayne used a servo with his DC coupled pre, and that is probably a good idea.

I have long ago stopped melting solder over my phono preamp and mic preamp musings while I am still interested academically.

JR

 

[volker]

I consider this one of the few "heavy lifting" tasks for capacitors (others are passive loudspeaker crossovers, and DC blocking in series with mic pre gain pot).

Yes I agree. 
As far as the mic-amp gain cap, I would personally rather have some "whoosh" sound as gain is adjusted and maybe have to replace the pot every few years?  Even though I don't have excellent ears, even I can hear a benefit without a cap in that spot. 

That is the paradox of designing for the mass market... they are far more sensitive to those easy to hear things like switch clicks and  scratchy pots, and judge the sound quality based on those, leading to the use of multiple blocking caps in every path. It is difficult to explain that the path could actually be better, when their ears tell them otherwise.

You just have to label it properly . It's not a bug, it's a feature!

 

[JohnRoberts]

I actually did something like that once during one of my marketing stints... I was product manager over power amps, and I had an old school heavy iron, heavy heat sink, large, lower tech product to merchandise (PV2000).  I flipped the detriments around and promoted the heavy transformer and TO-3 metal power transistors as features in an advertisement. People buy value products because of the price, but they don't want you to rub their nose in it... give them a good story to help rationalize their purchase decision. 

That said, I don't know how to promote switch clicks and pot scratch as good things, without a very long and winding story. In my experience customers do not have the attention span for obscure explanations for apparent flaws than can be dismissed as excuses by competitors, since they will attack apparent weakness at every opportunity.

JR

 

[ricardo]

That is the paradox of designing for the mass market... they are far more sensitive to those easy to hear things like switch clicks and  scratchy pots, and judge the sound quality based on those, leading to the use of multiple blocking caps in every path. It is difficult to explain that the path could actually be better, when their ears tell them otherwise.

You mean you prefer to design for subtle stuff that the unwashed masses (or even Golden Pinnae) can't hear and ignore stuff that they can? 

All measured via Double Blind Listening Tests according to Lipshitz & Vanderkooy of course.  8)

 

[JohnRoberts]

That is the paradox of designing for the mass market... they are far more sensitive to those easy to hear things like switch clicks and  scratchy pots, and judge the sound quality based on those, leading to the use of multiple blocking caps in every path. It is difficult to explain that the path could actually be better, when their ears tell them otherwise.

You mean you prefer to design for subtle stuff that the unwashed masses (or even Golden Pinnae) can't hear and ignore stuff that they can? 

All measured via Double Blind Listening Tests according to Lipsh*tz & Vanderkooy of course.  8)

I realize you are being sarcastic but my primary point is that consumers will dismiss a product based on temporary sonic issues like switch clicks, scratchy pots, and don't forget turn-on/off transients, without ever listening to the baseline sonic quality playing actual music.

So successful commercial designers must satisfy these IMO cosmetic issues, before even considering subtle improvements.

The irony is that mitigating these cosmetic issues without adding huge expense can actually degrade path performance.

I spent 15 years designing value products and learned to give the consumer what they want, based on how they vote with their money spent, not what I thought they really needed.

JR

 

[Gold]

I realize you are being sarcastic but my primary point is that consumers will dismiss a product based on temporary sonic issues like switch clicks, scratchy pots, and don't forget turn-on/off transients, without ever listening to the baseline sonic quality playing actual music.

If you are mixing a live show, having all that stuff is unacceptable. If you need to adjust the controls when you are recording, that stuff is also unacceptable. If you are doing a white glove audiophile recording where sound quality is more important than capturing a performance then you can get away with it. Equipment ergonomics rarely gets talked about around here but I consider it as important as sound quality.

 

[JohnRoberts]

Of course...  I guess I need to qualify that I don't consider these cosmetic issues unimportant, only that people tend to value them disproportionately relative to "measured" path performance (IMO).

A broader discussion of ergonomics would surely include things like gain laws, eq control laws and a whole panoply of subtle design decisions that people perceive often subliminally that color their perception of product quality, but never show up on data sheets, because the differences are transitory.

At Peavey we tooled up full custom parts and invested a great deal of engineering effort into simple things like mic preamp gain pots. Especially with next generation redesigns of high volume products. While of little consequence to actual performance, customers really care about things like how loud a product gets at 12 o'clock on the volume control... (I could tell stories about what happens when you miss by even a few dB).

A lot of EQ "quality" perception relates more to speed and range of boost/cut controls than static transfer functions.

everything matters, and the customer is always right. even when they aren't,,,

JR

 

[Gold]

Of course...  I guess I need to qualify that I don't consider these cosmetic issues unimportant, only that people tend to value them disproportionately relative to "measured" path performance (IMO).

It's much easier to spot a scratchy pot than describe a sound quality issue that a non technical operator doesn't have the language to describe. Many sound quality issues only reveal themselves after some time. Something would have to be pretty bad to notice right away. Dynamic audio is very different from test signals after all.

I agree that all the subtle ergonomic issues you described can add up to something being liked and useful, or not.

 

[ricardo]

That is the paradox of designing for the mass market... they are far more sensitive to those easy to hear things like switch clicks and  scratchy pots, and judge the sound quality based on those, leading to the use of multiple blocking caps in every path. It is difficult to explain that the path could actually be better, when their ears tell them otherwise.

I detect a certain amount of Golden Pinnae prejudice against Caps here.

I've spent a lot of time making switches popless & knobs scratchless and don't begrudge any of it.  In a Broadcast Desk, clicks, pops & scratchy pots are unacceptable and quite rightly so.

It's like making gear unconditionally stable & immune to RFI.  Gear that is immune sounds cleaner & better than stuff which isn't.  This is one of the things which separates the men from the toys though the usual specs won't tell you that.

There's places where even electrolytic caps have their place.  Cyril Bateman has chapter & verse and a brief executive summary appears in Self.  They point out where electrolytics may have some effect and more importantly, when they don't.

Having said that, one place I wouldn't have them is in series with the gain set resistor on an electronic balanced mike amp like Cohen, THAT etc.  They will almost certainly affect the LF response and hence introduce distortion bla bla and be unreliable.  The false prophet Jung, after pontificating at length about their evil, uses them in the one place where it would be audible on a Double Blind Listening test bla bla.  That tells me a lot about his (Golden?) pinnae.

I note Cohen DOESN'T have a cap there while practically all the DIY copies do.

Anyone here conducted Double Blind Listening Tests per Lipsh*tz & Vanderkooy on caps?  I have.  The results are not always what you expect.  And the Golden Pinnae aren't always Auric.

 

[JohnRoberts]

I'll make you a deal, don't put words in my mouth, and I won't put words in yours.

That said, i am not opposed to elegant solutions that finesse perceived or "real" problems. The best cap is no cap, and that is what my DC coupled level shifter does. I consider a capacitor-less phantom mic preamp as more of a marketing hook than advancing the state of the art. I stopped trying to sonically improve mic preamps almost 20 years ago.  Frequency response was already ruler flat, distortion hard to measure, and noise within a dB or so of theoretical perfection.

They already make capacitor-less speaker crossovers, it's called bi-amping.

In 15 years working for a value manufacturer, I learned to try to give the customer what they want.  Since lots of customers think caps are bad, that sounds like a product opportunity. Trust me, it's a lot easier to give customers what they want, then educate every last one of them that they are wrong. Even if you could, it would mostly piss them off, not put them in a mood to spend money.

JR

 

[ricardo]

Peace John.  I'm merely nailing my flag to the mast just like you.

I try to design buttons that don't click and pots that aren't scratchy.  I like certain electrolytics cos they help me do this.  Because of this, I've conducted Double Blind bla bla and other tests to find out where/which caps are evil and where/which  aren't.

IM very HO, gear that doesn't click or make scratchy noises are "better" than those which don't; even if they have loadsa caps.  You may disagree.

Perhaps its my poor facilitiy with English but it's difficult to translate

That is the paradox of designing for the mass market... they are far more sensitive to those easy to hear things like switch clicks and  scratchy pots, and judge the sound quality based on those, leading to the use of multiple blocking caps in every path. It is difficult to explain that the path could actually be better, when their ears tell them otherwise.

as other than a PC version of, "Caps are evil and the unwashed masses have wonky ears."

On the other hand, more than a decade of Blind Listening Tests tells me the unwashed masses (especially the female of the species) are far more perceptive than most people (including themselves) give them credit for, while generallly, most Golden Pinnae have wonky ears.
On the subject of mike preamps, input caps introduce problems with protection (for mike & preamp), CMR and also real estate.  In the end, the REAL performance of a mike preamp has many facets.

If the Unobtainium Transformer Co. had an equivalent to the Lundahls we used at Calrec in the 80's but the size & price of a THAT1510, I'd gladly go back to one of them + a NOS Mullard 5534.  Lose 0.7dB NF and loadsa measured spec but the real life robustness & RFI immunity would more than compensate.  And that would have caps galore too ...  ;D

 

[JohnRoberts]

While this is an unwanted veer from what I hoped was an inspection of an interesting circuit topology variant (my three input differential stage), I guess I do consider consumers if not to have wonky ears, at least to have questionable judgement regarding understanding what they do hear.

Allow me to expand upon the kind of engineering trade off that I have made to feed these consumer tastes that I would not make on a product for my own use. Turn on/off transients are annoying, and especially to live sound reinforcement where products may be connected to thousands of watt power amps and efficient loudspeakers. Generally experienced operators learn to turn on power amps last and turn them off first, but consumers can't be bothered with such things.

Historically I am not aware of speakers actually being damaged by these turn-on/off transients, only that it sounds bad. The value engineered solution is to increase the series output resistance and add a shunt element, like a JFET to ground.  Sometimes to get good transient kill you need to use two stages of this in series. Not to mention that these JFETs are sensitive to static hits and the like, so they need protection diodes, etc.  Back before these standard output driver ICs existed I asked them to build this mute capability into them but AFAIK they didn't.

I'll leave it to you to ponder if 50 ohm source impedance or 1k ohm source is better. A classic recent example of what happens when you don't address this was the first generation DSP/Speaker crossover controller, offered by dbx (sorry I don't remember the model number). They neglected to mute the turn-on/off transient and never heard the end of it from worried customers. They corrected that oversight in subsequent models.

I have long argued the electrolytic capacitors when properly used are not a problem.  I even wrote one of my Audio Mythology columns about them back in the '80s, so it's ironic to be painted as a cap hater now... (I also wrote about DC servos too). That said I wouldn't mind eliminating any parts (not just caps) that aren't really doing something useful. I used lots of caps in my designs of Peavey products  because the customers wanted them, whether they knew it or not. 

JR

Note: My personal power amp, that I designed and built in the early '70s (still works), uses a 3 position power switch, and relays, so the speakers only get connected after the power amp has stabilized, but get disconnected immediately at power off.

 

[ricardo]

Peace John.  I think we share many of the same prejudices though we differ on clicks & pops.  I agree Rs=1k is evil.

Having been a beach bum for more than a decade, I can't resist stirring things now as I was forced to be PC for most of my previous life.
On your LTT level shifter, I'm always interested in ways of getting rid of capacitors provided they are simple enough for my small brain and don't take up too much space.  You already know I begrudge the real estate for 2 resistors 

On the actual circuit, I don't really have anything to say except to point out Wayne's; which you've already done.