Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Teddy G.]

I don't care, long as it works!

For my first 5 years in the biz, my favorite mic was the EV RE-20(A dynamic). Then it was the RCA BK5(A ribbon), for another 10. Then, about 20 years ago, I tried an AKG C414, solid-state condenser(Against the RE-20, not the, then unavailable, BK5) and it became my favorite(Neuman's never "worked" for me.). That's the way it was until about a month ago, when I tried a Brauner Valvet. Now, it's my favorite. Of course, it's also possible that the Millenia Media solid-state preamp had something, or alot, to do with it? Though the Millenia's "addition" would have to be a "combo" thing? I've used lots of mics through the Millenia and never heard "it" until the Brauner..? Wonder what the Brauner would sound like through a millenia tube pre? Still, I like to have an RE-20 around - it has it's uses and then some.

Tubes/Schmoobs - what's it sound like?

TG

[Ronny]

danlavry wrote on Fri, 10 February 2006 15:44

The tube lovers want us engineers to be open minded to statements such as "I like it but I do not know exactly why". It is only fair to ask the tube lovers to be open minded to the possibility that tube like distortions can be made with circuits based on semiconductor technology, digital computer modeling technology and the rest of electronics... Regards Dan Lavry www.lavryengineering.com

Ever thought about getting into guitar amp modeling, Dan? They could use a lot of help.

First ss guitar amp that I heard that sounded like a tube amp was the Carvin Pro Lead 150 back around 1982, technology has been around awhile for solid state to emulate tube distortion for amps, but the modeling processors have a lot to be desired. They are getting close to the B3 sound with synths and horns are getting better, but still no cigar on stringed instruments and amp modeling, IMHO.  

[maxdimario]

Quote:

Tubes tend to be much nosier then transistors and FETs. Tubes age very fast... And if you like the sound, that is subjective, not a technical comment. All circuits based on tubes, transistors, FET's and Opamps Begin with parts containing some degree of non linearity, and the circuit design does correct for the non linearity to some degree, most often by use of negative feedback. Many tubes circuits have been incorporating feedback from day one (over a hindered years ago), and a tube designer knows that the reason for using feedback sparingly has been the fact that tubes have VERY LIMITED OPEN LOOP GAIN, thus tubes offer LIMITED AMOUNT OF NEGATIVE FEEDBACK. There is a TRAD OFF between circuit gain (closed loop) and linearity (more feedback).

What tubes are you talking about?

One EF804s can provide more gain with less noise than 1 trasistor or fet.

Funny that V76 or v77 preamps when working properly have an extremely high S/N ratio... all with 4 (v76) and 2 (v77) active components in total.

These are noise specs that transistor pres only dream of.. unless they use 8 or 10, or in the case of opamps, 25 to 100 active components in the signal path.

there are plenty of tubes of special quality that meet and exceed fet devices as far as linearity and noise.

High mu triodes and pentodes don't have limited open-loop gain.


does a transistor (one in number) utilized at high gain sound good? NO.. transistors are excellent switches for the same reasons.

does a tube (one) utilized at high gain sound good? YES.. depending on the QUALITY of the tube and the components around it.

anyway the proof is in the results.

you can build a simple amplifier suitable for music reproduction with a tube, a plate resistor linked to High Voltage, a grid resistor linked to a negative bias voltage and cathode to ground.

you can't do the same thing with a transistor.

with a fet you CAN build a similar circuit but there is a higher amount of distortion, and it a grating harsch distortion.

Early transistor designs, based on the simple concepts used in tube amps, sounded like crap, which is why engineers began to invent a whole bunch of ways to make them sound better by utilizing more active components to share the overall gain and current amplification.

preamp (low current) tubes can last for decades.

The 'cleanliness' of transistor devices lies in the amount of engineering and component count involved in making transistors suitable for passing a music signal.

since transistors are cheap, you can use as many as you like, splitting the total gain in many stages and keeping the internal impedances low.

but in a microphone every detail, including the distortion artifacts, gets amplified hundreds of times, so it's a good real-world display of what different active components really sound like.

[danlavry]

maxdimario wrote on Sat, 11 February 2006 16:59

Quote: Tubes tend to be much nosier then transistors and FETs. Tubes age very fast... And if you like the sound, that is subjective, not a technical comment. What tubes are you talking about? One EF804s can provide more gain with less noise than 1 trasistor or fet....

You said:

“What tubes are you talking about?
One EF804s can provide more gain with less noise than 1 transistor or fet.”
These are noise specs that transistor pres only dream of.. unless they use 8 or 10, or in the case of opamps, 25 to 100 active components in the signal path.

I say:

You missed my point completely. You are still talking about a device, and I am talking about a device in a circuit!  One can have, for example, a transistor, or a FET or an opamp with noise specs of less then 1nV/sqrt HZ (by the way it is better then any tube I know), and  of course one needs to account for noise current nA/sqrtHz (very tiny for FET). Still, it DOES NOT ADDRESS THE END RESULT! Take a state of the art transistor with say 1nV/sqrtHz and 2nA/sqrtHz, and put it in a mic pre circuit, with a mic impedance of say 150 Ohms, and the noise is swamped by  the resistance. A 150 Ohm resistance by itself is 1.56nV/sqrtHz at room temperature. Add to it the effect of the noise current, and you have another 3nV/sqrtHz. The combined noise figure is now 3.5nV/sqrtHz. Take the resistance out of the circuit and you end up with 1nV/sqrtHz.
So there is little value in looking at the device in a theoretical zero Ohm circuit. The shorted input yields .144uV over 20KHz, but with a good mic resistance (150 Ohms) you get .5uV, a loss of nearly 11dB.

And while at it, instead of claiming “open ended” claims that “one EF804 is…. “ come back with the DATA (real numbers) for both noise and gain. I believe you will be surprised. What I see in the data sheet for EF804S is a specification of 2uV for 25Hz to 10KHz bandwidth, so it is already WAY OVER over a state of the art transistor, even before you put it in a circuit, and at 1/2 the bandwidth. Now lets look at the gain. Say we have between 100 and 180V plate voltage, then the most you can get is about 1.5mA per volt change in the grid – that is gm = 1.5mA/V. Of course that is when running up to 3mA current. In reality that is too much (still VERY LOW COMPARED TO AVARAGE FET), and at typical current less then 1mA, the gm is lower, about 1mA/V…

To talk about gain, you must specify a load resistance! Lets go for 200KOhm (pretty typical), so the voltage gain is A=gm *RL = .001 * 200000 = 200, no more! That is low open loop gain!!!  Take a circuit made of 3 transistors, each one with hfe = 25 minimum, the combination has a potential of 25*25*25 = 15000 gain, all for 30 cents in parts. Some opamps (made of transistors inside) yield 1000000 (a million) open loop voltage gain. In fact, most modern FET single devices yield much more gain then your tube, and again, it is NOT ABOUT STAND ALONE DEVICES, IT IS ABOUT CIRCUITS AND RESULTS.

You said:

“High mu triodes and pentodes don't have limited open-loop gain.”

I say:

Please revisit that statement. You are simply very wrong. “High mu” is a relative term. In the days of tubes, going from a mu of 25 to 35 was a big deal. We have come some ways from the days of selenium rectifiers

You said:

“does a transistor (one in number) utilized at high gain sound good? NO.. transistors are excellent switches for the same reasons.”

That is a SUBJECTIVE STATEMENT!
Again:
That is a SUBJECTIVE STATEMENT!
AND again:
That is a SUBJECTIVE STATEMENT!

You seem too intent on staying on your messege, no matter what, to the point of refusing to register what I said. I am an older engineer. I started with electronics in the tube days, before transistors, therefore I am familiar with tubes. Your statements have no teeth.

When you state subjective opinions about what what sounds good, you go against the rules of this technical forum.

When you state technical statements, you are stating WRONG FACTS. I looked at the tube data and I am a circuit designer, so what I say holds, unless my calculations and my reading of the data sheet are wrong (they are not wrong).

I am open to be chalanged ON THE FACTS, though I do not belive a real design engineer, capable of understanding data, circuits and and signals, and quantifying things can chalange me. I am not into "loving or hating" devices. I am into objective technical facts.  

I put some of my valuable time into this post. I would expect any further arguments to be backed up by some real study, going beyond just throwing of subjective preferences and  
“relative terms” with no reference for comparisons.

Regards
Dan Lavry
www.lavryengineering.com

[Teddy G.]

Just to stop for a minute, here...

"...Tubes were usurped by transistors due to cost, size, etc."(Not a quote!)..

Were they? Yeah, sure. "Total cost", really. THE SAME THING(An AM radio for the kitchen, a practical space craft.) was almost wildly more efficient to build and operate. There is no end to at least the "practical advantages" of the transistor over mid-20th. century tube technology - which is where we are now, in tube-tech. Anyone have knowledge of how big a Pentium 4 processor would be, if constructed using "50's" tube design? Still, I'm not sure that even now one can build a truly practical 50,000 watt, solid-state transmitter for a clear-channel AM radio station(Then again, who listens to 50kw, clear-channel AM radio stations anymore?)?

Another strong reason for the move from tube to transistor? I have a theory - or at least a "fun" thought...

AT&T was a pretty big company... They must have used alot of tubes? Far as I know AT&T had to buy all their tubes from others(RCA, Sylvania, etc.). When AT&T(Bell Labs) came up with the transistor, all this changed... Now, not only did AT&T not have to buy tubes from outsiders, they even(Even... Ha!) got to license the transistor technology back to those same manufacturing companies from which they had purchased tubes. What a deal! Certainly, AT&T, alone(Not just with their invention but the "purchasing power/needs/technical furtherence abilities" of that single company itself!), had alot to do with at least the "speed" as well as the depth of the transistor transition? Though it was still 40 years before tubes were finally "dead" for almost all purposes(Even rather higher power devices like most BC xmttrs - of the 1kw kind, for instance, which until recently used several "quart bottle-sized" tubes and which, are now "complete-on-a-desk-top-size" units -- about the size of a large PC.). Matter of fact, the tubes' "total death"(Meaning here, no longer having, or having almost no "necessary" purpose, in any application, at least of a "normal" high-volume manufacturing/commercial or consumer nature.), came at just about the time(Early 90's?) as it's "revival" with the beginnings of "new" tube gear being released -- coinincidently(I'm sure) at just about the time that AT&T(Bell Labs), essentially, "died", or was in it's death throws, itself.

Of course, tubes never really "died", they never stopped doing what they do, they just stopped being a subject of truly major manufacturing(thus consumer/broad technical use, "parts" to make them work - high voltage xfmrs, etc., and furtherence of the "entire" technology and it's support structure "died", also.). With no major support(Mostly the support of millions of consumer-type devices to use a technology and support it's furtherence.), vacuum tubes were and are, indeed, dead.). Does that make it a "bad technology" or one that "can't compete" with modern technology. Well, bad, no. Can't compete? Yes - tube tech of the mid-1900's cannot compete. Are tubes "better" at anything? It no longer matters. Might as well argue the good/bad points of buggy whips. Matter of fact, designing, manufacturing and selling a "better" buggy whip would be a whole lot easier(Probably being done!? "Horse people have money!), though I'm sure PETA, would scream and cry - much as environmentalists would react in horror, if RCA "re-started" making tubes by the millions using the "same old tech". NO ONE is making "better" tubes. For that matter, NO ONE is making "better" recording tape(Seems an "allied" topic, for almost all of the same reasons?). Tube/Tape "bettering" is not something one can do in the "home shop"(Or at least no one has said they are doing so now - or at least have "backed-up" what they've said with any better supply.) and there is no more RCA and "The New AT&T" is not AT&T, by any stretch of the imagination and what there is left of it is, very happily I'm sure, all solid-state...

So, why does that Valvet sound so darned good - years after it's "tube"(I suppose?) was manufactured..? The technology is dead! For sure, one reason is that like so many technologies, tubes to camera film, did absolutely get to wonderful standards of quality and performance - though with "technical" tubes and film(Very low amounts produced for v-e-r-y specific purposes, NEVER even intended to be "recouped" financially by their-own sales -- the movies could NEVER have afforded to use Kodak's fabulous 35mm film(To say NOTHING about 70mm!), if so many consumers hadn't been cutting Auntie Fran's head off, so often using Kodak's consumer film by the billions and billions and billions(To borrow a phrase) of frames. Recording tape? There would never have been Ampex Grandmaster without the simple audio cassette to "backup" the "pro" tech R&D. Tubes and it's technology? Never really did make even a dent in the "world-buck" state of things.  Yeah, lots of folks had a big radio, maybe a record player, but not until solid state did consumers load each and every corner of their lives with electronic gadgets. "Pro's", of several and sundry varients where truly "lucky" just to have what they did, when they did. Still, along with very well documented "how too" information, and alot of the "remnants" around???(Lots of manufacturers apparently  caught entirely unaware of just how fast transistors would take over! And the manufacturers of these tubes made them by the millions! By the time RCA could shutdown the 6146b machine, here in Lancaster, PA - union agreements, dumb management and all - they had probably made an "extra" 150 million of them, I suppose...)((By the way: A simple request to an RCA worker/friend, a few years ago - in the IT department - got me 2 of the last 4, 6146b's, in the basement of the old RCA factory, "brand new", in the lovely red and white boxes. Pays to ask! They reside, and I get "great audio" compliments from it/them(?) - in my old Kenwood TS820.)). Still, the supply of many, many tubes we are trying to use - to make a living with, for God's sake, is finite, at best - rotting away at worst. There will be no more - certainly, no better - unless you..??? No, I didn't think so.  

In the end, tube or transistor, tape or film, it's the design and build of a piece that determines how it works(A great "film" is a great film, whether shot on film or made digitally.). Nothing to say(At least from me) that Mr. Brauner couldn't build a SS mic that would sound exactly the same as his tube Valvet. Harder to charge big money for it as the "romance of the tube" wouldn't be there? Still, today, if I get the 3 G's, I'll buy one of the Valvets, as I've never heard any mic "do me" so well(Is it me? Or is it Memorex? Or is it dreaming..?). 'Scuse me if I keep an AKG C414 as a "backup" and an RE-20 to back that up.  


Tubes are dead, long live the tubes.

Teddy G.

Back to the argument, I'm learning alot!

[maxdimario]

Quote:

You said: “does a transistor (one in number) utilized at high gain sound good? NO.. transistors are excellent switches for the same reasons.” That is a SUBJECTIVE STATEMENT! Again:

With a very small change in voltage, a transistor can go from a non-conductive state to a fully conductive state with a very low impedance, compared to tubes.

Tubes need more input voltage to work and they have a higher impedance so most tubes can't be used for switching purposes as well as transistors (although there are some specially designed tubes which were designed for that purpose).

this makes transistors ideal for switching relays on and off, for example, or in logic circuits, or to shunt a signal to ground depending on a control voltage.

if the transistor is used as a simple amplifier, where the full gain of the transistor is used in it's relative circuit with no feedback (say 200 times)will the results be acceptable from an audio standpoint? How about noise?

In tube mics, the only type of transistor that can be used possibly is the FET.

does a FET have a gain which is superior to a tube?

what has better inherent linearity over a wider input voltage, a FET with no feedback, or a special audio tube with no feedback?

What are the technical differences (S/N, Headroom, distortion) between a fet in a no-feedback circuit, and a tube (neumann tube quality) in a no-feedback circuit?

I am not slagging transistors for all applications, as there are indeed some great transistor designs (all of which are very simple circuits), but to say that tubes are noisier and more distorted is a fallacy.

12ax7's and the like are NOT audio tubes, they are cheap miniature tubes.

[Ronny]

maxdimario wrote on Sun, 12 February 2006 14:13

12ax7's and the like are NOT audio tubes, they are cheap miniature tubes.

Oh, brother. Tubes have to be a certain size now to qualify.

[danlavry]

[quote title=maxdimario wrote on Sun, 12 February 2006 19:13][quote title=Quote:]You said:

“...but to say that tubes are noisier and more distorted is a fallacy.”

I say:

A asked you to refrain from "general" statements, and come back with technical reply to my comments. I did not think you would or could raise to a level of technical discussion, but calling my comment "fallacy" without any technical backing simply stinks.  

You said:

“With a very small change in voltage, a transistor can go from a nonconducting state to a fully conductive state with a very low impedance, compared to tubes.”

I say:

Your comparison of a voltage input device (tube) to a bipolar transistor (a current driven device) is way off to begin with. I assume your comment was about bipolar because it does not cover FET transistors which are NOT abrupt…

But again, you are focused on a device, not on a circuit or the END RESULT. What is the problem? What does it take to make you register it?  Look at the WHOLE CIRCUIT. Indeed, when one wishes to switch fast, one would take advantage of the bipolar transistor ability to be switched easily between full conduction to no conduction. A typical switching circuit would have a grounded emitter, the base is driven by some current for conduction, or the base is “reversed” (or grounded) for cutoff.

But let us now take the beginning of the beginning of “transistor 101”. If I want to have a transistor work in the linear range, I can add an emitter resistor between the transistor emitter and ground. Say I add a 1KOhm, and drive a voltage change to the emitter, then the current through the device changes at 1mA per volt, very similar to a tube… That single resistor converted a current switching device to a voltage driven device. Of course there are other and better circuits, beyond “transistors 101”.

Yes, some transistors are oriented towards switching – very low capacitance, high gain bandwidth product and so on. Other transistors are more oriented for analog – low noise, higher voltage and or higher power, precision matched pairs and so on….

Your argument is STILL based on looking at “devices in free space” not in a circuit. Your statement shows that you are not a circuit designer. You arguments are coupled with vast lack of circuit knowledge.  

You said:

”if the transistor is used as a simple amplifier, where the full gain of the transistor is used in it's relative circuit with no feedback (say 200 times)will the results be acceptable from an audio standpoint? How about noise?”

I say:

No one will use a transistor with no feedback, and even after I have stated it, you still refuse to accept WHAT REAL CIRCUITS ARE LIKE. I stated twice, that it is a whole circuit that counts. You are coming back with “comparison” between a tube, and a transistor in an IMPOSSIBLE CIRCUIT. ANY ONE THAT KNOWS ANYTHING ABOUT  TRANSISTORS KNOWS THAT THEY REQUIRE FEEDBACK, THAT YOU CAN NOT EVEN BIAS TRANSISTORS WITHOUT FEEDBACK.

Your comparison is based on complete ignorance, or total refusal to concede ANYTHING AT ALL!

You say”

“In tube mics, the only type of transistor that can be used possibly is the FET.”
does a FET have a gain which is superior to a tube?

I say:

Many FET’s have higher gm, but again, one can put a whole bunch of FET’s and transistors and resistors in a lot less space then a tube, at a tiny fraction of the power required, a tiny fraction of the cost…
THE COMPARISON SHOULD BE BETWEEN CIRCUITS NOT DEVICES. You seem to have serious difficulty with that concept.

You say:

”I am not slagging transistors for all applications, as there are indeed some great transistor designs (all of which are very simple circuits), but to say that tubes are noisier and more distorted is a fallacy.”

I say:

I looked at the data sheet of the EF804S, and came up with the numbers. I took my dear time, and asked you to do similarly. The data clearly shows that tubes are noisier then low noise transistors and opamps!  The data shows that tube circuits are less linear because they have low gain, thus unless you are ready to “cascade a triplet or more” for just one low gain stage (with feedback), you end up with less linearity.  

Instead of answering with data, instead of doing your homework and filling the huge holes in your “thinking” you chose to end you post with a disrespectfull statement:
“but to say that tubes are noisier and more distorted is a fallacy.”
 
I am not stating fallacies, you are! Your comment is totally disrespectful to me and to all the participants. Your presentation of a comparison between a tube and an "open ended" transistor (no feedback) is either based on a complete ignorance of circuits, or worse, and attempt to stand your ground no matter what it takes. Which is it?

Dan Lavry
www.lavryengineering.com  

[compasspnt]

Ronny wrote on Sun, 12 February 2006 15:24

maxdimario wrote on Sun, 12 February 2006 14:13 12ax7's and the like are NOT audio tubes, they are cheap miniature tubes. Oh, brother. Tubes have to be a certain size now to qualify.

Well then, I guess the AC701 is out of play!

[Ronny]

compasspnt wrote on Tue, 14 February 2006 21:24

Ronny wrote on Sun, 12 February 2006 15:24 maxdimario wrote on Sun, 12 February 2006 14:13 12ax7's and the like are NOT audio tubes, they are cheap miniature tubes. Oh, brother. Tubes have to be a certain size now to qualify. Well then, I guess the AC701 is out of play!

Ya want big sound, get big tubes, it's a no brainer.  

[compasspnt]

But I will agree that the 12AX is not the best all time sounding valve.

[maxdimario]

The ac701 is a specifically designed miniature triode, and it sounds great.

there are people who will argue that the VF14, although not really a mic-amp tube does have better results from from the hi-mids down.

The reason why I compare transistors to tubes out of their circuits are the following:

first of all I begin with the idea that the least amount of active components make for the most direct sound.
this is due to the increasing complexity of the distortion as an amplifier increases in component count.
the product of any system becomes more unpredictable as the system becomes more complex.

if one is to pick what are deemed to be the best high gain amps made, usually the standard choices will go back to discrete gear with simple circuits.

I remember looking at the schematics for a HG desk built by a very famous german engineer for the german broadcast institute, in the late 70's or early 80's and the transistor circuitry inside had a topology similar to tube design.

this is in a mixer built with NO cost limits, where the wiring was in silver etc.

the best transistor stuff is very simple.

I have a V77 with new tubes and caps that can amplify a ribbon mic such as a 77d and sound convincing, with low noise, all with two active components (60-70 db of gain).

two-transistor preamps with such a gain have never been built. Maybe you can build one?

comparing a tube to an op-amp for noise is not the same thing, since an op-amp is an integrated circuit with more than 10 transistors inside, usually.

The FET transistor as you say is the only transistor that can be used without feedback, being a voltage-driven device and having the capacity for a higher input voltage. There is no need do convert voltage to current using a resistor.

So theoretically you can build an all FET design with no feedback, but no one has been able to do it with SUCCESS.

on the other hand there are tube designs with no feedback that work (some 50's gear springs to mind).

As you probably know, no-feedback mics such as c12 u47 etc. are generally thought of as being superior to mics with feedback in the amp.

[maxdimario]

Just wanted to add that once I did make a cicuit with triodes that was built like a transistor amp.

it was very clean and quiet, had (relatively)low output impedance etc, all the attributes which are often associated with solid state.

but unfortunately, even though the sound was 'cleaner' it sounded a little dull and processed, compared to the simpler traditional tube designs.

so a tube can be used like a transistor, as long as you don't try and drive low-impedance loads, of course.

you can even make tube op-amps if you like.

nobody would use them, but you could make them nonetheless..

[Gunnar Hellquist]

maxdimario wrote on Wed, 15 February 2006 16:40

first of all I begin with the idea that the least amount of active components make for the most direct sound.

This seems to be a common misconception. One problem with it is that it is provably not true in all circumstances:

If it was true, my cell phone microphone would have the most direct sound. A typical electret mic is only three components, you cannot really go below that: capsule, resistor, fet. Actually, electret mics can be made very good, at least I love the DPA4003 but they have a few more components.

The whole idea that the component count should be an important factor in selecting mics is sort of totally off in my world, but then again it is your choice.

To me, using mics is like eating food. Some times I like the Italian way, sometimes the French way of cooking. Italian typically builds on very few ingredients, French is complicated with lot of different ingredients. Both can be very nice, or extremely bad depending on the chef doing the cooking.

So, to wind it down, the chef is what counts. Sorry, the very mic design. Some mics use few components, some many, some uses tubes others not. What counts is how they sound and how well they fit with what you are recording. We live on a long tradition of Neumann (the man) that used whatever he could get his hands on (tubes, transformers, simple plastics) and cooked a great meal from that. Todays chef, sorry designers, has different ingredients but they may create a great meal anyway.

Gunnar

[danickstr]

to respond to the original poster's question:

The reason tubes are still in use is because the complexity of their sonic shaping still eludes the best designers of solid state attempts to emulate them, as a general rule.  You mention that tube harmonics and other alterations could be "easily duplicated".   Well, get to it.  You have millions of dollars to make for yourself.

They can be duplicated on a case by case basis with static signals, but there is no circuit that can react like a tube to a real live musical performance, because of the dynamic relationship between the electronics and the input signal.  If there were, then all the great producers would use it.  They don't.  They use tube gear, well the ones who like the sound of a good piece of tube equipment.

Til you are done showing the world how easy it is, however, I will stick with a good classic tube, for all the magic is has in it's little body.