Charles, nobody knows what you are talking about, except real time audio designers.
I learned about the nonlinearity of IC resistors when I took a night course in IC design given by Hans Camenzind, 40 years ago, in 1968. (You know, the guy who invented the 555). EVERYBODY knows that standard IC resistors are lousy! Except people here.
Also, I see that you have worked with Barrie Gilbert's IC's (non-opamp) with some success. I am not surprised that they would work as well as just about any IC out there. These guys don't appear to know the sonic difference between a fet and a bipolar device, including bias considerations. They also don't know about the differences between typical solid state and tube rectifiers. I prefer tube rectifiers for tube circuits, and I can prove why. However, for solid state circuits, I use high speed, soft recovery diodes (what a hassle) because I can hear the difference between those and standard rectifiers. In fact, it is my D Vendetta upgrade. I wish that I had the better diodes in when HP listened to my Vendetta 17 years ago. I'm sure that it would have modified his opinion, at least a little. Oh well, I live and learn, just like you.
I learned about the nonlinearity of IC resistors when I took a night course in IC design given by Hans Camenzind, 40 years ago, in 1968. (You know, the guy who invented the 555). EVERYBODY knows that standard IC resistors are lousy! Except people here.
Also, I see that you have worked with Barrie Gilbert's IC's (non-opamp) with some success. I am not surprised that they would work as well as just about any IC out there. These guys don't appear to know the sonic difference between a fet and a bipolar device, including bias considerations. They also don't know about the differences between typical solid state and tube rectifiers. I prefer tube rectifiers for tube circuits, and I can prove why. However, for solid state circuits, I use high speed, soft recovery diodes (what a hassle) because I can hear the difference between those and standard rectifiers. In fact, it is my D Vendetta upgrade. I wish that I had the better diodes in when HP listened to my Vendetta 17 years ago. I'm sure that it would have modified his opinion, at least a little. Oh well, I live and learn, just like you.
john curl said:
To Charles: Would you be willing to tell us a bit more about those? Are these commecial IC's or are they custom developed for you?
BTW I read about this in your interview on the Ayre website. That interview seems to me (among other gems) a great plug, (coming from you!), for the advantages of IC circuitry over discretes.
Jan Didden
john curl said:
Arrgh! I HATE it when you make me go do all that cut-n-paste
But seriously, I know he doesn't say that IC's are the solution to everything, but he gives quite a few advantages and even calls it a 'breakthrough'. Reading from his posts above I got quite another impression on his views on this. But indeed, maybe I read it wrong, it has happened before. So, I yield to you without further ado.
But I would be very interested to know more about these IC's. Please, Charles?
Jan Didden
quote:
Originally posted by Charles Hansen
But you are absolutely right. It is *much* easier to design good sounding equipment with tubes. In fact, I would go so far as to say that as long as you don't do something stupid (liking using high-feedback solid-state voltage regulators) it's almost impossible to design bad-sounding tube gear. You can even get away with solid-state rectifiers, although tube rectifiers will generally sound better.
...to add some more to support G.Kleinschmidt
:bs: :bs: :bs: :bs: :bs: .... never enough for such statements!
Originally posted by Charles Hansen
But you are absolutely right. It is *much* easier to design good sounding equipment with tubes. In fact, I would go so far as to say that as long as you don't do something stupid (liking using high-feedback solid-state voltage regulators) it's almost impossible to design bad-sounding tube gear. You can even get away with solid-state rectifiers, although tube rectifiers will generally sound better.
...to add some more to support G.Kleinschmidt
:bs: :bs: :bs: :bs: :bs: .... never enough for such statements!
Re the comment a few posts back about resistors. Semiconductor companies would definetly not make stand alone resistors in their fabs to compete with common discrete resistors because it just does not make economic sense - absolutely nothing to do with performance or anything else. You can build very good discrete resistors with some relatively cheap equipement (=low investment) . . . . so why tie up a multi-million semi fab to do this. The payback would not be there. Thats the reason Charles that you cannnot buy discrete resistors made in a semi fab.
WRT IC resistor peformance . . . if needed, very good resisors can be made - linear, laser trimmed, ppm stability - but these cost even more than standard polysilicon resistors, so are only used where absolutely necessary. BTW, the use of laser trimmed resistors in IC's has declined over th e last 10 years or so as IC designers and process specialists have improved theie techniques to get around a.) the requiorement for accurate resistors and b.) tighned up processes so that resistors are much more accurate and have predictable performance.
Interwting to see that Charles has resorted to using 'dirty sand' in his products.
Hmmmm
WRT IC resistor peformance . . . if needed, very good resisors can be made - linear, laser trimmed, ppm stability - but these cost even more than standard polysilicon resistors, so are only used where absolutely necessary. BTW, the use of laser trimmed resistors in IC's has declined over th e last 10 years or so as IC designers and process specialists have improved theie techniques to get around a.) the requiorement for accurate resistors and b.) tighned up processes so that resistors are much more accurate and have predictable performance.
Interwting to see that Charles has resorted to using 'dirty sand' in his products.
Hmmmm
SY said:
Not that I'm aware of, although I've never investigated that avenue. The tube circuit will have 5 resistors (if properly done), one triode, and one coupling cap. It would require only a positive supply (although at high voltage) and heater supply (something of a can of worms in itself, as it can have such a large impact on the sound of the circuit).
The solid-state circuit would have 12 transistors, 12 LEDs, and 10 resistors. It would require a bipolar supply. The main advantage would be the elimination of the coupling cap. I have found that these add more coloration than does a current mirror that eliminates the need for the capacitor.
But I agree with your point completely. A wonderful sounding line stage preamp can be made with 1 or 2 triodes per channel. A wonderful sounding power amp can be made with 1 or 2 input tubes and 2 output tubes per channel (I'm a push-pull kind of guy). There was a "dark era" when good transformers were hard to come by. In the old days, Dyna sold theirs (William Johnson used these in his first products!). Then it became really hard to find good ones. Now it is easy to find good ones, possibly the best ones ever made.
Plus these designs can be built very easily with point-to-point construction, eliminating the need for CAD software and making PCBs.
Finally, the simplicity of the designs makes it easy to experiment with different brands of components and fine-tuning of circuit values. (One of the few benefits of the "chip" amps, in my opinion.) So I would agree that the average DIY'er would best off to at least start with tubes. (That's what I did, although I have to admit that I also built a Leach amp in the mid-70's that sounded really great once I removed the stupid low-pass filter cap on the input -- I used a mil-spec ceramic part, and it really screwed up the sound.)
Cheers,
Charlie
Charles Hansen said:
Charles,
My view on this is that tubes sound better (in part) because they can swing 100V or more relatively linearly. If you can do 100V well, then your ability to swing a mere 1 or 2V is assured. So it's not that you'd need the 100V, per se, it's that it's a nice way to back into low level linearity.
SY's views may or may not coincide with mine.
(Now, if I could only get my solid state stuff to sound as good as my tube stuff, I'd be a happy camper...)
SoNic_real_one said:
Resistor series? There are making ultrasensitive HALL sensors for a long time now
Unfortunately, there are only three options. You can load the line capacitively, inductively, or resistively. At some point, you're going to find that what you're using as a sensor is going to have an effect on the thing you're measuring. How much? That will depend on the sensitivity of the circuit behind the sensor. Use a sensitive enough circuit and you can use a .0001 Ohm resistor, which is going to have a pretty minimal effect on the measurements.
Or is it?
That's the problem. I don't know how many decades of sensitivity will be required before useful information turns up...and that's assuming that there's anything there to begin with. My suggestions are conceptual only. I prefer to use what little time I have to work on electronics working on audio circuits, not measurement circuits. Some people here seem more fascinated with the measurement side of things, so perhaps they'd be willing to spend the time to actually build and try some alternate test equipment.
The reason I specified resistance is that with current technology resistors are closer to their ideal than capacitors or inductors. I'd rather keep things as pure as possible.
Using the Source or emitter resistors seems tempting, but they're generally only of medium quality (and often 5% tolerance, sometimes worse--I would prefer something like .1% as a sensing resistor [yes, they're expensive, but you'll only need one]) and at any rate, you're only seeing the behavior of one output pair. I want to know what the amp as a whole is doing.
Grey
SY said:
I have three pieces of data that support my assertion. The first was back in the dawn of high-end history. The Audio Research SP-3 had a capacitance multiplier voltage regulator. I think the reference was a string of zeners (bypassed with an electrolytic cap), and the pass device was a bipolar power transistor in a TO-3 can. (By the way, this was probably the main reason that it sounded better than the otherwise very similar Dynaco PAS-3.)
Then there was a small company in Paoli, PA that made a HV regulator using a floating 3-pin regulator as an add-on PCB for the SP-3. My friend put it into his preamp (I certainly couldn't afford one then!). The sound become much more detailed, the bass tightened up, and the soundstaging become much more focused. But after a few months we noticed that we simply weren't listening to as much music and when we did we weren't enjoying it as much.
The second time this happened, I took a Dyna Stereo 70 and rebuilt it with the Curcio input board from the mid-'80s. This had a cascoded driver stage with a diff-amp as the phase splitter. It also had an on-board regulator for the front end circuit. Of course, this was a high-feedback design based on some op-amp. It was funny to listen to -- we called it "goo and shine". You could hear the "gooey" signature of the electrolytic cap used across the voltage reference, along with the zingy "shine" provided by the high feedback regulator circuit. I redesigned the regulator to a zero-feedback design using only PP film caps, and then that amp really began to sing. Running it as a triode output stage made it one of the best amps I'd ever heard up to that time, although it would only put out 18 watts per channel.
Finally, New York Audio Labs used to make a phono stage called the "Super-It". It had a JFET input for low noise and a tube as a cascode for the JFET. I think it had another tube or two for more gain. Once again it had a high-feedback op-amp based regulator. Replacing with a zero-feedback regulator (again using JFETs so that a PP cap would be sufficient to bypass the voltage reference), and that thing sounded absolutely superb. (I also replaced all of the coupling caps with better parts.)
Anyway SY, you may want to try some experiments. E-mail me the schematic of your existing regulator, and I'll e-mail back something that I think will sound better. You can try it and report back to me.
Cheers,
Charlie Hansen
There have been numerous posts in addition to Bob Cordell's letter in Stereophile claiming that the passive components in chips are of equivalent quality to stand alone discrete components. Are any of you guys going to post accurate detailed specifications for actual chip passive components or are we just supposed to take your word for it that they're wonderful?
Stand and deliver, guys. You're always chanting, "Prove it! Prove it!" to those who listen...now the shoe's on the other foot. I think most people (myself included) would say that some actual numbers on those parts would be useful, so either post hard numbers or desist with extolling the virtues of chip passives. I, for one, am unpersuaded.
Grey
Stand and deliver, guys. You're always chanting, "Prove it! Prove it!" to those who listen...now the shoe's on the other foot. I think most people (myself included) would say that some actual numbers on those parts would be useful, so either post hard numbers or desist with extolling the virtues of chip passives. I, for one, am unpersuaded.
Grey
janneman said:
I can't get your link to work for me. Can you give me a URL? I'll be happy to respond in detail when I know what you are referring to.
In general, we use some ICs in the audio path of our entry level products to keep the price down. It turns out that there is one Analog Devices part that is very similar to the circuit we were using five years ago (we've moved on since then). By using two of these parts together in a way that Barrie Gilbert never intended, we were able to create a zero-feedback, balanced gain stage.
(I had a brief e-mail correspondence with Mr. Gilbert. He is a brilliant engineer, but he thought I was fool to avoid the use of feedback. His main argument was that no matter how linear of a circuit I made, the distortion could always be lowered by using feedback. I had the same question for him as I do everyone else here -- if I can achieve 0.001% distortion without feedback, why bother adding feedback? Do you really think that we can hear 0.001% distortion, but not 0.0001% distortion?)
- Status
- Not open for further replies.