I believe that whenever there is a minimum for something to become operational, then there is most probably also a maximum, when the desired effect starts to turn into its own opposite.
I do not know what is the minimum, or the maximum, I simply assume they exist, as they exist everywhere else.
Defining the maximum, off hand, is in my view like playing lottery, pure guesswork. Many factors are involved, from the size, to type, to specific behavior of each type, making it impossible to actually know each and every combination possible. One simply has to try.
Ever since the beginning of the 70ies, Japanese tweakers had proposed and used the idea of a composite "super" capacitor. Say you need a 10 uF cap someplace - their idea is to use in parallel one cap made of say polyethylene, another made from polyester and a third made of say polycarbonate. They should preferably be all of the same value, i.e. 1/3 of the desired end value. The theory is that in this way, one avoids any specific characteristic and /or sonic imprint of any single capacity, while combining their good sides into one. By this method, you arrive at one value of a "super" capacitor.
And, if for some reason, they cannot all be of the same value, then make sure that the biggest in the lot be made of polycarbonate, since these will, by virtue of the material they are made of, be the fastest, sometimes by far.
I have tried this out just once, since I prefer DC coupled circuitry, and then on a phono RIAA device I made according to Burr-Brown's project specs, though somewhat upgraded by adding current booster transistors. Two were made, one directly as described by BB, and another by using "super" capacitors.
I have to admit that the one using "super" capacitors was that n-th bit cleaner and clearer, I'd say just over the threshold of audiability. Small, but noticeable.
Last edited:
Never heard them but I loved the idea, both the lens allowing a planar radiator behind and the OTL amp, which made perfect sense.
Floyd Toole remarked to me that Harold Beveridge was in some ways too smart for his time. It's nice to see his sons carrying on.
OTOH I tried, years ago, to persuade my friend the local audiophile to allow me to construct a direct drive amp for his then-Acoustat panels. This guy is heavily motivated to have things no one else has, modifies everything, and initially he didn't realize I was talking about getting rid of the stepup transformer completely. I could see him agonizing, as I described that the amp would be developed and made absolutely stable by using a carefully duplicated dummy load based on small-signal measurements of the panels.
But in the end he demurred. "No, you'll have to get your own speakers."
When I visited him recently after a long lapse, he had moved on to the next-to-largest Sound Lab 'stats (the largest wouldn't fit in the room). If anything he is more nervous about arcing now than ever, and despite the big amp he had which he deemed necessary to properly drive the speakers, he was very cautious about levels when I gave him some materials with which he was not already familiar (despite my assurances that there was not that much low frequency energy).
I think I'll stick with conventional cones, surrounds, and motors myself
Tim de Paravacini offered to lend me a pair of Quad electrostatics direct driven by one of his valve amplifiers . He then said the big surprise was the rising treble when used that way if uncorrected . The very soft sound of the Quads a choice of the company . One could almost say a transformer which normally would be seen as rather poor was in symbiosis . I don't know much more than that . I want to make a pair soon . I fancy just to make a mid top unit . I have some ideas from a friend who when a school boy visited Alex Shackman who made low priced speakers using an electrostatic tweeter ( Dynastatic ) .
DVV . Boosted op-amps . I guess you might be getting lower distortion if you take the load off of the op amp ? 10K is a better load if it is possible . I use op amps in parallel , easy enough and seems better . MC33079 is cheap and rather OK . 33079 is reasonably happy doing wrong things , sometimes wrong works very well . There are special op-amps for driving headphones . I wonder if they are better than some would imagine ?
If you find a rare example of a house where Quads are allowed to work switch the lights off . Ask a friend to find them ( use BBC R3 or similar ) . More difficult than you might think . I once installed a pair half way up a 40 foot Georgian room ( 17 apart slightly angled in , touching the side walls which can work ) . The lady who bought them insisted before I had a chance to ask . I had hoped just to try them , she insisted based on physics we should . For her sins I think she taught Margret Thatcher ? Mrs T invented a desert called Instant Whip which I remember enjoying as a child . Now you probably would need exorcising if feeding it to kids !
DVV . Boosted op-amps . I guess you might be getting lower distortion if you take the load off of the op amp ? 10K is a better load if it is possible . I use op amps in parallel , easy enough and seems better . MC33079 is cheap and rather OK . 33079 is reasonably happy doing wrong things , sometimes wrong works very well . There are special op-amps for driving headphones . I wonder if they are better than some would imagine ?
If you find a rare example of a house where Quads are allowed to work switch the lights off . Ask a friend to find them ( use BBC R3 or similar ) . More difficult than you might think . I once installed a pair half way up a 40 foot Georgian room ( 17 apart slightly angled in , touching the side walls which can work ) . The lady who bought them insisted before I had a chance to ask . I had hoped just to try them , she insisted based on physics we should . For her sins I think she taught Margret Thatcher ? Mrs T invented a desert called Instant Whip which I remember enjoying as a child . Now you probably would need exorcising if feeding it to kids !
My comment on op amp current boosting was a general one. I did not claim that each and every one of them would become a super audiophile solution, but I did claim that each and every one of them will behave better when current boosted, even into relatively high loads, such as 47k.
Ultimately, do not take my word for it, Nige, I encourage you to try it for yourself. All it costs you is 4 resistors, 1 diode and 1 transistor per channel. Possibly more if you decide to develop the circuit further, but either way, it's still a simple job. Just make sure what you use for your output trannies - I strongly suggest some of those which can do lots of current (for their size, i.e. for TO-92 plstic packaging), such as, for example, BC 639/640, or MPSA 056/06, etc.
Note that much more than classic measurements will ever show, there's a difference in sound. They become clearer, as if they were more focused, so to speak. It's not exactly a revelation, but it's easily noticeable, especially on higher resolution gear.
Specifically, I base these comments on the op amps I have used, and these would be, in random order: OP37, LF356, LF353, LF357, AD826, AD829, AD818, AD828, OPA604, OPA2604, etc.
Just my 2 cents' worth.
And on purpose, a different posting.
For driving headphones, I honestly say to you: you would not believe at what your headphones are capable of until you have driven them for a high current capable amp.
For this specific purpose (and for driving low impedance loads), I strobly suggest two diodes to each supply line, a driver transistor and an output transistor. BD 139/140 will do just fine, so long as you swing them towards class A as much as your heat sinking will permit, at least 40 mA.
The subjective bass depth WILL surprise you at least in terms of what you previously thought your earphones could do, and coupled with much better bass control, your cans may well slip into a different overall tonal balance simply because now, there's some muscle behind the bass. My own Sennheiser 598 can really surprised me, they are often said to have wonderful mid and treble (true), but are somewhat lacking in the bass region, also true when driven from the usual sources. However, with a dedicated headphone amp, which can provide all the current they could ever use before making you permanently deaf, this changes in no small degree. They don't invent bass, but what bass there is, they will do with ease, and sometimes, it's a VERY deep bass, down around 30 Hz.
But, as ever, there is a caveat. Once you do it, and hear what your say cheap cans can do, the bug raises its head and makes you wonder - if the cheapos can do that, what would you hear with a pair of better phones? Where "better" also means "more expensive"?
For driving headphones, I honestly say to you: you would not believe at what your headphones are capable of until you have driven them for a high current capable amp.
For this specific purpose (and for driving low impedance loads), I strobly suggest two diodes to each supply line, a driver transistor and an output transistor. BD 139/140 will do just fine, so long as you swing them towards class A as much as your heat sinking will permit, at least 40 mA.
The subjective bass depth WILL surprise you at least in terms of what you previously thought your earphones could do, and coupled with much better bass control, your cans may well slip into a different overall tonal balance simply because now, there's some muscle behind the bass. My own Sennheiser 598 can really surprised me, they are often said to have wonderful mid and treble (true), but are somewhat lacking in the bass region, also true when driven from the usual sources. However, with a dedicated headphone amp, which can provide all the current they could ever use before making you permanently deaf, this changes in no small degree. They don't invent bass, but what bass there is, they will do with ease, and sometimes, it's a VERY deep bass, down around 30 Hz.
But, as ever, there is a caveat. Once you do it, and hear what your say cheap cans can do, the bug raises its head and makes you wonder - if the cheapos can do that, what would you hear with a pair of better phones? Where "better" also means "more expensive"?
That is interesting and you did say it before Dvv . I think it is more than likely the big op-amp up grade might be a booster ? Might it be argued that a humble op-amp with booster will outperform an expensive one ( ears ) ? I wonder also if power op-amps ( better than TDA 2040 ) might make decent high level op amps ? The distortion curves are when giving high power . It will be lower into a smaller load , also cables can be very long if required . Some power could be drawn off in class A with a pull up device ( resistor or CCS usually to the - ve rail ) . Also potted down to use the sweet spot ( 5 W equivalent , ) . You can do that with op-amps also . I have never seen it done either way .
I can't answer that, Nige, because I've never worked with so-called high current op amps (that I know of, anyway).
With the scheme I am suggesting, you determine the bias current yourself. In my several year long experience, about 40 mA is quite enough to kepp you in pure class A 99.9% of the time for cans of 24 Ohms and more, and 100+% for cans of 50 Ohms and above. Remember, cans are usually declared as xx dB at 1 mW.
With the scheme I am suggesting, you determine the bias current yourself. In my several year long experience, about 40 mA is quite enough to kepp you in pure class A 99.9% of the time for cans of 24 Ohms and more, and 100+% for cans of 50 Ohms and above. Remember, cans are usually declared as xx dB at 1 mW.
Coolest meter ever! And I like the Venturi heat sinks, and the use of copper.
But I wonder why the noise is so high? I calculate 356uV rms output noise based on the HiFi News measurement sidebar for A weighted noise below 1W (presumably into 8 ohms, hence referenced to 2.83V rms out). And from the "distortion" plot, which begins at 1W, for the stated 0.015% presumably with a wider bandwidth than A weighting, I calculate 425uV rms. Folks, this is not particularly low-noise, and conceals whatever may be going on with distortion per se.
And the highish noise is not because the gain is high. In fact that is a somewhat low 23.3dB. I don't know what the magazine's convention for unweighted bandwidth is, but supposing it's 30kHz this translates into an equivalent input noise density of 167nV/sqrtHz, which is very high (equivalent to a ~1.7M resistor
).I think he'd better send me a pair immediately so I can figure out what's going on
Best amplifier in the world ....
Only fair he sends one out , i'm sure you could make it the quietest best amp in the world ...
Only fair he sends one out , i'm sure you could make it the quietest best amp in the world ...
Coolest meter ever! And I like the Venturi heat sinks, and the use of copper.
But I wonder why the noise is so high? I calculate 356uV rms output noise based on the HiFi News measurement sidebar for A weighted noise below 1W (presumably into 8 ohms, hence referenced to 2.83V rms out). And from the "distortion" plot, which begins at 1W, for the stated 0.015% presumably with a wider bandwidth than A weighting, I calculate 425uV rms. Folks, this is not particularly low-noise, and conceals whatever may be going on with distortion per se.
And the highish noise is not because the gain is high. In fact that is a somewhat low 23.3dB. I don't know what the magazine's convention for unweighted bandwidth is, but supposing it's 30kHz this translates into an equivalent input noise density of 167nV/sqrtHz, which is very high (equivalent to a ~1.7M resistor
I think he'd better send me a pair immediately so I can figure out what's going on
Well, that is mean. Plus that style of bathroom scale is usually not very accurateSomeone compared it once to a weighting scale. ...the bathroom style kind
A line of early aftermarket autosound amps I worked on a little for JBL, circa 1990, looked a lot like carpet sweepers, one of those old Bissell jobs, something noted as well by Tom Jacoby iirc.
- Status
- Not open for further replies.