Impressive THD and noise..... If you need a gain of one circuit.
THx-RNMarsh
As a unity gain buffer, it's a dream come true.
That's the game of the competition. Research and engineering. If a wine completion was made with supermarket graphs you can be sure someone will do something spectacular. As said TIP3055 was doing 8 MHz is 1980. That was nearly state of the art at the time. John Dawson said officially 4 MHz. I love 3055's as they punch well above their weigh. Some say the 3055H is less good sounding as it has traded ruggedness for bandwidth. 5534 is OK and that's the challenge. I would not reject signals going in through pins 1 & 8 using BC337 for example. 5534 is about 4 nV/Hz. With 6 x BC337 < 1 nV/Hz ? 3 devices should be about 0.7071 the noise of one device (0.6 nV/Hz). You can attach the transistors via the normal input transistors of the input pins. When I did this it was lower in noise, however the sound was less good. Doing that is a dream come true technically as the self servo action of an op amp helps bias them and set DC offset. If with TL071 or LF351N you will need to bias them as the J FET's do not interface in the way of 741 and it's generics. In the old days it was worth the effort to do that. Even a 741 or worse still LM324 can be made to perform to a reasonable standard. The output in SE class A is important. A 741 with those tricks was often used in studios and was respectable in performance (microphone amp). The designer often highly skilled in getting something totally unsuitable to give it's best. I dare say some people can make the world best chips sound worse ? Don't fool yourself. A Shure microphone through the worlds best chip will not begin to beat a ribbon microphone through transistors plus 741. Although openness can be lost that is not the same as making it happen when not there. Sometimes the Shure is the right microphone to use, I don't say otherwise.
I can't think of anyone right now which has made a clear statement that speakers / transducers have achieved perfect transparency and surpassed all upper limits.
In fact, I'm not sure if anyone in some sort of advanced position believes in this. The article I linked does to a certain extent. You know, I vaguely wonder if he's right or not.
I don't really have any specific 'position' in all of this, I'm only interested in the truth.
I have seen internet users say that perfect frequency response and perfect CSD, achieved via acoustic treatment, will create a a perfect speaker setting.
I don't really want to say particular names.
In any case, that's not the critical part to me personally, neither is the microphone or the ADC, it's the DAC and the amplifier.
I believe I can hear DAC and amplifier differences and I really like them, I think they're aesthetic and emotional.
However, there isn't very much science to specify the differences. I'm looking for it.
In this thread we discussed THD spectrum and settling time recently.
Note, when I say I believe in the differences, it's not true belief, it's like an annoyed belief, if I had some kind of way to know that what I hear is real, I'd feel happier.
There is a very small possibility with the NE 5534. That is to deliberately phase shift and use positive feedback. I doubt if anyone would have a use for that? It would be open loop. A super fast comparator perhaps or Schmidt trigger? You might find that LM339 is faster. Has anyone ever used one as an op amp? I guess you end up with an LM324 in SE with tricky compensation problems and buckets of noise? It will allow 50 V 50 mA through the open collector output.
This is the other bootstrapping that is interesting. Sometimes more difficult in reality. It should be possible to do this with TDA 2050 and a set of dumpers or even your favorite Douglas Self or whatever amplifier. You might get a Quad 405 to swing 60 Vrms. Now if that was 2N3055 we might swing 40 Vrms. As we will most likely be in class AB why not make the amplifier small devices work super hard and get this high power. Or we build a Crown DC 300?
Pimping an Op-Amp for wide voltage swings joebrown.org.uk
It's a difficult and interesting problem that lots of people are working on. Even more difficult when one considers that the target may be different from recording to recording.
A perfect example of the intellectual vacuousness of equating that, because some problems have been long solved (e.g., how to get an electrical signal from point A to point B and make it larger or smaller without audible effect), there are "many people" who believe that ALL problems have been solved. No one believes that, it's just an audiophile trope, akin to the proverbial (and fictional) engineer who believes that THD is the only important parameter.
The problem with distortion is it has a number of routes that create it. It is very hard to say "distortion" and give an accurate picture. Simple curve distortion is not so bad. Some people say that triode valves are very linear. This is not really true. They do not have much gain and the pentode device says why. Some pentodes are great in fake triode ( EL 34 ). The gain exchanged for better linearity. You might say the triode has internal negative feedback although people will argue about that. Next we might by various methods convert our pentode to triode. The latest in the RH amps Kitic is using shunt input feedback to g1 as a neglected version ( a la Schade of RCA 1938 for 807). We now have the problem of not exactly knowing what we are listening to. Is it distortion %, distortion spectrum or output impedance? Often it is all three. Sadly the output impedance might be the big cause and will form a prejudice in the builders mind. A little knowledge is a dangerous thing and that applies to all of us no matter how experienced. With the pentode we have a unique possibility. That is to find it's UL compromise. This is where the pentode /triode curves are almost placed in mirror image and a middle curve is found. On a curve tracer these curves can be seen and by the visual route a curve inferred. It must have occurred to someone in 1927 that this would be possible. In 1935 Blumlein says how the new Harries Beam Tetrode might be made to work this way. The Hafler patent is the one most often referred to. UL is the most difficult to use (Not so SE with no loop feedback). Like the famed NE5534 not always a success and for identical reason of ringing.
All of this had become a science when valves. When transistor amps came alone some designers noticed problems and cured them. Leak did AB public tests and were convinced they were no different valve or transistor. I would say they were justified for the ordinary listener to say that. Many transistor amps came out that were much more than awful. This caused public rejection of the new devices. On top of this they were hyper unreliable. Some made great efforts not least John Curl and I am sure a vast number of friends of to name these new distortion mechanisms. Some remain a partial mystery. Anecdotally we cure them. Speed seems our best tool and bags of surplus current. As with most witchcraft as many things that support a theory also disprove it. To be clear UL almost belongs in the transistor camp of problems. Please don't say this amp and that amp doesn't have this, we know that. Dynaco?
My own take on this is a very recent and I am ashamed to say an obvious thing. I built some very simple speakers. The only real difference between amps via them is tonal. The Audio-Lab amplifier I usually detest is very nice if a little bright. Then I corrected their frequency response. Now they are a different animal. They have become amplifier fussy. It is only too obvious this should be true. I needed to do it this way to find out. Transistor class AB amps and valve class A amps share no great common ground. Problem is the same words are used to describe a number of illnesses. Loudspeakers have what looks like curve distortion. Alas that is the minor distortion of them. Re-radiation and room interaction plus polar distortion are in the picture. That's before we consider how they attack the amplifiers. Some say the back EMF is irrelevant. That can't be true. They will not produce square-waves ( most). Suddenly that doesn't matter if them. It does matter. Even the ugliness of a speaker makes for less pleasure. That might be going too far and is not a distortion I can name. Mood distortion? I guess that the greatest one of all?
People say cheese give funny dreams. I never found that.
I tried the MUSES8920 and it sounded nearly identical to a JRC chip, I think JRC4580, not sure but a specific JRC chip I had in my case. If I didn't know it was a Muses I would have stopped listening, no difference, I kept listening and tried to hear something, at last, I recall I could hear some kind of hazy natural sound connected to a specific CD player. This is around 1.5 years in the past now.
Recently, I finally tried the MUSES01, even though I spent a fair bit on it, I was hoping it would sound normal, because the specs are normal, LoL.
I think it's definitely a little coloured or different. In fact it was so coloured it almost sounded like some reverb processing, I wondered a little if it was a fake I received.
After lots of trial and error, fake OPA627 in the buffer channel provided the best synergy with it for me.
Voices seemed involving, the female singer wants to pull out her heart and deliver it to you through the speaker, sort of.
Not normal.
Not transparent either, lme49990 is clearly more transparent in my equipment, in my listening.
"They claim they are made from better materials so they may not even measure better as such."
The day of reckoning cometh.
Last edited:
Actually Nige, chronologically speaking, the above statement is incorrect. In 1978, Pioneer was selling RETs in their amps, and these were 30 MHz+ devices from Sanken.
Sansui and Kenwood also used their versions of Hift transistors (from NEC and Toshiba), so even by then, the 2N3055 was already antiquated.
Not to even mention Motorola's MJ15xxx series, as used in many US made amps and just as many German made amps (e.g. Dual).
Blumlein.
Isaac Schoenberg lent Mr B to television research. We used it as a cover for radar development. The 405 lines of Blumlein far better than the German system. This allowed pre war joined UK/German companies to make valves with no other obvious use. Blumlein died testing the idea that Goering said cost them the war. Blumleins radar for bombers. Churchill demanded it and the slightly primitive Blunlein system had the small advantage of actually working. The success was posthumous and his family never told (a disgrace).
Isaac Schoenberg said for patent busting the RCA royalty Alan would never had needed to do another thing and still be paid. Isaac Schoenberg said Allan was not content to save them money, he made the cutting lathe work better. Alan developed the op amp circuit called the Long tail pair, he most likely was the inventor. Many fancy cables look like Blumlein designs. He was a sane Tesla.
https://www.youtube.com/watch?v=rqaMiDqE6QQ
Isaac Schoenberg lent Mr B to television research. We used it as a cover for radar development. The 405 lines of Blumlein far better than the German system. This allowed pre war joined UK/German companies to make valves with no other obvious use. Blumlein died testing the idea that Goering said cost them the war. Blumleins radar for bombers. Churchill demanded it and the slightly primitive Blunlein system had the small advantage of actually working. The success was posthumous and his family never told (a disgrace).
Isaac Schoenberg said for patent busting the RCA royalty Alan would never had needed to do another thing and still be paid. Isaac Schoenberg said Allan was not content to save them money, he made the cutting lathe work better. Alan developed the op amp circuit called the Long tail pair, he most likely was the inventor. Many fancy cables look like Blumlein designs. He was a sane Tesla.
https://www.youtube.com/watch?v=rqaMiDqE6QQ
8 Mhz to 30 isn't much when a 20 khz bandwidth.
My reason for choosing 2N3055 is how "old dog" they are. Do you know most of the super ordinary amps I have heard use very fast transistors. Nothing to do with them. Conversely.....
They say you can't teach an old dog new tricks. I say you can.
Also when saying 30 MHz important we are saying versions of H C Lin amp are they only valid type ( yes I know he wasn't LTP, rumour is he invented modern amps circa 1960 for the military). A proper Quad 405 answers that. I bought some 3055's for 3 for $1 recently. I blow them up first when designing things.
I would have thought that a wider bandwidth at least implies faster switching.
But where we agree, Nige, is that with wide badwidth transistors the issue of high frequency stability is more pronounced than with their older brethern, which could do like 4 or 6 MHz "only". Much more robust, they were.
But where we agree, Nige, is that with wide badwidth transistors the issue of high frequency stability is more pronounced than with their older brethern, which could do like 4 or 6 MHz "only". Much more robust, they were.
There is a thread on DIY audio where Ahuja compensation is mentioned :http://www.diyaudio.com/forums/soli...r-amplifier-design-handbook-douglas-self.html
Nigel, bipolar transistors really did improve from the 60's (2N3055) to the late 70's RET's) and the amp performance changed considerably because of this.
For the record, very high F(t) complementary power devices DID EXIST in the late '60's but they cost about $100 each. They just were not practical.
Design engineers had to cope with better safe area usually giving lower F(t) with a few (expensive) exceptions, so an F(t) of 1-4 MHz was fairly standard before 1978.
It took the Japanese to make the breakthroughs that made high F(t) and therefore higher slew rate amplifiers relatively easy to make.
For the record, very high F(t) complementary power devices DID EXIST in the late '60's but they cost about $100 each. They just were not practical.
Design engineers had to cope with better safe area usually giving lower F(t) with a few (expensive) exceptions, so an F(t) of 1-4 MHz was fairly standard before 1978.
It took the Japanese to make the breakthroughs that made high F(t) and therefore higher slew rate amplifiers relatively easy to make.
Last edited:
I'll remember that the next time I use an op-amp open loop. Which is like never.
- Status
- Not open for further replies.