I am not 100% sure, but I think it was (a bigger guru in the recoding industry) Rubert Neve.
With kind regards,
Bas
one follow up. I just asked my co-worker who is also deeply into pro-audio. According to him, Rubert Neve had a hand in the NE5534 design. The first chip that would become the NE5534 was the TDA1034.
you can read some history here: neve
A nice note is, that int he newer NEVE recent design they do use the pull down Class A tricks, to force the chip to be in class A. Neve is a company that truly goes for sound quality, but they still swear with the NE5534's when they use opamps. Neve also use discrete class A designs in some applications.
With kind regards,
Bas
you can read some history here: neve
A nice note is, that int he newer NEVE recent design they do use the pull down Class A tricks, to force the chip to be in class A. Neve is a company that truly goes for sound quality, but they still swear with the NE5534's when they use opamps. Neve also use discrete class A designs in some applications.
With kind regards,
Bas
Last edited:
I don't think Neve actually designed the 5534. He may have spec'd it through the roadmapping process and been the lead customer though (normal approach in the semi industry). Neve was running a professional sound company - the 5534 was put together by a very good analog guy who was used to working with silicon.
Dear Bonsai, AS far as I understood Neve didn't designed the silicon indeed, but the first TDA1034 was based on a Neve op-amp design. The TDA1034 was followed up by the NE5534.
About the class A biasing. See quote from this review: Rupert Neve Designs Portico
"The circuitry is built almost entirely using conventional-sized components, with the old familiar NE5534 op-amps doing the bulk of the work, supplemented with separate transistors in appropriate places. Rupert Neve's designs traditionally employed single-sided class-A topologies, and it is hard to square that approach with these Portico circuit boards covered in op-amps. The secret missing ingredient is that he uses a circuit technique with the 5534 op-amps which offsets the DC point of their output stages, so that for signals below about 0dBu they are effectively running in a single-sided class-A mode. This removes crossover distortion artifacts completely and is a significant contributor to the sound of this preamp."
Wonder if they bias through PIN5, or the add current on the output.
With kind regards,
Bas
The 'winners' of D. Self depend on noise and (measurable) distortion. H. ten Pierick has a total different method!!!
Of course 'bad sound' is seldom produced by op-amps. Life is not that simple. In the AD and DA-domain, eg. the jitter on the clock is a much, much more important issue!
It might sounds a bit black & white, and not very musical, but I am agree with Mr. Self. For a one opamp unity or 2x gain application noise might not be such a big issue. But in e.g. an active speaker system, all noise add's up. The NE5534 is hard to beat here. Personally I would choose low noise over everything.
I am agree with you on this "'bad sound' is seldom produced by op-amps" A good designer can get much better results out of a NE5534 opamp, then a bad designer out of a OPA627.
With kind regards,
Bas
Hi.... ,
until some years ago, the NE5534 was unbeatable (4nV/sqrtHz) and it's THD (-100 dB ?) was not that bad. Ti (BurrBrown) used them in their high end DAC stages and made the measurements on it (PCM1792 (up to 130dB SNR !!!!)).
But since some years, there are the LME49xxx Opamps with ~ 1nV/sqrt Hz and THD+N better than 120dB.....
But in the prof. audio industries, the only limiting factor was the noise (Yeah, wanna build an ADC with 130dB SNR
it's hard task with 4.5nV/sqrtHz... that's where 1kOhm resistors make more noise than the Opamp.... But for "standard" Audio, the NE5534 is a good choise (and its price is affordable
Try also OPAx34 (more noise, better THD) and the LM4562 (same noise, better THD) or the MC3307x devices (same noise, same THD, but DUAL and QUAD)
until some years ago, the NE5534 was unbeatable (4nV/sqrtHz) and it's THD (-100 dB ?) was not that bad. Ti (BurrBrown) used them in their high end DAC stages and made the measurements on it (PCM1792 (up to 130dB SNR !!!!)).
But since some years, there are the LME49xxx Opamps with ~ 1nV/sqrt Hz and THD+N better than 120dB.....
But in the prof. audio industries, the only limiting factor was the noise (Yeah, wanna build an ADC with 130dB SNR
it's hard task with 4.5nV/sqrtHz... that's where 1kOhm resistors make more noise than the Opamp.... But for "standard" Audio, the NE5534 is a good choise (and its price is affordable Try also OPAx34 (more noise, better THD) and the LM4562 (same noise, better THD) or the MC3307x devices (same noise, same THD, but DUAL and QUAD)
Who/when created NE5532/34/35 ?
I'm not sure when the NE5532/34/35 family was first created. Philips bought Signetics in 1975.
I'll bet Hans R. Camenzind had a part in it's invention and design. He was at busy at Signetics with the 555 timer in 1971. Patents were not used in Silicon Valley back then. I sent him an email...
His website Designing Analog Chips by Hans Camenzind has a free pdf book Designing Analog Chips.
IEEE Spectrum: 25 Microchips That Shook the World
It's from an era of hand-cutting sheets of Rubylitha masking film to make chips !
I'm not sure when the NE5532/34/35 family was first created. Philips bought Signetics in 1975.
I'll bet Hans R. Camenzind had a part in it's invention and design. He was at busy at Signetics with the 555 timer in 1971. Patents were not used in Silicon Valley back then. I sent him an email...
His website Designing Analog Chips by Hans Camenzind has a free pdf book Designing Analog Chips.
IEEE Spectrum: 25 Microchips That Shook the World
It's from an era of hand-cutting sheets of Rubylitha masking film to make chips !
What concerns me is whether those THD figures are really necessary in diy audio circuits (as opposed to consoles). Clearly if a signal is passing through dozens of opamps, the distortion is cumulative. But for diy circuits with just a handful of devices, I'm sceptical (but nevertheless open to persuasion) that we need to aim for such low figures. In the case of the 5534 the price is enough to sell it, but for the LM4562 which is perhaps 10X as much, isn't it overkill ? My view is that a lot of it is psychology - because the paper specs of the LM are so amazing, the power of suggestion makes it sound good to designers.
I second Jan's request - can you tell us more about his method?
This much is obvious - bad sound is produced by systems, not so much by individual opamps. So the focus on ultra-low THD+N opamp performance for me takes away attention from what matters just as much - design and layout.
No, life is not that simple. There are different kinds of jitter - some are more important than others. Jitter caused by random noise is much less of an issue when compared to data-correlated jitter. Our brains are adept at picking out patterns and simple noise doesn't produce any.
Hi Bas - a Benchmark DAC1 upgrade? That intrigues me because about a year ago I read a very long thread (I think it was on headfi) where the designer of the DAC1 explained that no upgrades were possible for the DAC1. Every tweak, he assured us, would degrade rather than upgrade the sound because he'd covered all the bases. In particular the I/V stage uses NE5532 rather than anything else because that was the best part in listening tests.
I have a really cheap PCB using a TDA1543, I rarely listen to it as I find it much less satisfying than my heavily modded AD1955. It has passive I/V - initially it sounded clearer than my AD1955 but listening more I noticed the acoustic space isn't reproduced well.
I agree with that, he's a top flight designer and makes an entertaining read too. Maybe someday I'll write a complementary book to his from my 'subjectivist' standpoint
In my experience, every IC op amp sounds different. I gave up on the 5532/5534 over 30 years ago. However, we found that adding a jfet input stage though the compensation pins made an acceptable op amp. Siliconix was the first to show how to do this. We ultimately put them into equalizers for the Wilson WAMM system with some success.
We're back to Mullard and the TDA1034
The roots seem to go back to the Mullard TDA1034 (TO-99 metal can package). Hans Camenzind wrote back "... I know nothing about it " which means Signetics was merely manufacturing them (after Philips acquired Signetics in 1975). So the chip design is from across the Atlantic
maybe some folks from the UK can fill in the gaps.
Late 1976 they (TDA1034) went into a new custom console for Pink Floyd
TDA1024 (around 1974) later became known as the NE5532.
The roots seem to go back to the Mullard TDA1034 (TO-99 metal can package). Hans Camenzind wrote back "... I know nothing about it " which means Signetics was merely manufacturing them (after Philips acquired Signetics in 1975). So the chip design is from across the Atlantic
maybe some folks from the UK can fill in the gaps.Late 1976 they (TDA1034) went into a new custom console for Pink Floyd
TDA1024 (around 1974) later became known as the NE5532.
typo?
+-21Vdc supply rails can at best output 14.5Vrms. Into a 600r load this maximum drops to about 12Vrms and that demands an output current of 28mApk
The TDA1034 was designed by Philips. At one point Philips bought Signetics and the TDA1034 re-appeared as the Signetics NE5534. Since it was hugely successfull, everybody and their grandmother (sorry grandma
) jumped on the train and designed a 'me too 5534', which had nominally the same specs but was different inside and sometimes also measured differently.jd
