Very low noise op amp recommendations?

[bikerhifinut]

Originally Posted by mhennessy

But is it any better than an NE5543?

Assuming your workshop has a reasonable supply of passive components, it would take 30 minutes tops to knock up something like the Self circuit (don't need the sub-sonic filter for a quick test) on a breadboard or even a bit of Veroboard. Why not do that first, before sourcing obsolete ICs?

That will establish a baseline - a reference point. After than, feel free to experiment with different op-amps and different topologies - you never know, you might find something you prefer, even if it's not as good on paper.

I'm worried you're on the edge of an audiophile rabbit hole here. Regular audiophiles get it bad enough, what with all those magazines, forums and sometimes iffy marketing materials, etc, but it's even worse for the DIY audio crowd. Many of them love finding the most obscure and difficult solutions, as if the pain of the journey makes for a better end result - and in their minds, it absolutely does. Which is fine in itself, but unfortunately it often develops into a faith-based believe system to push - sometimes quite forcibly - onto others. Which is why I no longer frequent the DIY audio forums...

To steal the signature of a well-known poster from the Usenet days, music is art, audio is engineering. Wise words

Don't worry on that score Mark!
I lurk on a couple of audio fora and often end up biting my virtual tongue, especially when people advocate mass replacement of electrolytics because they might be time expired even if the equipment is working properly, as a for instance.
That and the "signal path" capacitor arguments. and as you have mentioned elsewhere whether a caddock or a Shinto (Shinko?) or whatever resistor sounds best...………. I can't tell the difference between a standard Alps RK27 and a fancy switched attenuator so its wasted on me!
I'm simply curious so this is all of interest to me. I have a very good phono stage in my regular solid state commercial preamp and I doubt whether I will find an improvement on that, its my normal everyday listening.
As I said, I will be trying out the simple stage with a 5534 and I must get a good look at the Douglas Self design, I knew about Self a long time ago and completely appreciate his approach to design.
I have fun playing around with different circuits and component technologies but I never expect to improve on proven and reliable designs. That's not the point for me.

Thanks to all of you though.

Very best regards.

Andy.

[Synchrodyne]

Originally Posted by turretslug

The HA12017 was quite heavily lauded in one of Ambit's famous catalogues of the late '70s- whilst I enjoyed reading some of William Poel's insightful and shibboleth-slaying blurb, there's always the chance that they just wanted to sell the things.... It does have unusually high supply voltage capability, no doubt fitting into the then must-allow-vast-pre-clipping-capability tick-box. If it's cheap enough, no reason not to try it and SIL means easy layout!

Originally Posted by mhennessy

But is it any better than an NE5534?

Attached the HA12017 page from Ambit Catalogue #4. Note the “purple prose”: “At risk of being contradicted, the HA12017 is the best audio preamp IC yet. Certainly it has no peer at the price – and even when compared to the NE5534 series devices, the HA12017 is no worse - and by virtue of its low price and 10v output capability - probably a lot better.”

Certainly, Bill Poel was refreshingly iconoclastic, but in this case he seems to have overlooked the adage that one should choose one’s target wisely.

As far as I know, the HA12017 was more-or-less an improved version of the widely-used Toshiba TA7122 (which I think had a Hitachi counterpart with a four-digit number). This was essentially the three-transistor circuit in integrated form, intended primarily for the MM/RIAA job. I somehow doubt that it would or even could have done better than the NE5534 in that role, or even better than the erstwhile LM381, which was also included on the same Ambit page. The LM381 was also more versatile. But these low-noise preamplifier ICs were effectively made obsolete by the NE5534/2, which delivered about the best realizable MM/RIAA performance via an industry standard, multi-use opamp that could also be deployed for most other functions in a domestic audio control unit. Nonetheless the 1979 Signetics Applications, under stereo preamplifiers, referred not only to the NE5534, but also the NE542 and LM381 family.


Cheers,

[bikerhifinut]

Well that does get things in perspective.

I'd hazard a guess that in practice there'd be nothing to choose between that and a 5534 in a real world application.

A.

[Craig Sawyers]

Originally Posted by Ted Kendall

It's irritating in a way, but the fact is that the 5534A is still the best chip you can get for an MM preamplifier, even if it has been about for forty-odd years, and very close in noise terms to the best attainable with discrete devices, an approach which brings other and greater problems. The single stage EQ approach is the optimum, but to calculate the exact component values is non-trivial exercise, because of the complexity of the relevant equations. Still, Self has done the sums, so that's that problem kicked into touch. He's written a tome just about electronics for vinyl, incidentally...

Just use Baxandall's values for R and C and you'll be fine. You don't need to get under the bonnet of calculating the values unless like me you are a bit of an analysis masochist,

Baxandall also shows an inverse RIAA. You plug that into the RIAA input and use your favourite audio sweep generator (or audacity via your computer sound card) and you can check your RIAA accuracy.

There is also a good paper on inverse RIAA here http://www.hagtech.com/pdf/riaa.pdf . Both Baxndall's and Hagerman's designs need careful value selection for accuracy of course.

Craig

[bikerhifinut]

Originally Posted by Craig Sawyers

Just use Baxandall's values for R and C and you'll be fine. You don't need to get under the bonnet of calculating the values unless like me you are a bit of an analysis masochist,

Baxandall also shows an inverse RIAA. You plug that into the RIAA input and use your favourite audio sweep generator (or audacity via your computer sound card) and you can check your RIAA accuracy.

There is also a good paper on inverse RIAA here http://www.hagtech.com/pdf/riaa.pdf . Both Baxndall's and Hagerman's designs need careful value selection for accuracy of course.

Craig

That's timely Craig and answers what my follow up questions would be after doing a prototype. It would be an essential part of my learning curve to see on a scope what the circuit was doing in real life.

Andy

[Radio Wrangler]

Originally Posted by Ted Kendall

It's irritating in a way, but the fact is that the 5534A is still the best chip you can get for an MM preamplifier, even if it has been about for forty-odd years.

It might always be.

The semiconductor industry was seduced by cellphones. Products that sell in zillions and have a production lifetime of about 6 months per model.

The vinyl revolution will never compete with that, and the 5534 is good for other applications.

Because the semiconductor industry is wholly concentrating on whatever the latest in gadget may be, there are whole areas of technology suffering neglect, and seeing the last devices suited to them going obsolete.

All the Avago RF ICs and diodes got cancelled with a single stroke a few years ago. Almost all the varactors have gone, except those in cellphones.

David

[turretslug]

Originally Posted by Synchrodyne

Attached the HA12017 page from Ambit Catalogue #4. Note the “purple prose”: “At risk of being contradicted, the HA12017 is the best audio preamp IC yet. Certainly it has no peer at the price – and even when compared to the NE5534 series devices, the HA12017 is no worse - and by virtue of its low price and 10v output capability - probably a lot better.”

Certainly, Bill Poel was refreshingly iconoclastic, but in this case he seems to have overlooked the adage that one should choose one’s target wisely.

As far as I know, the HA12017 was more-or-less an improved version of the widely-used Toshiba TA7122 (which I think had a Hitachi counterpart with a four-digit number). This was essentially the three-transistor circuit in integrated form, intended primarily for the MM/RIAA job. I somehow doubt that it would or even could have done better than the NE5534 in that role, or even better than the erstwhile LM381, which was also included on the same Ambit page. The LM381 was also more versatile. But these low-noise preamplifier ICs were effectively made obsolete by the NE5534/2, which delivered about the best realizable MM/RIAA performance via an industry standard, multi-use opamp that could also be deployed for most other functions in a domestic audio control unit. Nonetheless the 1979 Signetics Applications, under stereo preamplifiers, referred not only to the NE5534, but also the NE542 and LM381 family.


Cheers,

Thanks- that's refreshed me on the article and how it sounded somewhat defensive in nature. Whenever debate turns toward uncalled-for trench-digging, hyperbole and ad hominem, I start to smell a whiff of brass (short a).

[Craig Sawyers]

Originally Posted by Radio Wrangler

It might always be.

The semiconductor industry was seduced by cellphones. Products that sell in zillions and have a production lifetime of about 6 months per model.

The vinyl revolution will never compete with that, and the 5534 is good for other applications.

Because the semiconductor industry is wholly concentrating on whatever the latest in gadget may be, there are whole areas of technology suffering neglect, and seeing the last devices suited to them going obsolete.

All the Avago RF ICs and diodes got cancelled with a single stroke a few years ago. Almost all the varactors have gone, except those in cellphones.

David

That was when Broadcom bought Avago that the excellent high speed MMIC's, Schottky barrier diodes and sampling bridges got killed off. Like you say David - it it doesn't sell in products that have volumes in >100m semiconductor industries of today lose interest fast. Mobile phones, TV's, vehicles etc etc is where they see the market.

There are a few glimmers of hope. There is a vibrant research group at TI who are progressively rewriting the collection of audio products - opamps of varying flavours and performance, and upcoming a range of audio DACs.

Craig

[Radio Wrangler]

What was Avago used to be HP's internal semiconductor group. After HP split up into HP computers and Agilent (instruments), Agilent sold off the semiconductor group principally to a financial investment consortium. Its name changed then to Avago.

Avago, then took over Broadcom, a semi firm much more into mainstream stuff. It looks like David eating Goliath, but Avago was just the front for a huge financial consortium. Broadcom didn't have a chance against that much clout.

Avago now owned Broadcom and the rights to that name as well as their own. They thought Broadcom sounded more sensibe than 'Have-a-go' so they opted to change over to that name. They owned the rights and the web rights etc. so why not.

As a result many people think Broadcom bought Avago. But it was the other way round. Which makes it all the more irritating that one of their first actions was to dump almost the entire Avago portfolio. It makes more sense when you realise it's the financial consortium's hand in the puppet.

Broadcom got going with chips for digital TV. HP used some of them and I saw evidence that they didn't much understand phase locked loops. Since then they've become a big name in Wifi and other peripheral devices in computers, though they're probably known to more people for their devices instantiating ARM cores... the one in the raspberry pi for example.


David

[Jon_G4MDC]

If I could only write here what I want to say...
The light is wan. They must have been wearing blinkers.

[Restoration73]

If you want noise use MC1303.

It is possible to connect to a microphone input. Then the pre-amp output could be
filtered actively using a graphic equaliser set to give the RIAA curve.

[Synchrodyne]

Originally Posted by Restoration73

If you want noise use MC1303.

The MC1303 must have been one of the earliest ICs developed for audio preamplifier use, dating from c.1968.

As far as I know the first such IC for which it was claimed to match discrete circuitry on the noise front was the LM381A of 1972. I am not sure that it was much used in domestic equipment, although I have a vague recollection that Hi Fi News published a suitable circuit when it was new. There is some evidence that it was used as a microphone preamplifier in professional equipment.

Originally Posted by Restoration73

It is possible to connect to a microphone input. Then the pre-amp output could be
filtered actively using a graphic equaliser set to give the RIAA curve.

Basis what had already been said in this thread, that approach, flat gain followed by equalization, would be noisier than wrapping the equalization around the first stage. That said, Wireless World (WW) 1976 September, pp.81,82 carried an article for a reduced-noise MM/RIAA preamplifier that followed that pattern as a pathway to lower noise than could be achieved with the conventional single-stage series feedback circuit. It used a µA747 opamp in each stage. The SNR was said to be 65 dB referred to a 5 mV input. That was a case of non-QED when the theoretical best (noiseless amplifier) for a series feedback circuit is 80 dB referred to 5 mV, and H.P. Walker’s discrete single-stage circuit provided 78 dB referred to 5 mV. Oddly, the 1972 HPW article was quoted as a reference.

An irony was that the Mullard TDA1034 low-noise opamp (similar/identical to the NE5534) was announced in that same issue of WW.

Also in that WW issue were announced a suite of Hitachi ICs, namely the HA1137 (CA3089 type), HA1196 (MC1310 type) and HA1197 (AM radio), supplied by Ambit. It gave the impression that Ambit was a main supplier of Hitachi ICs in the UK, which would explain its later enthusiasm for the HA12017 MM/RIAA IC. Perhaps the HA12017 was welcomed by the Japanese equipment makers, basis their prior use of the TA7122, but it was probably a difficult sell in geographies where the NE5534 had established itself.


Cheers,

[Radio Wrangler]

Hitachi ICs hadn't been seen on the UK scene until Ambit started offering them, and their wording was a bit pushy suggesting they were improvements on the incumbents in those particular slots.

My FM tuner at the time was a PW design because someone at work had laid out a PCB and made several. CA3089 and CA3090 pretty plain-vanilla flavour that worked as well as needed in a good signal area. My concentration was in trying to eliminate the spurs from a frequency counter I'd done for it. So I never tried the Hitachi ICs. If I'd needed higher performance, I'd have gone discrete or to individual ICs for limiter stages.

At the time, the uA747 being a dual 741 had never been associated with low noise, and its gain.bandwidth product was a bit too limiting for best audio, so that WW article was just dismissed out of hand by all those I discussed it with. It was a bit odd that it got through their filter.

David

[bikerhifinut]

And so the wheels turn, and the circle is completed and we end up where we started, ergo the 5534 and its equivalents really are as good, if not better than they need to be given the constraints imposed by source materials and transducers and noise from resistors etc.?

One thing I get great pleasure from is the discovery that so much of my vintage equipment really does cut it in the modern world, given that its maintained and fettled to the condition it would have been in when new.

But I do enjoy very much the learning process and benefiting from the professionals on here insight and experience. And a bit of experimentation.

It would appear that Armstrong used the HA12017 in their 700 series amplifiers.
Lovely workmanship too.
http://www.audiomisc.co.uk/Armstrong/700/700page5.html




A.

[Radio Wrangler]

There was plenty to like in those last Armstrong amplifiers and tuners, but the company's name had been on the wane for some time before them and they came out just as the audio world was starting to go silly. So they never succeeded in the market to the extent they deserved.

Some vintage equipment wasn't so good, but some was genuinely excellent and trying some shows that there may have been lots of technology changes, but there has been relatively little progress in the prime purpose, of reproducing sound.

To put this in perspective ask yourself how much progress human hearing has made in the same period

The people back in the fifties and sixties weren't stupid and they knew what violins, voices and electric guitars sounded like. They could sort the good stuff from the crap.

Loudspeakers are where some useful progress has been made.

David

[John_BS]

I can't add anything to the excellent theoretical & practical posts in this thread, other than the following peripheral tip.

Back in my youth I made an RIAA preamp which was housed in a tin can (for screening) and located up under the turntable. With all that gain down at 50Hz it's very easy to end up with hum issues, and I'd decided to remote the RIAA stage close to the cartridge.

Now, the problem of supplying ultra-clean DC supplies? I used simple (decoupled) zeners on the RIAA board, and fed the unit from constant current sources back at the main amplifier. Hence no ripple DC current in the earth lead. It worked a treat.

John

[John_BS]

Also in that WW issue were announced a suite of Hitachi ICs, namely the HA1137 (CA3089 type), HA1196 (MC1310 type)

The HA1196 was an excellent decoder and capable of professional performance: much superior to the MC1310. In fact I used it in the BBC's stereo decoder design for the odd place where decoding to L&R was necessary (most re-broadcast FM installations demodulate to multiplex and then remodulate back up to Band II)
John

[Craig Sawyers]

The Armstrong last gasp of the 7000 series was a truly spectacular swan song.

They only come up for sale very occasionally. Amazingly the 521 seems to fetch £80 to £120; the world has gone truly bonkers. I have one in the attic, and it measures truly badly. RIAA off by a mile, noisy and a really bad distortion power amp section. I only keep it for nostalgia's sake.

The 600 series was a different kettle of fish. Looks great, slimline, nice feel to the controls and switches, and measures reasonably.

But the 700 series was in a different class entirely. I'd like a set, but think they are likely to fetch good money on the rare occasion one comes up for sale.

Craig

[Restoration73]

Seriously I did look into low noise for microphone pre-amps a few years back I seem to
get good results from Ferranti ZN459, and OP-27, PMI I think.
When we were building RIAA amps as a classroom project, we settled on the inexpensive
RC4558 / 4559 and went on to plot the frequency/gain a response close to ideal. These
might not meet your s/n but they were popular in audio mixers. If anything tape head
preamps are more critical as the levels can be very low.

[john.north]

I have a PW Winton amp with toroidal transformer and fet output transistors which I built in 1979. It’s pre stages were tlo72s.It has been in continuous use with sme3009 arm connoisseur bd1 deck and Nagaoka MP110 cartridge, and Telefunken 710s. I had slight hiss with the volume control at about 11 o’clock.
A few weeks ago I replaced the first stage with a surface mount OPA1642 dual jfet and stage 2 and 3 with OPA2134. I Can now turn it up to about 2 o’clock before I hear any hiss. This set up sounds just as good as a new Denon PMA600NE with Dali spektor2s.