"Classic" Chips

[DrG]

General Instrument AY-3-8910...

These were carried in Radio Shack when I was a kid and made me lose my mind a little bit. ....

That GI chip was the most sophisticated sound chip that I could get my hands on..better than the available TI chips. The GI chip literally proved to me that I did not understand music no matter how much I loved to listen ....didn't even bother trying with the SID chip.

I think that the 8912 was what was carried at  RS and it is listed here as the most widely used variant https://en.wikipedia.org/wiki/General_Instrument_AY-3-8910#Variants

Later, I tried to convince Yamaha to send me samples of their 'new' sound chips, but they would have none of that - despite my "creative explanations", they, apparently, concluded that a guy from a medical research institute did not need samples of our sound chips.

Here is my All-Shack project from 1987/8 with on-board amplifier (LM386) and ready to run from a PC parallel port - I was quite happy with it at the time. I even dug up a demo program that I wrote in Borland's Turbo C. BTW, 32-33 years later, I still can't cut a circle for a plastic case

Code: [Select]

void c_chord()
{
  reset();
  load(7,248); /* enable tones on A,B,C      */
  load(12,80); /* envelop coarse [about 0.4 Hz] */
  load(8,16); /* amplitude of a on envelope control */
  load(1,1); /* C-4 MSB */
  load(0,172); /* C-4 LSB on a */
  load(13,0); /* envelope shape/cycle[long delay] */
  timer(250); /* wait */
  load(3,1); /* E-4 MSB */
  load(2,83); /* E-4 LSB on b */
  load(9,16); /* amplitude of b on envelope control */
  load(13,0); /* envelope shape/cycle[long delay] */
  timer(250); /* wait */
  load(5,1); /* G-4 MSB */
  load(4,29); /* G-4 LSB on c */
  load(10,16); /* amplitude of c on envelope control */
  load(13,0); /* trigger      */
  timer(1700); /* wait */
  load(13,0); /* trigger all three */
  timer(1850); /* wait */
  reset();
}
Good times, thanks

[free_electron]

TBA120
TDA2020 TDA203 and family
8051 family and especially 8052-ah-basic. single chip microcontroller with 8k floating point basic on board. and not one of those namby pamby second sources. real intel with mask rom !

[schmitt trigger]

The difference between a 6L6 or any other tube vs the K2-W is that the K2-W is not a component but a functional block.
The circuit comprised of active devices plus several passives (including neon lamp level shifter) are integrated into a package. This package becomes itself a component on a much larger circuit.

That is why I mentioned “integrated circuit” in the broader sense, although it is indeed vacuum state electronics.

[magic]

The Signetics 5534 and 5532, which may be the first successful op amps specifically designed for audio.

... and by Philips, apparently

The LM10 lives on as the LT1635.

Interesting, I didn't know there were clones. That being said, the original is still in production, at least the comercial grade seems to be.

[TimFox]

The 5532 and 5534 (and the 555) were developed by Signetics, who were acquired by Philips (now NXP) in 1975.

[TimFox]

The difference between a 6L6 or any other tube vs the K2-W is that the K2-W is not a component but a functional block.
The circuit comprised of active devices plus several passives (including neon lamp level shifter) are integrated into a package. This package becomes itself a component on a much larger circuit.

That is why I mentioned “integrated circuit” in the broader sense, although it is indeed vacuum state electronics.

Yes, the K2-W is an integrated circuit, but not a monolithic one.

[magic]

The 5532 and 5534 (and the 555) were developed by Signetics, who were acquired by Philips (now NXP) in 1975.

I assumed it was designed at Philips because it first appeared in Philips databooks and without cross reference to Signetics.
But the acquisition definitely muddies the water.

The specific designer(s) remain completely unknown to this day, as far as I know.

edit
Studio Sound magazine, August 1990, p. 55 confirms that it first appear under Philips brand.

Part One of this series charted the initial development of op-amps and their fledgling application in pro-audio equipment. In 1977, Philips began making an op-amp called TDA1034. Probably intended for telephony / telecom equipment (Philips have a comic way of seeming to stumble on winning designs by accident), it wasn't long before 1034s were being shipped out of Eindhoven to London, to upmarket console manufacturers. Packed in extruded aluminium tubes, the initial price was high (£5.00/$8.00) but then the 1034 combined a host of features that were desirable for audio and not available in any other op-amps. For the first time, you could build a compact, very simple and respectably quiet mic amp without resorting to discrete transistors. Harmonic distortion (THD) was commendably low, well below 0.01% and the 1034 (`ten thirty four' to its friends) could drive a 600 S2 load without driving a distortion meter off-scale. Moreover, you could use a ±22 V supply to achieve a high headroom of above +24 dBu.
In 1978, production was switched to an ailing stateside IC manufacturer that Philips had recently bought out, called Signetics. The Mullard/Philips TDA1034 was then relabelled as the Signetics NE5534. The 5534's price fell swiftly between 1980 and '82, after Raytheon and later Texas and other makers began to second source their own ` klones'. By now, the 5534 was poised to become the de facto standard for pro- (and much later, domestic) audio, a position it still holds 13 years later. It's the only IC op-amp designed in Europe that's met with any continued success. The introduction by 1980 of a dual version (NE5532) in an 8-pin package did a lot for the 5534's future popularity and utility. Today, 5534 and 5532 klones are widely second-sourced and cost below 50p (804) in manufacturing quantities.

[m k]

Fonz-11

[TimFox]

I was trying to find a definitive history of the 5534, but missed that article.  I had first encountered the device under the Signetics name.  In my clouded memory, I thought that Jung had endorsed the new device earlier than 1977, but I stand corrected. 
Harris also had some interesting high slew rate op amps ca. 1980 that were used in high-performance audio test equipment, but didn't seem to attract much audio attention. 
See the 2500 series in the 1977 catalog  http://bitsavers.trailing-edge.com/components/harris/1977_Harris_Linear_and_Data_Acquisition_Products_Volume_1.pdf
Jung had a rule of thumb that the slew rate (in volts per microsecond) should be more than one-half (or better, one times) the peak output voltage of the device for audio use.  My favorite theoretical explanation for that factor pointed out that with Miller capacitor feedback on the second stage for stability, as the slew rate limit is approached the input differential stage goes into large-signal operation, inducing distortion.  The 5534 combined good slew rate and high output current (suitable for 600 ohm load) with low input noise voltage (but relatively high input bias current).

[David Hess]

Harris had a fast complementary bipolar process in the 1970s which gave them more options for fast designs.

Signetics had their 531 and 535 high slew rate operational amplifiers in 1976 if not earlier but to me their design looks like how I would add emitter degeneration to a 741/301A for higher slew rate which is consistent with their low gain-bandwidth product.  They are definitely not precursors to the 5534.  The 318, not Signetics, was an earlier fast operational amplifier with high slew rate but used emitter degeneration for transconductance reduction contributing to its high input noise so not suitable for audio; it might be considered the first "video" operational amplifier.

[Zero999]

Does anyone know why a quad NE5532 was never produced? It seems logical, since it would safe space and possibly power too.

[David Hess]

Does anyone know why a quad NE5532 was never produced? It seems logical, since it would safe space and possibly power too.

I would hazard a guess that the large die would not have fit into the width of the package.  94 picofarads of MOS capacitors per amplifier does not a small chip make.  Quad amplifiers from that era like the LM324 had more like 5 picofarads of MOS capacitors per amplifier explicity to minimize size and thereby cost, which is why a true quad 741 was either never produced or very expensive.

[magic]

I was trying to find a definitive history of the 5534, but missed that article.  I had first encountered the device under the Signetics name.  In my clouded memory, I thought that Jung had endorsed the new device earlier than 1977, but I stand corrected. 

There appears to be a consensus among EEs that the part originated from Signetics and a similar consensus among audio engineers that it's from Philips, see DIYAudio

The earliest datasheet I have is dated June '76 so the chip was clearly announced before '77. Somebody on DIYA posted that NDA'd engineering samples were available (from Philips) a year or two prior; dunno if it's credible.

Does anyone know why a quad NE5532 was never produced? It seems logical, since it would safe space and possibly power too.

Almost 2×3mm die size in old Signetics duals. Not only capacitors (maybe ⅓ of the die) but also a fair bit of circuitry.
Power dissipation: worst case quiescent power is 6mA·36V (that's if we give up 44V operation) - already 850mW for a quad just doing nothing. Add minimum 450mW reserve for >25mA short circuit to ground on one channel. Not a great thing for DIP14.

[nfmax]

Just found 7-off TDA1034 date coded M8233B in the junk box. I think they were being branded as Mullard (owned by Philips) at the time - they certainly don't have the Philips badge.

[magic]

pics?

[VEGETA]

L200
TL431
LM324
LM358
2N2222
etc...

[srb1954]

I was trying to find a definitive history of the 5534, but missed that article.  I had first encountered the device under the Signetics name.  In my clouded memory, I thought that Jung had endorsed the new device earlier than 1977, but I stand corrected. 

There appears to be a consensus among EEs that the part originated from Signetics and a similar consensus among audio engineers that it's from Philips, see DIYAudio

The earliest datasheet I have is dated June '76 so the chip was clearly announced before '77. Somebody on DIYA posted that NDA'd engineering samples were available (from Philips) a year or two prior; dunno if it's credible.

To add to the confusion I have a Philips data book labelled Signetics Integrated Circuits 1976. This must have been a reprint of a Signetics data book immediately after Philips took over as everything inside the covers is labelled Signetics and none of the data sheets are in the Philips format. The only part of this data book that is in the Philips format is the outer cover.

This data book appears to cover the complete range of Signetics chips including logic, linear, interface and microprocessor but there is no mention of the NE5534 in this data book.

From this I would conclude that the NE5534 was perhaps introduced after Philips took over but maybe before complete assimilation took place and during which time the Signetics name was still being regularly used.

[floobydust]

I thought Rupert Neve has approached Philips to develop the NE5534? I have Philips, Signetics, Raytheon ones in my bins.
The modern TI NE5534 part has little to do with the old original part- different die and process, considered "equivalent". Only spec'd to +/-15V rails, not 20-22V.

Does anyone know why a quad NE5532 was never produced? It seems logical, since it would safe space and possibly power too.

A quad package part is not feasible due to the high power dissipation. The op-amp (5534) was run at +/-19V in Neve consoles and high quiescent current with the Class AB bias. Say 4mA/38V 150mW/section = 0.6W for a DIP-14, let alone more heat driving 600 ohm loads.

[SilverSolder]

LM10. Decent op amp (works down to 1.1V supply) and a band gap reference.  Thanks to Bob Widlar  What's not to like.   Also the old 723.

Hmm the LM10 is an unusual item, is there a modern equivalent?

[Edit:  LT1635.  Note to self:  read entire thread before posting]

[David Hess]

High power 14 pin and larger DIPs were available then which used a larger lead-frame however the cost with the more expensive package would have been greater than a pair of normal 8 pin DIPs containing duals so there would be no reason to make a quad NE5534 when they already had the dual NE5532.  Dual parts are often preferred over quads anyway because they simplify layout.

For instance the Fairchild uA706 was available in the otherwise normal seeming 14 pin DIPP, dual inline power package, which was good to 1.7 watts.  A 14 pin ceramic DIP package could also have been used like many 723 regulators did for higher power.

[tchicago]

MC34063 buck/boost.

Don't forget it can also convert positive to negative. Truly universal chip, although not as efficient as modern MHz-range bucks.

[BrianHG]

NE592 differential in and out video amplifier with 1 resistor set gain from 0 to 400.  90Mhz bandwidth at a gain of 100, 12ns rise and fall time, 10ns delay.

UA733 is almost the same, except the gain begins at 10 instead of 0 & it reaches 200MHz.

[tchicago]

TL714 high speed comparator. Classic and still manufactured.

[tchicago]

Well... I did not grow up with 6502 and in our postUSSR country, 6502 was never a thing.   

Here almost all was 8080 clones or Z80 clones.

uA702 aka 140UD1. First single-chip opamp, courtesy of Bob Wildar. The western civilization has quickly evolved from this chip into using much better ones, but the USSR got stuck with it for decades. 140UD1 (clone of uA702) was everywhere...

[magic]

I thought Rupert Neve has approached Philips to develop the NE5534?

That seems to be an oft-repeated story, and I now found another post about it.

Well, I first heard it from Graham Langley at a trade show about 10 years ago... Graham abbreviated it (by allusion) to Rupert essentially having designed the 5534. -It seemed outrageous to me, but about 4 or 5 years ago, Rupert came for a visit and I found myself seated at the dinner table with him when that tale came back to me, so I asked him if what Graham had said was true.

-He answered that it was actually 'essentially true, I suppose.' and then enquired as to where I happened to come by that nugget, so I told him that I'd heard it from Graham. (We had the first M3000 from Amek back in 1979, and worked fairly closely with Graham and Greg Hogan at the Amek Salford location, back LONG before Rupert -much later- began his design association with AMEK)

He expanded for illustration, explaining that -at the time- the Neve company had been bought out, (Bonochord/Bon-accord?) and he was an employee of the company, so his objections carried little weight. However, the new comapny owners -I forget exactly who was the driving force- wanted to produce an upgraded alternative to the 741 op-amp, which was finding its way into audio equipment at the time. I think that the design was to be the property of Philips, (Were Philips assiciated with Mullard?) but I may be off on that one... none the less, the licensed variants would bring revenue to the company who commissioned the design (whoever that was), and not to Neve.

Rupert suggested that this almost amounted to giving other people the tools to put Neve out of business, but he was overruled, and the design went ahead. For his sins, he was instructed to oversee large parts of the project.

I may be off a little on the details, but the essence of the tale I heard from Rupert's own sober self, and not via any third party.

https://repforums.prosoundweb.com/index.php/topic,15555.msg203117.html#msg203117

I have Philips, Signetics, Raytheon ones in my bins.
The modern TI NE5534 part has little to do with the old original part- different die and process, considered "equivalent". Only spec'd to +/-15V rails, not 20-22V.

TI has been making these chips for an eternity too and the dice look about the same. If you don't like their version, there is still ON Semi, recommend voltage up to 20V, but only in SOIC. JRC made DIPs but they killed it last year.

MC34063 buck/boost.

Don't forget it can also convert positive to negative. Truly universal chip, although not as efficient as modern MHz-range bucks.

The output can only swing within 2V of VCC, so the buck mode is extremely inefficient. And there is no thermal shutdown.