TI NE5532 Audio Opamp Changes

[Alex Nikitin]

INA128U with the same PCN problem 2022: slew rate and settling time degradation depending on random chip site origin.
"INA128U: The die is manufactured in CSO: SHE or CSO: FRE."
"Changed typical slew rate specification from 4V/μs to 1.2V/μs in Electrical Characteristics"
 
https://www.ti.com/lit/ds/symlink/ina128.pdf

Just as a joke I've checked the current datasheet for the device I am very familiar with - INA133 . The same 5HIT version, depending on the place of production, the slew rate can be 5V/us or 20V/us, the noise can be 3 time higher, the output short circuit current can be a half of what it was and even the internal configuration is different (see attached diagram, and I'm pretty sure that they have marked the diagram wrongly*  ).

Cheers

Alex

* P.S. - I've checked on an old "proper" INA133 I have here and my guess was right, the original version is BJT, not CMOS, and has diodes between inputs. So even the information given is wrong!

P.P.S. - and small things I've missed on the first read:

Changed 0.01% settling time test condition from "CL = 1000pF" to "CL = 100pF" in the
Electrical Characteristics

   

[floobydust]

Even the RC4580, RC4560, LM358, MC33078, LM833 got the die design change. Ref. PCN 20231114002.1, 20230627002.1

I have also got burned by TI cheapening their IC packages - different epoxy and metals. One IC failing in a product and after a lot of grief, TI concluded the flux was too aggressive, never mind how it was getting into the epoxy and attacking the bonding wires.

It's some kind of "race to the bottom" that TI is doing.

[tom66]

It would seem that any TI parts that used an advanced bipolar process at Sherman fab are at risk, as well as some devices made in Lehi, UT (LFAB).  Some have been redesigned, in many cases with significantly degraded parameters.  Perhaps others have been silently respun without any noticeable degradation.  It seems a lot of this came in after Haviv Ilan took office (from Rich Templeton who had served as CEO for 19 years), an aggressive cost saving programme was introduced which saw the closure of several fabs and revamps of others; I imagine the engineers at TI have had very little time to respin these devices.  With more time, they could have likely kept parameters closer, but sub-optimal designs were chosen to save time and this is the result.

Corporate America strikes again.

[David Hess]

It would seem that any TI parts that used an advanced bipolar process at Sherman fab are at risk, as well as some devices made in Lehi, UT (LFAB).  Some have been redesigned, in many cases with significantly degraded parameters.  Perhaps others have been silently respun without any noticeable degradation.  It seems a lot of this came in after Haviv Ilan took office (from Rich Templeton who had served as CEO for 19 years), an aggressive cost saving programme was introduced which saw the closure of several fabs and revamps of others; I imagine the engineers at TI have had very little time to respin these devices.  With more time, they could have likely kept parameters closer, but sub-optimal designs were chosen to save time and this is the result.

Corporate America strikes again.

Besides the loss of optimized for specialized semiconductor processes, Texas Instruments has likely sacrificed institutional engineering knowledge in favor of less expensive younger employees.

[16bitanalogue]

Whatever gave you the idea that it is college students and hobbyists who are concerned about this issue??

Write an angry letter to your TI FAE and let them know or go onto their E2E and raise an issue. I'm saying that from TI's perspective small customers are in the same bin as college students and hobbyists. If I were the business manager who approved the decision, or perhaps it was forced upon the business manager because of a change in fab, I wouldn't lose sleep over people complaining on an internet forum. I am sure they are well aware of how many units are sold and to whom. Since no MIN/MAX actually changed in the ECT (ABS MAX doesn't count), this isn't a controversy. Annoyance and a good rant, sure.

The changes themselves are not that bad:
1. ABS MAX supply changed from +/- 22V to +/-18V. The MAX operating is still the same: +/-15V. Anyone operating beyond the maximum operating range was always asking for trouble. The datasheet specs do not hold outside of the ECT.
2. The ABS MAX input range changed from +/-VCC to +/- 15V; although, input common mode range remains the same.
3. GBW increased on a typical spec.
4. Slew rate reduction on a typical spec. This one spec honestly is the biggest kick, but it still is a typical.

So, purely as an example, you were using the old version of the N5532 with +/20V supply and you called me to rant. I would point out to you that the maximum operating voltage of +/15V has not changed which governs the ECT.  That decision to operate beyond that operating point purely rested on your engineering judgement, and you should have been told as such.

I tell every young engineer the same spiel. No one guaratees limits or behavior outside of the ECT even if the operating point is within the ABS MAX. Not simply because it is not tested there. One must consider over time shifts in performance (I have seen it), or migration to other fabs, test facilities, etc. The goal is to make the changes transparent. As mentioned above, the changes are really affecting ABS MAX and typical specs in the ECT.

I am sure some customers will be affected, and in the immortal words of the remaining greybeards at TI/Burr-brown, you'll lose some customers while winning others. It's an op-amp, not an ASIL-D buck-boost converter for ADAS.

[David Aurora]

Whatever gave you the idea that it is college students and hobbyists who are concerned about this issue??

Write an angry letter to your TI FAE and let them know or go onto their E2E and raise an issue. I'm saying that from TI's perspective small customers are in the same bin as college students and hobbyists. If I were the business manager who approved the decision, or perhaps it was forced upon the business manager because of a change in fab, I wouldn't lose sleep over people complaining on an internet forum. I am sure they are well aware of how many units are sold and to whom. Since no MIN/MAX actually changed in the ECT (ABS MAX doesn't count), this isn't a controversy. Annoyance and a good rant, sure.

The changes themselves are not that bad:
1. ABS MAX supply changed from +/- 22V to +/-18V. The MAX operating is still the same: +/-15V. Anyone operating beyond the maximum operating range was always asking for trouble. The datasheet specs do not hold outside of the ECT.
2. The ABS MAX input range changed from +/-VCC to +/- 15V; although, input common mode range remains the same.
3. GBW increased on a typical spec.
4. Slew rate reduction on a typical spec. This one spec honestly is the biggest kick, but it still is a typical.

So, purely as an example, you were using the old version of the N5532 with +/20V supply and you called me to rant. I would point out to you that the maximum operating voltage of +/15V has not changed which governs the ECT.  That decision to operate beyond that operating point purely rested on your engineering judgement, and you should have been told as such.

I tell every young engineer the same spiel. No one guaratees limits or behavior outside of the ECT even if the operating point is within the ABS MAX. Not simply because it is not tested there. One must consider over time shifts in performance (I have seen it), or migration to other fabs, test facilities, etc. The goal is to make the changes transparent. As mentioned above, the changes are really affecting ABS MAX and typical specs in the ECT.

I am sure some customers will be affected, and in the immortal words of the remaining greybeards at TI/Burr-brown, you'll lose some customers while winning others. It's an op-amp, not an ASIL-D buck-boost converter for ADAS.

I know you're just using this as a single example, but I think it sums up a bit of the flawed logic going on here. You (as in TI) can't assume that everyone using these at voltages higher than +/-15V is doing that in a new design. 99% of my work is repair work. And I'd say the amount of circuits I see using 5532s at or below +/-15V is maybe 30%.

We're talking about an IC that's got a nearly 50 year history being used in pro audio equipment (quite reliably) at higher supply voltages (generally 16-18V). That's a 20-30% safety margin from the absolute max and has never been an issue. And, a bigger point- the +/-15V recommendation is NOT the original recommendation, that's from more recent revisions. When a lot of equipment I regularly service was designed and built, running them on 17V rails was actually conservative based on the data sheets of the time.

And this gaslighting is the whole issue. If they just discontinued the IC it'd be annoying but you couldn't be angry at them. But to re-write the rules on how it can be used throws a spanner in the works on maintaining 50 odd years worth of equipment that WAS designed within data sheet specs of the time.

[Someone]

And, a bigger point- the +/-15V recommendation is NOT the original recommendation, that's from more recent revisions.

How far back do you have access to?
2002: +/-15 recommended
1990: +/-15 recommended

On the other hand they tacitly support +/-18V  operation by having some specifications at that operating point.

[0xFFF0]

Modern chip structures are simply thinner and more fragile. Consequently, you cannot apply high voltages or currents. Still, it’s a disgrace that TI didn't assign a new part number.

[floobydust]

Modern chip structures are simply thinner and more fragile. Consequently, you cannot apply high voltages or currents. Still, it’s a disgrace that TI didn't assign a new part number.

I don't agree, there are many op-amps in the lineup that are HV and high current. But they are expensive or command a high markup on the price.


What op-amp can you use when functional safety is a concern? Have to ask because the old jellybean parts have a lot of failure information relevant to the FMEA you did... and have to redo with no FIT info on the "new chip" 

[David Aurora]

And, a bigger point- the +/-15V recommendation is NOT the original recommendation, that's from more recent revisions.

How far back do you have access to?
2002: +/-15 recommended
1990: +/-15 recommended

On the other hand they tacitly support +/-18V  operation by having some specifications at that operating point.

Sorry, I definitely mis-spoke there. Lots of the old data sheets do use +/-15V as typical or recommended operating conditions, the point I was trying to make was that they explicitly advertised it as capable of running at higher voltages. A handful of data sheets I pulled up in 2 seconds:

https://www.qsl.net/w9sz/files/PDFs/NE5532.pdf 2001
https://www.farnell.com/datasheets/81865.pdf 2006
https://www.onsemi.com/pdf/datasheet/ne5532-d.pdf 2013
https://www.ampslab.com/PDF/ne5532p.pdf 2000
https://www.elektronikjk.pl/elementy_czynne/IC/NE5532-6.pdf 1988

When you're going through op amp data sheets and almost all of them list +/-18V as absolute maximum with +/-15V as typical (3V margin), and then you have 5532 data sheets saying +/-22V absolute maximum and +/-15V typical (7V margin), I (and plenty of manufacturers) think it's beyond reasonable to assume that running them at 16-18V should be absolutely fine. Hence so many products out there that work this way, that can no longer use TI IC's as replacements

Hell, the 2000 TI data sheet literally says "Peak-to-Peak Output Voltage Swing
32 VTyp With VCC± = ±18 V and RL = 600 Ω"

[Dr. Frank]

Until August last year, I have been working for a big Automotive Electronics Supplier company as the Component Expert and PCN/PTN manager for my business unit.
Together with SQM, Purchasing, silicon experts and several experienced design engineers, we defined a PCN  matrix.
This defined, which PCN cases could apply, and what to do in which cases. Basically, a PCN could only define changes on the production methods of the device, but it was strictly forbidden to make any basic design changes on the silicon / device itself, i.e. no change in Form, Fit, Function. This could and would always lead to problems in the many applications, i.e. failures of one or more applications. Even an improvement of parameters was not allowed, or under very strict rules only. Allowed were slight technology updates or compatible  movement of fabs even with slight changes of electrical parameters. Even die shrinks were allowed, when electrical parameters wouldn't change severely.

Such A level-changes, or major changes required a full re qualification of the changed device in EVERY project which used this device. Qualification cost for each project started at about 50k€ for E.M.C. testing and could easily approach 1M€ / project, when performed outside the regular automotive re-qualification runs.

Such complete redesigns with obviously completely new silicon technology were strictly rejected, as these were no regular PCNs any more, and would become a PTN afterwards.

Here, T.I. openly claims a design change, but that was not evident in each case, so you had to be very skeptical on every PCN,  especially with T.I.
After they acquired several other analog companies, their company policy changed as well, and they were not a reliable and quality oriented supplier any more.

Frank, retired Component Manager / Expert

[Someone]

https://www.elektronikjk.pl/elementy_czynne/IC/NE5532-6.pdf 1988

That is the Signetics data sheet which is (on page one) word for word identical with the current ON Semiconductor data sheet.

Large Supply Voltage Range: ±3.0 to ±20 V

Unchanged to this day.

Where as the TI 1990 data sheet there was already a derating of the supply voltage range, when that was first dropped? needs more data sheet archives.

[David Aurora]

https://www.elektronikjk.pl/elementy_czynne/IC/NE5532-6.pdf 1988

That is the Signetics data sheet which is (on page one) word for word identical with the current ON Semiconductor data sheet.

Large Supply Voltage Range: ±3.0 to ±20 V

Unchanged to this day.

Where as the TI 1990 data sheet there was already a derating of the supply voltage range, when that was first dropped? needs more data sheet archives.

Wanna link that 1990 data sheet? Because that 2000 TI data sheet I posted above explicitly advertises use at +/-18V, and has the +/-22V max rating in it. I'm not really sure what you're trying to say here

[tooki]

Collage students are the lesser issue - maybe you confuse them with the Ti NE5532 being so much different from most others. They tend to get a new datasheet for a new design.
The big issue is with old designs used by commercial customers. There is a good chance that some audio designs go higher and maybe run the NE5532 at 18 or 20 V to get extra headroom. This would be OK for the old part, but bad for the new one.
There are also a few users that use the NE5532 for other things than audio and they may be effected by the other changes:
lower slew rate, missing input clamping diodes, other sign of bias current. This would likely not be large volume users of the 5532, but may still be valuable customers for other parts.

There is absolutely no excuse to not have markings on the part to tell the old and new version apart. Just the paper lable on the box is not enough. As just shown not even Digikey can't tell for sure if the parts they have are the old or new chips.
Those parts are essentially useless for professional users - right on par with parts from the gray market.
This stunt may help with a few sales of NE5532 (until they get the reputation for frequent failure), but it will hurt the brands reputation.
They will soon have a good reason the EOL the NE5532 and maybe have to sell the new part and a different name.

I expect commercial customers will simply say "to hell with this" and order 5532's from OnSemi.

[Messtechniker]

5532's from OnSemi: the way to go for now.

[pope]

5532's from OnSemi: the way to go for now.

No through hole option though.

[Someone]

https://www.elektronikjk.pl/elementy_czynne/IC/NE5532-6.pdf 1988

That is the Signetics data sheet which is (on page one) word for word identical with the current ON Semiconductor data sheet.

Large Supply Voltage Range: ±3.0 to ±20 V

Unchanged to this day.

Where as the TI 1990 data sheet there was already a derating of the supply voltage range, when that was first dropped? needs more data sheet archives.

Wanna link that 1990 data sheet? Because that 2000 TI data sheet I posted above explicitly advertises use at +/-18V, and has the +/-22V max rating in it. I'm not really sure what you're trying to say here

You're all over the place with numbers, brands, specs.

Signetics were the originator of the 5532 part/spec.

TI came along later with a different spec.

And, a bigger point- the +/-15V recommendation is NOT the original recommendation, that's from more recent revisions.

How far back do you have access to?
2002: +/-15 recommended
1990: +/-15 recommended

On the other hand they tacitly support +/-18V  operation by having some specifications at that operating point.

You say +/-15V is not the original recommended operating conditions, but where is the evidence? TI have the same +/-15V recommended operating conditions working back as far as I can see.

Wanna link that 1990 data sheet? Because that 2000 TI data sheet I posted above explicitly advertises use at +/-18V, and has the +/-22V max rating in it. I'm not really sure what you're trying to say here

Comes up on google for "Texas Instruments 5532 1990", the document is SLOS075A
http://www.richard-dj1pi.de/Techn_Data/NE5532.pdf
That has the +/-3V to +/-20V marketing puff on page 1 but still recommended +/-15V.

[Yansi]

You can call it puff as you wish, but professional audio equipment uses +-17, sometimes even +-18V rails pretty commonly, because higher headroom and SNR is highly desired. Even relatively cheap entry level pro equipment uses more, than +-15V rails. And no, those engineers were not stupid, those are designs proven by decades of operation.

[David Aurora]

https://www.elektronikjk.pl/elementy_czynne/IC/NE5532-6.pdf 1988

That is the Signetics data sheet which is (on page one) word for word identical with the current ON Semiconductor data sheet.

Large Supply Voltage Range: ±3.0 to ±20 V

Unchanged to this day.

Where as the TI 1990 data sheet there was already a derating of the supply voltage range, when that was first dropped? needs more data sheet archives.

Wanna link that 1990 data sheet? Because that 2000 TI data sheet I posted above explicitly advertises use at +/-18V, and has the +/-22V max rating in it. I'm not really sure what you're trying to say here

You're all over the place with numbers, brands, specs.

Signetics were the originator of the 5532 part/spec.

TI came along later with a different spec.

And, a bigger point- the +/-15V recommendation is NOT the original recommendation, that's from more recent revisions.

How far back do you have access to?
2002: +/-15 recommended
1990: +/-15 recommended

On the other hand they tacitly support +/-18V  operation by having some specifications at that operating point.

You say +/-15V is not the original recommended operating conditions, but where is the evidence? TI have the same +/-15V recommended operating conditions working back as far as I can see.

Wanna link that 1990 data sheet? Because that 2000 TI data sheet I posted above explicitly advertises use at +/-18V, and has the +/-22V max rating in it. I'm not really sure what you're trying to say here

Comes up on google for "Texas Instruments 5532 1990", the document is SLOS075A
http://www.richard-dj1pi.de/Techn_Data/NE5532.pdf
That has the +/-3V to +/-20V marketing puff on page 1 but still recommended +/-15V.

The data sheet you linked to and your own post says clear as day that the IC can run on 20V rails. Yes, +/-15V is typical/recommended in it (I corrected myself about that a few posts ago and clarified my point), but that didn't mean you couldn't use 18 or 20 or 5 or 9 volt rails as they were all within the safe operating range.

What is the actual argument you're trying to make here? That a 5532 should never be used at any voltage except +/-15V? Because that's moronic.

TI's different spec came DECADES after the original IC, and long after their own version. That's the point of this thread.

[pope]

Everyone who's been into audio design knows that it is was absolutely fine to run the 5532s on +/-18V.

[HwAoRrDk]

Did they decap RC4580? Would now be very interesting to compare the dies. 

Zeptobars has not (at least that I could find), but someone did in one of Noopy's threads here:

https://www.eevblog.com/forum/projects/opamps-die-pictures/msg6158555/#msg6158555

[tom66]

Comes up on google for "Texas Instruments 5532 1990", the document is SLOS075A
http://www.richard-dj1pi.de/Techn_Data/NE5532.pdf
That has the +/-3V to +/-20V marketing puff on page 1 but still recommended +/-15V.

The front page literally states:

Peak-to-Peak Output Voltage Swing
32 V Typ With VCC± = ±18 V and RL = 600 Ω

In other words, it's fine to use ±18V supply and even earns a "typical" parameter spec.

The new part completely deletes this specification; it doesn't even specify what the swing would be with 2kohm load or at ±15V supply.

TI took a part that had a defined top-line specification of performance, changed the part materially and deleted the specification and all references to it because they knew it was not comparable and would be noticeably worse than the competition.

[Alex Nikitin]

The main problem IMHO is that: there is now what looks like a complete chaos with many industry standard parts from TI, not just the NE5532. I might take some time before this situation will blow up properly, as the counterfeit parts will take a while to perpetrate supply chains and affect the industry, however the effect can not be underestimated. As an example scenario I can easily imagine, the company I used to work for some years, a pretty large US corporation producing critical instrumentation for many applications, including semiconductor manufacturing (the marketing line was and I suppose still is that all semiconductor devices currently produced are made with the help of my former employer's products). I know that current products I've worked on use many TL072 and INA133 devices from TI to name but two. The counterfeit devices will not work correctly in many positions. One can not expect to have 3 times the noise when it is a critical parameter in the circuit. Or a quarter of a slew rate. And there is no easy way even to check which type of the device is used when soldered on a pcb  .

Cheers

Alex

[iMo]

The problem with the 5532 at TI could become an opportunity for other smaller vendors, imho.. There is around 100 foundries around the world, afaik, most of them with older nodes. So - would it be a large problem for a smaller talented vendor to produce the 5532 within all that original specs? With even 3-5x the price for such a unique opamp people will buy it, sure (as all the discussions show)..

[Alex Nikitin]

Why am I having a feeling that some new additions to the "NE5532" fake version data sheet were written by AI?!

6.3 Feature Description
6.3.1 Unity-Gain Bandwidth
The unity-gain bandwidth is the frequency up to which an amplifier with a unity gain can be operated without
greatly distorting the signal. The NE5532, NE5532A, SA5532, and SA5532A devices have a 10MHz unity-gain
bandwidth.

Cheers

Alex