TI degrading specs of existing parts without changing part numbers

This looks horrendous. How to destroy your reputation in one easy lesson. John

Supposedly they enshittified the TL072 as well.

They changed the process on a part that's 47 years old.

Scandalous.

John Larkin Highland Tech Glen Canyon Design Center Lunatic Fringe Electronics

When you significantly degrade a couple of parameters that are important for many users of that part and don't provide a means of knowing whether you have the original part or the degraded imitation, then yes, it is scandalous. Especially when the ECN tells buyers that the changes will have no impact. John

It is impressive that TI keeps parts in production for decades, almost

This part belongs in the 555 timer and uA741 and 2N3055 generation, getting creaky before most of my enginers were born. Or before their parents were born.

John Larkin Highland Tech Glen Canyon Design Center Lunatic Fringe Electronics

Well you could use the tlv9362.

Cheers

I hate U-tube and didn't watch it, but I suspect Texas Instruments is still doing what it did back in 1974 - changing the numbers in the specification of an industry standard part.

At George Kent back in 1974 we had our own part numbers, and the engineers had the job of setting up a specification sheet and filling in the numbers and identifying which manufacturers parts could be bought against that part number.

I repeatedly found that the Texas Instruments "industry standard" parts weren't. Their data sheet for that part number would specify looser limits and poorer typical values than the rest of the industry.

The U-tube link starts off on this this part

I couldn't see any obvious differences between the TI and the On-Semi data sheets, but I didn't look all that hard. I don't think that Texas Instruments is destroying their reputation - they are just confirming it.

The key differences are that the same TI part number has reduced its absolute max supply voltage from 22V to 18V and substantially reduced the output drive capability. They have removed the spec for driving into 600 ohm loads. The main use for these parts is in audio system where driving 600 ohm loads is often a requirement. I don't have a problem with them stopping the manufacture of an ancient part and recommending a replacement. They should not take a completely different design and try to pass it off as the original and keep the exact same part number they were using previously. The design really is completely different. The comparison is between different editions of the same TI data sheet. One advantage of using automotive parts is that such changes are very unlikely. Car manufacturers would not tolerate it. I was once sitting in on a meeting when a car manufacturer told a well known semiconductor manufacturer to spend another USD500k on requalifying a part that had been changed. The semi manufacturer immediately agreed to do it.

Thanks for this warning. (S.

All great parts - technically obsolete but will be available forever from thousands of sources - if your application can use them then why use anything else?

piglet

That's really bad. Purchasing departments aren't expected to read data sheets, and they really shouldn't have to cope with data sheet changes from one order to the next.

Shrinking a part so that it is more likely to blow up if briefly over-loaded does create problems, but if the transient handling capacity isn't specified on the data sheet you don't have any right to sue.

Changing the data sheet is criminal.

People like us, and small scale manufacturers in general, exploit the mass market create by people who buy in bulk.

If the people who buy in bulk change their ordering habits, we can be left high and dry.

The disappearance of wide-band (5GHz or better) PNP transistors is a recent example.

Perhaps they were telling the truth and the others were stretching it a bit?

A firm I worked for often quoted specs that were so far below the actual performance that one of our sets almost met spec. with one of the transistors soldered in backwards. This understatement meant that their sets had a reputation for always meeting the specification - even after the customer had tried to 'improve' them.

It was only when the bean-counters began over-ruling the engineers that sets with better specs on paper but worse or unreliable performance began to out-sell them.

Seems unlikely.

Specs tend to reflect what you can test quickly rather than actual performance.

MOSFETs typically had picoamp leakage through the gate, but waiting for long enough for the leakage current to build up to a respectable voltage wasn't an option for production line testing.

The bean-counters mostly don't know what they are doing. This is - in practice - where the problems usually come from.

Educating managers is definitely part of an engineers job, and managers are remarkably resistant to the idea that they need education.

I'm not saying there isn't an alternative, I am saying that if they have to reduce the maximum supply voltage spec, or similar important ones, they should just say the old part is obsolete, and provide a new part number for whatever they can make.

NE5532

. . . . just in case somebody is actually searching for info on an actual part number.

If you're wired into the ON/TI/NXP PCN channels, you'll pick this up in time to remove the TI part number from your list of alternates.

Audio mfrs are pretty paranoid about sources. Chances are TI never showed up, in the first place.

RL

If I hadn’t stopped talking audio design seriously in about 1980, I certainly would have when crappy compressed MP3s took over the world.

Golden ears, my left buttock.

Cheers

Phil Hobbs

Your buttocks have ears?

I was very impressed with mp3 compression first time I heard it So many audio files on a small memory.

Same goes for MPEG compression for video.

Have you ever heard a Quad electrostat?

Have some old Sennheiser HD210 headphones too, good when practicing playing the keyboard.

...but it doesn't sound so good by the second or third generation.

The difference between the original AIFF file and a first generation mp3 at a moderate compression rate is audible when both are electrically recorded to a wax cylinder. Sometimes, when a 78 is transferred to mp3, the results can be excruciating because the algorithm favours the scratch over the music.