We're Getting Old. Infineon just killed all classics ...

[CatalinaWOW]

Time moves on.  The big guys don't make 01A, 76, 5U4GB etc. anymore.  And nobody would make things like 6L6 if it weren't for a price is no problem audio market.

[jitter]

Good question. I wasn't aware of it, until now.

One of the parts going obsolete is a 74HCT221N. I think this would make one of our customers complain (unless the company I work for informed them on time and we did a last time buy). I don't seem to easily find a drop in replacement...

That one is still made by TI (CD74HCT221E).

Thanx.
Is TI the last of the big boys interested in selling DIP 74 series logic? If so, I wonder what the prices are going to do...

[batteksystem]

BC547 BC557 BC517 1n4148 2n2222 BAV99 and tons more . All the stuff i grew up with ... all gone. in one shot ...

http://media.digikey.com/pdf/PCNs/Infineon/PD_012_16.pdf

next i'll read somewhere the 741 and 555 are dead too ...

Usually when they kill off the old, that often mean they also offer some kind of upgrade to the market at the same time. Unless it is really an odd chip.

[Cervisia]

Is TI the last of the big boys interested in selling DIP 74 series logic?

"TI maintains a firm commitment to remain in the market with both leading-edge and mature logic lines."

"TI Logic’s stable no end-of-life policy secures that all products in our portfolio will always be available."

If so, I wonder what the prices are going to do...

As always, the prices are set by the Supply & Demand Fairy. Don't annoy her by using unpopular parts …

[tszaboo]

So they dont want so sell cheap transistors, and diodes with generic name. They want to sell vendor specific packages, and part numbers for higher margin. Vendor lock in is brilliant. Synergy!

BTW some of these diodes are relatively important for high accuracy analog front ends. I dont know what were they thinking. Please tell me when there is a replacement for the small leakage BAV199

[Miyuki]

BTW some of these diodes are relatively important for high accuracy analog front ends. I dont know what were they thinking. Please tell me when there is a replacement for the small leakage BAV199

There is still lots of small manufacturers who make these parts and they don't disappear ever. Its like with vinyls as big manufacturers close, small one and enthusiasts start their business. Today these simple semiconductors can be simply made by group of enthusiasts with old equipment.

[edavid]

BTW some of these diodes are relatively important for high accuracy analog front ends. I dont know what were they thinking. Please tell me when there is a replacement for the small leakage BAV199

BAV199 is still in production from multiple sources 

http://www.findchips.com/search/bav199

[David Hess]

BTW some of these diodes are relatively important for high accuracy analog front ends. I dont know what were they thinking. Please tell me when there is a replacement for the small leakage BAV199

The replacement is called the collector-base junction of a non-gold doped small signal transistor.  If you can live with its low breakdown voltage, then the collector-emitter junction is also low leakage and very fast as well; not coincidentally, fast low leakage diodes, which I do not think anybody makes anymore, have reverse breakdown voltages of about 10 volts.

Notice that the room temperature leakage specification at 75V for the BAV199 is 3pA typical and 5nA maximum.  They are not even testing them for low leakage and for a good reason; leakage tests into the picoamp range take significant time which costs money.  Non-gold doped small signal transistors offer similar performance and if you want reliable picoamp leakage, you will have to grade devices yourself even with the BAV199.

[T3sl4co1l]

The replacement is called the collector-base junction of a non-gold doped small signal transistor.  If you can live with its low breakdown voltage, then the collector-emitter junction is also low leakage and very fast as well; not coincidentally, fast low leakage diodes, which I do not think anybody makes anymore, have reverse breakdown voltages of about 10 volts.

E-B junction*..?

E-B is faster, but lower breakdown voltage and higher leakage than C-B.

Recovery, C-B:


E-B:


E-B leakage is higher.  I was playing with a semi-precision analog voltage clamp circuit, and couldn't get anything reasonable from the E-B junction.  (I ended up going with inverted transistor operation, which gives precise Vec(sat), at the expense of needing more bias on the "reference" side of the circuit.)  This follows from the junction's at least partial zener character: leakage is higher, and covers a wider voltage range -- becoming significant around 4-5V, with proper "breakdown" around 7V.

Notice that the room temperature leakage specification at 75V for the BAV199 is 3pA typical and 5nA maximum.  They are not even testing them for low leakage and for a good reason; leakage tests into the picoamp range take significant time which costs money.  Non-gold doped small signal transistors offer similar performance and if you want reliable picoamp leakage, you will have to grade devices yourself even with the BAV199.

There are also PAD-1 and etc. (maybe discontinued now?), which ironically don't even work as well as a BJT C-B junction.  (But, as you note -- they don't need to be qualified for low leakage, they already guarantee it.)

[David Hess]

You have to qualify and grade for the right transistor; it depends on the manufacturer and process.  Not all 2N3904s are created equal.

The collector-base junctions are pretty slow but high voltage low leakage diodes are slow as well which the BAV199 datasheet shows.  When used for an input protection circuit, this does not matter unless you also want fast recovery from overload.

The good emitter-base junctions I tested had about equally low reverse leakage compared to the collector-base junctions in the picoamp to 10s of picoamp range but of course the reverse voltage is low.  The good ones had a recovery time which was close to my best reference level pulse generator at about 600ps.  The fast low leakage but low voltage diode I was thinking of was used as a clamp in the z-axis amplifier in a Tektronix oscilloscope and was rated at I think 600ps, 15 volts, and 50pA.  I wonder if I can find it going by memory since I did not keep my written notes ...

Tektronix 465 - CR1464 - 152-0153-00 - 15V 50mA - Fairchild Semiconductor FD7003,PG7026,TID777

Well, I was close.  15 volts maximum but 10 volts working, 100nA (not low leakage but I think that is the maximum specification), and 750ps which is the fastest silicon diode Tektronix listed a reverse recovery time for.

Next time I will get some photographs of my results but I am holding off until I have a faster pulse generator and I want to use a faster oscilloscope.

[T3sl4co1l]

Ah.

Also, what does matter for protection: forward recovery, which is proportionally worse for higher voltage (and slower) diodes.  This manifests as series inductance, so the overshoot can be significant!

Tim

[David Hess]

Also, what does matter for protection: forward recovery, which is proportionally worse for higher voltage (and slower) diodes.  This manifests as series inductance, so the overshoot can be significant!

Robert Pease mentioned and complained about that in his book.  I have never encountered it myself (or did not recognize it) but from what I remember, he said he was working on a voltage-to-frequency converter where long forward recovery from a bad batch of diodes manifested as nonlinearity.

I am reading what he said now and the overshoot was not inductance because it changed with repetition rate; it was worse at lower repetition rates and got better at higher repetition rates.  The "bad" diodes he encountered came from a process change by the manufacturer.

A little later he mentions the 2N3904 base-emitter diode:

One “diode” that does turn ON and OFF quickly is a diode-connected transistor.  A typical 2N3904 emitter diode can turn ON or OFF  in 0.1 ns with negligible overshoot and less than 1 pA of leakage at 1 V, or less than 10 pA at 4 V.  (This diode does, of course, have the base tied to the collector.)  However, this diode can only withstand 5 or 6 V of reverse voltage, and most emitter-base junctions start to break down at 6 or 8 V.  Still, if you can arrange your circuits for just a few volts, these diode connected transistors make nice, fast, low leakage diodes.  Their capacitance is somewhat more than the 1N914’s 1 pF.

I will add that you need to qualify and grade them yourself for this kind of performance.

[T3sl4co1l]

Oh and, the waveforms above are the junctions only -- other pin open.  With the other pin tied to base, IIRC, recovery all but goes away (you're using the transistor in its linear mode, which is fast).

Tim

[tszaboo]

BAV199 is still in production from multiple sources 

http://www.findchips.com/search/bav199

I'm aware. If they have managers, who decide to do the same, than what.

There is still lots of small manufacturers who make these parts and they don't disappear ever. Its like with vinyls as big manufacturers close, small one and enthusiasts start their business. Today these simple semiconductors can be simply made by group of enthusiasts with old equipment.

The last VHS manufacturer stopped the production. Tubes are not really made anymore. The issue is with the discrete semiconductors is that we still use them. Glue logic, analog stuff, power supply control and so on. If they say, BAV199 is not made anymore, but there is MBJ199HC (made up number) instead, with the exact same characteristics, who cares? Except if you work in a big company, and you need to create the part in the ERP system, build a prototype to verify that it is exactly the same, verify that the "soft characteristics" you relied on are exactly the same, step the revision of your design, and do it every time this happens. Suddenly the who cares becomes weeks of your time.

[jitter]

On the other hand, that much work may mean that you and maybe some other people get to keep their jobs...