"Aged" bipolar transistors in Sony/Tek 323 scope, ca. 1968; any experience?

[zike]

I just fixed up a couple battery-damaged 323 mini-scopes and one had a peculiar (non-battery-related) fault. Two Tek 151-0220-00 (aka 2N4122) PNP transistors were still 'functioning,'  but with only a third of their original specified current gain. This rendered their transitions too sluggish for the intended functions (blanking and horizontal retrace).

These were Q356 and Q326 in sheet 4 of the 323 scope schematic (manual is available here).

Based on context and annotated voltages/waveforms, operating conditions are nowhere near the device limits. Q356 is strictly a switch, always cut off or saturated, and Q326 is briefly saturated at the end of each sweep, but currents are mild and there's no appreciable overdrive or reverse bias. The whole scope runs on 4 watts; nothing in it gets appreciably hot.

I'm curious about similar "soft" or progressive failures with these parts, or silicon bipolars in general? Is there an aging mode that's unrelated to stress?

 I have some 1980's vintage parts on order, it will be interesting to compare. For now I got everything working with a couple 2N3906 (selected for beta). 

[David Hess]

One of my Tektronix frequency counters had a pair of bad 2N3565s which made me suspicious.  Testing the rest of the 2N3565s revealed that about half were "soft", with poor Vce saturation so they had significantly lower gain at low Vce.  I ended up replacing them all with 2N3904s.  A curve tracer is invaluable for this sort of work.

Usually this sort of thing is caused by a manufacturing problem or contamination during or after packaging.

I would not bother with NOS (new old stock) parts unless the part is something like a fast saturated switch which has no alternatives.  The 2N4122 can likely be replaced with the common 2N3906.

[floobydust]

I have noticed two things - old (power) transistors have low h_FE, technology has greatly improved that.
But they seem to age, h_FE dropping to pathetically low numbers with NOS parts in the bin. Example: 1978 TIP29/TIP30 lucky to get an h_FE of 15, one was 12 so I laughed and threw them all out.
Not sure if hermetic packages also "age" implying moisture is a factor with old plastic parts 

[David Hess]

I have noticed two things - old (power) transistors have low h_FE, technology has greatly improved that.
But they seem to age, h_FE dropping to pathetically low numbers with NOS parts in the bin. Example: 1978 TIP29/TIP30 lucky to get an h_FE of 15, one was 12 so I laughed and threw them all out.
Not sure if hermetic packages also "age" implying moisture is a factor with old plastic parts 

Were those tested under the conditions given in the datasheet?

[ondraN]

I have refurbished several Tektronix oscilloscopes from those years (422,453, 454A...), but I have not come across anything like this. It is possible that the gain of some transistors has decreased, but it was not noticeable in the function of the oscilloscopes.

[Agent24]

A long time ago I repaired a 1980s TTi DSA-524 which (among other faults) had a failure of a power transistor (I think TIP41) in one of the power supply rails. From the darkened base resistor, I wondered if it had failed with low gain also, requiring more current to turn it on, overheating the drive circuit. When I load tested the transistor, it failed open-circuit completely. Visibly it looked totally fine IIRC.

[zike]

Not sure if hermetic packages also "age" implying moisture is a factor with old plastic parts 

I feel like I've also found unused old-stock TO-39's with curiously low gain, in the teens, but never followed up whether it was allowed by spec or evidence of decay.  I'm talking pre-1970's, like 2N1613 and such. No real application in mind, just going through hoarded junk.  Wonder if i still have them someplace.

I had to give away my 575, downsized, but I held on to a 7CT1N and some jigs.

[wn1fju]

I've seen this a few times with my repairs, mostly on power transistors.  However, this kind of fault is minimal compared to the number of times I've seen open or shorted transistors.

[ondraN]

I was once repairing a vintage turntable, I think it was a Yamaha 800. It had a direct drive with purely discrete components. One of the transistors in the differential comparator had occasional shot noise that caused the speed to fluctuate. I don't know the exact type, but it was a PNP transistor with high gain (about 900) in a plastic case. It behaved normally during measurements.

[Hr_Satch]

Hi there, strange, a die of a Si-semiconductor that derates with age ? ? ?

Over +10.000 years perhapse yes, but here in a time periode laps of merely max. a 100 years ? ? ?

That's new to me.

It is not the semiconductor that sets the gain, it are the surrounding components thou !
The required setup: common Emitter / common Base / a source follower, etc . . .

The Datasheet just predicts what kind of Max. gain, Voltage, current, etc . . . can be achieved, per model-type, . . .
a 100Watt transistor for example, doesn't perform a 100 Watts just like that.
You also need to supply the necesary electrical power for it, and the required setup, in a ceratian Load, . . .

So when suspecting a so-called transistor gain drop: isn't it rather pointed to check the surrounding components ?
And applied voltages, that drops often due to capacitor exhaustion, read increased ripple, which drops the average ;-)

And that is Electronics ! The Gain of a transistor setup, is due to its surrounding resistors bias, . . .

Regards, harry.

[ftg]

Many older semiconductors degrade with time because the packaging is not perfect, the manufacturing process is not perfect and crucially the passivation is not perfect.

[zike]

... and crucially the passivation is not perfect.

This. There's quite a bit of research but most of it (e.g., this) measures loss of gain due to repeated breakdown or ionizing radiation bombardment, leading to carrier recombination in the base which robs carriers. But I've seen comments (still looking for original sources?) suggesting imperfect surface passivation or cumulative contamination can shunt carriers around the junction. I'd have thought this would also produce reverse leakage, though, which I don't see.

And here I thought the Ge tin whiskers would be the end of it... I reckon lots of things go to pot after 60+ years. Me included.

[floobydust]

I have noticed two things - old (power) transistors have low h_FE, technology has greatly improved that.
But they seem to age, h_FE dropping to pathetically low numbers with NOS parts in the bin. Example: 1978 TIP29/TIP30 lucky to get an h_FE of 15, one was 12 so I laughed and threw them all out.
Not sure if hermetic packages also "age" implying moisture is a factor with old plastic parts 

Were those tested under the conditions given in the datasheet?

No, not the 300usec pulse test V_CE=4V I_C=0.2A spec h_FE min. 40, I used 10V with a resistor to supply I_B, I_C not to make too much heat. Much lower test current.
But the decider was all of them measuring very low hFE - compared to all other parts in the bins, except on par with some old RF PA ones as well. All from the 1970's.
Actually I think the TIP29 parts were dated 1974 and sorta TO-220 with the plastic cap and epoxy filling.

I have no idea of the aging mechanism, have not tried testing hermetic parts to see if it's aggravated by moisture ingress.
I also see ion-implanted PMOS IC's age, the outputs go leaky and lower transconductance. At the 50 year mark.

[factory]

Failure of plastic small signal transistors for no apparent reason, is common with some HP equipment from the 1960s & 1970s, sometimes they fail open, or go very intermittent.

Also heard of some Japanese transistors in consumer AV equipment, failing due to silver migration, the leads being plated with silver.

Edit: This thread on another forum was where I heard about problems with silver plated transistor leads;
https://www.vintage-radio.net/forum/showthread.php?t=205318

David