Reverse Vbe exceeded, long term gegradation

While reviewing a design (not mine) I came upon a situation where repetitive pulses exceed reverse Vbe so much that it'll cause breakdown. It's only around a milliamp at 10-20% duty cycle but this can go on for weeks, months, years. Some studies show a steady deterioration of beta but they are looking at much higher reverse base currents such as 60mA here:

formatting link

A slight deterioration which then reliably plateaus would be ok while a steady slow drop in beta would be problematic. How about 1mA with a regular run-of the-mills 2N3904 type?

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg
Loading thread data ...

I can't give you numbers off the top of my head, but it's a function of temperature as well. The basic rule is DON'T DO IT. ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
 Click to see the full signature
Reply to
Jim Thompson

That's what I already have in the draft of my review. I'd just like to back that up with something. Like a longterm test at very low current. Otherwise I'll just write "Jim in Arizona said that a major catastrophe could ..." :-)

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

tell em it'll work for a while and that some of the damage can be reverse by touching the transistor with a lit cigarette, or so I've heard

Reply to
Lasse Langwadt Christensen

AFAIR that only worked with hand-rolled Bull-Durham cowboy tabaccah. Aside from all the non-smoking laws this also leaves an odd stench when trying it on SOT23.

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

A milli amp at ~10V doesn't sound too bad. I'm doing ~0.1 mA at 100V. That's a fault condition so most of the time not active.

George H.

Reply to
George Herold

That 1 mA is not a problem for electromigration of aluminum wires on the silicon device. Little heating will occur, so no diffusion of dopants will be any problem. Run away heating can happen and the 1mA can increase to 2mA. Rejoice.

Reply to
Alan Folmsbee

Good grief, they applied 8 volts and got 60 ma of base zener current. That's half a watt. They fried the poor thing.

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
John Larkin

I used to have a few white noise generators that used bc547 - the breakdown current was .3 to .5 ma. The transistors were useless for anything else after an hour or so. The gain seems to drop of logarithmically so after a week or so they "looked" very stable. but were really just trouble waiting to bite someone in a year or two. A vga might be a solution or a diode to stop the reverse current.

--
Using Opera's mail client: http://www.opera.com/mail/
Reply to
David Eather

It's not the current, it's the surface-of-the-transistor electric field that does the damage (it moves crud around). When the collector junction is a buried layer, it can stay biased without pulling crud from the surface Base and emitter connections on the same surface, though, mean you can get ion transport, even with 'good enough' passivation.

So, old germanium alloy transistors had surfaces for BOTH junctions, and were a mess. Best bet is to use a 2N3904 that has low voltage breakdown (maybe

5V or so?), and maybe that isn't enough field to matter. National actually had app notes on using the breakdown of their 2N3904's B-E junction as a zener, but it's off-datasheet behavior.

What's ON-datasheet, is acceptable aging if you (for instance) diode-clamped the reverse spike. Costs you another component, though, or maybe a PMD4001K would work .

Reply to
whit3rd

Quote:

"The reverse-biased emitter-base junction is usually spec'd at 5V or 6V. A higher voltage (as Jony showed) causes the junction to have avalanche breakdown (like a zener diode). The avalanche breakdown causes hot spots on the emitter-base junction and since it is not designed to dissipate heat like the collector is then the transistor is slowly damaged."

formatting link

Much more info is available on google.

Reply to
Steve Wilson

I looked into this a few years ago and posted results here, I think latest was Jan 22 this year in response to a query by John Larkin about operating BJT in reverse.

Yes, beta does degrade: rapidly at first then levels off. In my application the victim was exposed to capacitive discharges that zenered the b-e junction with low energy. I exposed DUTs to the equivalent of decades of operational use periodically measuring beta. Final beta was about 65% of the initial beta. The app was a simple switching circuit where beta only needed to be above 50 and has proven absolutely rock solid.

However I have read that even briefly zenering b-e junctions wrecks low-noise devices which become noisy. So while reverse biasing may be alright in switching circuits I think protection should be used in linear usages.

piglet

Reply to
piglet

In 7/27/97, Tony Williams (RIP) studied zener noise at low currents. He clearly showed the existence of microplasma hot spots by the random noise they created. The thread, "Zener Diode Oscillation" is available here:

formatting link

Some example waveforms are shown in "Microplasma Noise as a Tool for PN Junctions Diagnostics", at

formatting link
et-al.pdf

formatting link

Reply to
Steve Wilson

Base-emitter breakdown is bad news, for sure. How about a nice

2SD2704K? Superbeta, low V_CEsat, -25V V_BEmax. A bit on the slow side, but there it is. A Schottky diode in series with the base (maybe with a speedup cap in parallel) would be the other approach.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

Doesn't work as well with SC-74s. ;)

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

What transistor has a reverse base-emitter current that low at 100V? Sounds like a very useful device!

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

The problem is hot-carrier damage in the base region, not dopant migration. You get vacancies and interstitials, which cause recombination. Since the problems are local, annealing gets rid of them.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

Interestingly the degradation was less at higher currents, so they were probably annealing the defects as they went along. The lower-current results were fairly horrifying.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

Oh sorry a photodiode... I was just thinking in energy/ power terms. (Reading reports of others, there seems to be more going on with a transistor.)

George H.

Reply to
George Herold

Yes, but we aren't supposed to smoke in the labs anymore. Question is, if deployed in the field and annealing never takes place, does this process creep on and on until one day the beta is so low that things fall off the cliff?

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.