effect of reverse b/e current over time

I'm looking at using an astable multivibrator, like the one in the
diagram at this link
http://www.play-hookey.com/digital/experiments/rtl_astable.html
running probably at a few tens of kiloherz
to drive a mosfet in a forward converter.
I'll probably run the multivibrator somewhere between 12 and 18 volts.
That will put a reverse bias on the base-emitter junctions of the
transistors
in the multivibrator as the base voltage gets pulled below ground
during oscillations.
I've heard that reverse biasing the base emitter junction affects gain,
but that the effect
is of limited severity. I can live with the loss of a little bit of
gain. But what about the long term effects -- will it damage the
transistors, or will they operate reliably for something like their
normal lifetime if they are not abused otherwise?

 

I have seen such circuits fail, when the gain fell so low that
oscillation stopped. It is a reliable circuit with a 5 volt supply,
but one that damages the transistors when run from a 12 to 18 volt supply.

 

I would imagine a resistor (comparable to the collector resistors) in series
with the base and a diode across E-B to clamp the negative voltage to 0.6V.

I don't know if a series resistor would stop it from oscillating.

Tim

 

I'll put a zener in series with each transistor's base.

 

You could get around the issue of breakdown by using FETs instead - a
VN2222LL (TO92 package for through hole) or 2n7002 (SOT-23 SMT) should
work fine. They are both rated at Vgs of +/- 20V continuous, 40V
non-repetitive pulse.

Cheers

PeteS

 

Beta degradation as a result of b-e zenering is a legend, but I've
never seen a quantitative estimate of the effect. Somebody could do an
experiment or two...

John

 

Not an urban legend, at least.

I just have made the experiment with a brand new BC547C.

New, @VCE=2V, IC=1mA, (beta ~ 370).

Carefully zener it (no cap discharge) for 15 seconds at 10mA (BE broke down
at 10.1V)

Then under the same conditions, IC dropped to 965uA, a 3.5% beta drop.
That's not insignificant at all.

What's more noticable is that as I'm writing this I can see the beta still
sligthly dropping to 957uA (more than 4%). Perhaps the starting 965uA
benefited from the 100mW heating from BE zenering.

I now have to get my little daugther to a painting exhibition.

More zenering this evening...

 

"Fred Bartoli"

So, more zenering...

30s @ 10mA, IC dropped to 950uA -> 5% beta drop;
2min @ 10mA, IC dropped to 938uA -> over 6%.

Time to be a bit more nasty.
1 min @ 30mA, IC dropped to 922uA -> 8% beta drop.

2 hours @ 30mA, IC dropped to 843uA -> almost 16% beta drop.

One noticable thing is that the normal working VBE constantly dropped with
the cumulative zenering, from 615mV down to 604mV.

It could also be interesting to see what happens with cap discharge like
with a multivibrator but I've no more time now.

Maybe someone else?

 

Nice, first numbers I've seen on this. Looks like the beta degradation
is falling off as the current-time product increases.

I'll have to try this myself one of these days.

I wonder if LEDs are degraded by reverse current.

Thanks.

John

 

I've become amuzed with this now and I'm emulating the 20V supply astable
case with a 50ohm/20V generator @ 10kHz, and a 22nF cap. Ouch, about 160mA
peak zenering current. Probably about what it'd be in such a circuit.

10 min at this regime and beta dropped further down to -18%
30 min later again brings us down to -20% and 603mV VBE.

I don't know which region of the BE junction is zenering first and what the
degradation mechanism is.
Maybe Jim has some knowledge about this.

When I've time I'll open a TO18 or a TO5 (if I can still find one) and try
to have a look.

You'll have to be pretty persuasive with LEDs. The ones I've tried happily
withstood 60V with under 10pA leakage IIRC.