What is the real maximum voltage for a BJT?

[Norm Dresner]

Data sheets give information like BVceo, BVebo, ... and I think I've seen
an occassional BVcer ... as well.

What's the absolute maximum safe value to use ? Okay, I don't mean absolute
safe valeu because that's zero, but realistically, if I keep the voltage
between the C & E less than BVceo, am I guaranteed that the transistor won't
break down (that way)?

Is there any safety margin in these figures?

TIA
Norm

 

[Don Klipstein]

Data sheets give information like BVceo, BVebo, ... and I think I've seen
an occassional BVcer ... as well.

What's the absolute maximum safe value to use ? Okay, I don't mean absolute
safe valeu because that's zero, but realistically, if I keep the voltage
between the C & E less than BVceo, am I guaranteed that the transistor won't
break down (that way)?

Is there any safety margin in these figures?

When switching with the base being instantly shorted to the emitter or a
voltage slightly lower than the emitter voltage (more positive for PNP),
the CEs rating or the CB rating apply.

If a CEr rating is available, I consider it a with-safety-margin rating
for maximum voltage with no current flowing.

When any current is flowing through the collector, the maximum voltage
is the CEo rating. Furthermore, you should get a safe area curve. Many
transistors have their maximum safe power dissipation compromised at
higher voltages - typically above something like 40 volts.

Maximum safe power dissipation of a heatsinkable transistor is with
heatsinking that cools the case to 25 degrees C. (Niagara Falls +/- 500
GPM.) Even then, life expectancy of some may be less than that of some
incandescent lamps. I would avoid going over half the rated power
dissipation even with really favorable heatsinking.
Power MOSFETs are more forgiving in my experience, and I feel
comfortable pushing to 60% of their rated power dissipation with really
good heatsinking.

- Don Klipstein ([email protected])

 

[John Larkin]

Data sheets give information like BVceo, BVebo, ... and I think I've seen
an occassional BVcer ... as well.

What's the absolute maximum safe value to use ? Okay, I don't mean absolute
safe valeu because that's zero, but realistically, if I keep the voltage
between the C & E less than BVceo, am I guaranteed that the transistor won't
break down (that way)?

Is there any safety margin in these figures?

TIA
Norm

BVcbo or BVcer are safe numbers to use in the real world in most
circuits. I assume the manufacturer applies the margin so I don't have
to, but I do have to worry about worst-case transient situations.

BVceo is worst-case, a bit pessimistic unless you really intend to run
it with the base open. But to be conservative, use this one.

John

 

[Winfield Hill]

Don Klipstein wrote...

Maximum safe power dissipation of a heatsinkable transistor is with
heatsinking that cools the case to 25 degrees C. (Niagara Falls +/-
500 GPM.) Even then, life expectancy of some may be less than that
of some incandescent lamps. I would avoid going over half the rated
power dissipation even with really favorable heatsinking.
Power MOSFETs are more forgiving in my experience, and I feel
comfortable pushing to 60% of their rated power dissipation with
really good heatsinking.

Given the datasheet's power dissipation rating is for a 25C case,
which is never realized in the real world (with high case-to-heat-
sink thermal resistances, and warmed-up heat sinks), operating at
60% of the FET's rating no doubt means you're pushing the junction
temperature well over its maximum value. Yes MOSFETs are forgiving
and operating a 150C-rated FET at 175, 200 or even 225C will not
result in any noticeable changes in most circuits, and probably will
not result in a dramatically shortened operating life, leading you
to think it's OK. But if you're not going to carefully make all
the appropriate measurements and calculations, it's better to use
a more modest rule of thumb, e.g., 30 or 40% of rated dissipation.

Thanks,
- Win

(email: use hill_at_rowland-dot-org for now)

 

[Kevin Aylward]

Don Klipstein wrote...

Given the datasheet's power dissipation rating is for a 25C case,
which is never realized in the real world (with high case-to-heat-
sink thermal resistances, and warmed-up heat sinks), operating at
60% of the FET's rating no doubt means you're pushing the junction
temperature well over its maximum value. Yes MOSFETs are forgiving
and operating a 150C-rated FET at 175, 200 or even 225C will not
result in any noticeable changes in most circuits, and probably will
not result in a dramatically shortened operating life,

Ahmmm... did we not discuss this before?

Tut, tut, tut. Plug the *actuall* numbers into the Arrhenius equation.
Major differance dude!!!

R = A.exp(Ea/KT)

You can easily knock a life of 10,000 hrs down to a few. I get a factor
of 1000 for a 150C to 200C change with an Ea of 1eV.

I would never run a device at 150C as standard. Its essentially
guaranteed to fail over time. I'v watched it happen.

Kevin Aylward
[email protected]
http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

 

[Winfield Hill]

Kevin Aylward wrote...

You can easily knock a life of 10,000 hrs down to a few. I get
a factor of 1000 for a 150C to 200C change with an Ea of 1eV.

I should have chosen my words more carefully. What I wanted to
convey was that the FET's life won't be shortened sufficiently
for the user to realize his error. You know, the old "Well, it
worked OK for me" argument. My point was that using 60% of the
datasheet power rating likely means the junction will be pushed
over-temp, even though it seems to work OK. The cool-operation
ten-million-hour life will unknowingly be reduced to thousands.

Thanks,
- Win

(email: use hill_at_rowland-dot-org for now)

 

[Jim Thompson]

BVcbo or BVcer are safe numbers to use in the real world in most
circuits. I assume the manufacturer applies the margin so I don't have
to, but I do have to worry about worst-case transient situations.

BVceo is worst-case, a bit pessimistic unless you really intend to run
it with the base open. But to be conservative, use this one.

John

Only **I** know for sure ;-)

...Jim Thompson

 

[John Larkin]

Only **I** know for sure ;-)

...Jim Thompson

But Jim, we're all sure that you are conservative.

John

 

[Bill Sloman]

But Jim, we're all sure that you are conservative.

If not actively reactionary ....

 

[John Larkin]

Only **I** know for sure ;-)

...Jim Thompson

Jim: my "conservative" comment was meant to be whimsical, not nasty (I
consider myself to be fundamantally conservative, using my own
definitions of course.) But Sloman seems, of late, to follow me around
like a deranged puppy, amplifying or clarifying my comments in
directions I never intended. He used to insult me, which I actually
preferred because that, at least, verified that we have nothing in
common. He speaks only for himself.

John

 

[Jim Thompson]

But Jim, we're all sure that you are conservative.

John

When it comes to circuits I'm a pessimist.

...Jim Thompson

 

[Bill Sloman]

Jim: my "conservative" comment was meant to be whimsical, not nasty (I
consider myself to be fundamantally conservative, using my own
definitions of course.) But Sloman seems, of late, to follow me around
like a deranged puppy, amplifying or clarifying my comments in
directions I never intended. He used to insult me, which I actually
preferred because that, at least, verified that we have nothing in
common. He speaks only for himself.

Who else might I have been speaking for?

And can you recall when I insulted you? I don't share your political
opinions, but your ideas about engineering have always struck me as
very respectable.