2N3055 barrier height potential

[Paul Burridge]

Hi,

Why does this device seem to have such a low inherent Vbe for a
silicon tranny?

Thanks,

p.

 

[Spehro Pefhany]

Hi,

Why does this device seem to have such a low inherent Vbe for a
silicon tranny?

0.72-0.78V typical @ 0.1A is low?

Best regards,
Spehro Pefhany

 

[John Larkin]

Hi,

Why does this device seem to have such a low inherent Vbe for a
silicon tranny?

Thanks,

p.

There is no inherent Vbe, just a diode exponential curve and some
extra base and emitter resistances; the resistances aren't important
at low currents. The larger the device, the higher the exponential
curve and the lower the resistances, generally.

John

 

[Paul Burridge]

0.72-0.78V typical @ 0.1A is low?

I'm only seeing 350mV!
See the schematic on a.b.s.e under the same subject header as this one
for the test circuit This is a simulation I'm running. I can
substitute *any* other BJT in the LT Spice model library and get the
usual 0.6V Vbe., but for some reason, the 2N3055 collapses to just
over half this value. I'm just investigating distortion in grounded
emitter amp stages and this odd result cropped up for the '3055 alone.
There must be something unusual about the characteristics of the '3055
I'm not aware of. Or there's something else I've not allowed for, of
course. Anyone spot the problem?

thanks,
p.

 

[Paul Burridge]

There is no inherent Vbe, just a diode exponential curve and some
extra base and emitter resistances; the resistances aren't important
at low currents. The larger the device, the higher the exponential
curve and the lower the resistances, generally.

Thanks, John. I've provided a bit more background info in my follow-up
to Speff. See what you make of it.

 

[Tim Wescott]

I'm only seeing 350mV!
See the schematic on a.b.s.e under the same subject header as this one
for the test circuit This is a simulation I'm running. I can
substitute *any* other BJT in the LT Spice model library and get the
usual 0.6V Vbe., but for some reason, the 2N3055 collapses to just
over half this value. I'm just investigating distortion in grounded
emitter amp stages and this odd result cropped up for the '3055 alone.
There must be something unusual about the characteristics of the '3055
I'm not aware of. Or there's something else I've not allowed for, of
course. Anyone spot the problem?

thanks,
p.

Wait, you're doing this in simulation? And you believe it?

First rule of simulation: If the simulation doesn't match expectations,
check it against reality.

First rule of using 3rd-party software: If the software seems buggy, it
probably is.

Have you troubled to grab a 3055 out of you junk box and _check_?!? I
did, and I see about 0.52 volts with my low-current VOM, which is about
right for a big-ass power BJT.

 

[Spehro Pefhany]

I'm only seeing 350mV!
See the schematic on a.b.s.e under the same subject header as this one
for the test circuit This is a simulation I'm running. I can
substitute *any* other BJT in the LT Spice model library and get the
usual 0.6V Vbe., but for some reason, the 2N3055 collapses to just
over half this value.

At what current? Google on "diode equation". A realistic BASE current
for that old POS is 100mA.


Best regards,
Spehro Pefhany

 

[Paul Burridge]

At what current? Google on "diode equation". A realistic BASE current
for that old POS is 100mA.

Looks like you might be on to the answer there, Speff. I'd quite
forgotten how current-hungry this er, "marvellous old classic" is.

 

[Ian Stirling]

There is no inherent Vbe, just a diode exponential curve and some
extra base and emitter resistances; the resistances aren't important
at low currents. The larger the device, the higher the exponential
curve and the lower the resistances, generally.

To amplify this, with high value resistors, a transistor multivibrator
will start up well below 0.6V.

 

[Paul Burridge]

Have you troubled to grab a 3055 out of you junk box and _check_?!? I
did, and I see about 0.52 volts with my low-current VOM, which is about
right for a big-ass power BJT.

Thanks, Tim. I'm wouldn't leave your "low-current VOM" hooked up to
this SoB for too long as my revised simulation shows it takes well
over 100mA to get the Vbe to that level. I don't want to feel obliged
to buy you a new battery.

 

[Tim Wescott]

Spehro Pefhany wrote:
-- snip --

At what current? Google on "diode equation". A realistic BASE current
for that old POS is 100mA.


Best regards,
Spehro Pefhany

"Precious Old Stuff" -- right?

And yes, I do have vacuum tubes in my office, as well as germanium
transistors, surface-mount transistors and nearly-new Xilinx parts.

 

[John Larkin]

Looks like you might be on to the answer there, Speff. I'd quite
forgotten how current-hungry this er, "marvellous old classic" is.

The b-e junction of a 2N3055 is just another diode. Check the Is
parameter in the Spice model... they are sometimes absurd values,
don't know why. Measure the Vbe of a real 3055 at, say, 1 or 10 mA and
adjust the Spice Is value to get identical voltage drop in an
equivalent simulation. Expect roughly 0.55 volts at 1 ma.

I think some people believe that Is is the same thing as real-world
reverse leakage current, so they read the datasheet leakage value
(worst-case to boot) and plug that into Is. The result is absurdly low
apparent junction drop in the simulated part. I've used schottky diode
models where people apparently did just that, and the numbers were
silly.

John

 

[Tim Wescott]

Thanks, Tim. I'm wouldn't leave your "low-current VOM" hooked up to
this SoB for too long as my revised simulation shows it takes well
over 100mA to get the Vbe to that level. I don't want to feel obliged
to buy you a new battery.

It's a Fluke 77, and it sources less than a milliamp. Better review the
first rule of simulation again.

 

[Ban]

The b-e junction of a 2N3055 is just another diode. Check the Is
parameter in the Spice model... they are sometimes absurd values,
don't know why. Measure the Vbe of a real 3055 at, say, 1 or 10 mA and
adjust the Spice Is value to get identical voltage drop in an
equivalent simulation. Expect roughly 0.55 volts at 1 ma.

I think some people believe that Is is the same thing as real-world
reverse leakage current, so they read the datasheet leakage value
(worst-case to boot) and plug that into Is. The result is absurdly low
apparent junction drop in the simulated part. I've used schottky diode
models where people apparently did just that, and the numbers were
silly.

John

I just simulated the 2N3055 and got Ube=627mV at 8mA. The gain at this
operating point is almost 100. That seems realistic. I bet you didn't use
the transient analysis and your transistor is cut off. Or you have modified
the models. For a beginner not advisable.

 

[Paul Burridge]

It's a Fluke 77, and it sources less than a milliamp. Better review the
first rule of simulation again.

I think it's down to me, I suspect. I really shouldn't have continued
to use that same circuit fragment in an attempt to investigate Vbe.
It's better achieved with the current source technique Genome posted
elsewhere.

 

[Paul Burridge]

I just simulated the 2N3055 and got Ube=627mV at 8mA. The gain at this
operating point is almost 100. That seems realistic. I bet you didn't use
the transient analysis and your transistor is cut off. Or you have modified
the models. For a beginner not advisable.

Hi Ban,

No I didn't alter the model. But I'd be interested to see the model
*you* used. Can you post it here? Thanks.

P.

 

[John Larkin]

I just simulated the 2N3055 and got Ube=627mV at 8mA. The gain at this
operating point is almost 100. That seems realistic. I bet you didn't use
the transient analysis and your transistor is cut off. Or you have modified
the models. For a beginner not advisable.

I used a real 2N3055 and a voltmeter. Only beginners need to simulate
stuff this simple.

John

 

[Fred Bloggs]

I used a real 2N3055 and a voltmeter. Only beginners need to simulate
stuff this simple.

Nothing like a good mental midgeting session over some crap pedant
subject matter introduced by that decrepit UK genderless hag Burridge, eh?

 

[John Larkin]

Nothing like a good mental midgeting session over some crap pedant
subject matter introduced by that decrepit UK genderless hag Burridge, eh?

Well, he did reinforce the point that simulation sometimes produces
wildly unrealistic results. People need to get out of their chairs,
get vertical, and measure a real part once in a while. Seems like some
goodly fraction of the youngsters playing with this stuff these
believe that electronics *is* simulation. I'm told that many college
EE "lab" courses just simulate.

John

 

[Fred Bloggs]

Well, he did reinforce the point that simulation sometimes produces
wildly unrealistic results. People need to get out of their chairs,
get vertical, and measure a real part once in a while. Seems like some
goodly fraction of the youngsters playing with this stuff these
believe that electronics *is* simulation. I'm told that many college
EE "lab" courses just simulate.

John

Actually I kind of favor the idea of making the physical circuit conform
to the simulation- providing the models are accurate- a little filtering
here and there and other forms of limiting should do it. You increase
the tolerance variation on those models and the real world will be exact