Low-voltage rail for transistor bases?

[Lauri Alanko]

Hello.

When power transistors are used as switches, the base current can be
significant, as the saturation beta may be as low as 10. On the other
hand, the base _voltage_ need only be at Vbe(sat), and any additional
voltage needs to be dropped by e.g. a resistor. Given the high
currents, this voltage drop may be a significant loss.

This suggests to me that ideally the base of a low-side NPN switch
would be powered by a low-voltage rail, only slightly higher than
Vbe(sat). This way the conduction loss would be minimized.

Has such a rail been used with bipolar switches? If not, what's wrong
with my thinking?

(Please don't tell me to use MOSFETs because they are better. I know
that already. I'm looking for understanding, not solutions.)


Lauri

 

[John Fields]

On Sat, 21 Sep 2013 19:30:44 +0000 (UTC), Lauri Alanko <la@iki.fi>
wrote:

Hello.

When power transistors are used as switches, the base current can be
significant, as the saturation beta may be as low as 10. On the other
hand, the base _voltage_ need only be at Vbe(sat), and any additional
voltage needs to be dropped by e.g. a resistor. Given the high
currents, this voltage drop may be a significant loss.

This suggests to me that ideally the base of a low-side NPN switch
would be powered by a low-voltage rail, only slightly higher than
Vbe(sat). This way the conduction loss would be minimized.

Has such a rail been used with bipolar switches? If not, what's wrong
with my thinking?

(Please don't tell me to use MOSFETs because they are better. I know
that already. I'm looking for understanding, not solutions.)


Lauri

---
If you had a low voltage rail just to use as a base-to-emitter supply,
you'd still have to switch it into the base, which would require
another transistor: (View using a fixed-pitch font.)

Vcc>-----------------------+
|
Vbb>-------------+ |
| [RL]
E |
| O--[R]--B PNP C
-->| C--[R]--B NPN
| O-+ E
| |
GND>----+-----------------+

Which would work, and the dissipation would be lower than driving the
NPN's base from Vcc.

--
JF

 

[John Larkin]

On Sat, 21 Sep 2013 19:30:44 +0000 (UTC), Lauri Alanko <la@iki.fi> wrote:

Hello.

When power transistors are used as switches, the base current can be
significant, as the saturation beta may be as low as 10. On the other
hand, the base _voltage_ need only be at Vbe(sat), and any additional
voltage needs to be dropped by e.g. a resistor. Given the high
currents, this voltage drop may be a significant loss.

This suggests to me that ideally the base of a low-side NPN switch
would be powered by a low-voltage rail, only slightly higher than
Vbe(sat). This way the conduction loss would be minimized.

Has such a rail been used with bipolar switches? If not, what's wrong
with my thinking?

(Please don't tell me to use MOSFETs because they are better. I know
that already. I'm looking for understanding, not solutions.)


Lauri

Couple of issues:

Vbe_sat isn't really well defined. If you jam a fixed voltage onto the base, the
base current could vary a lot with temperature, or between different
transistors.

Transistors have equivalent emitter resistance on-chip, and more in the leads
and circuit board traces. At high currents, the drop in the total emitter
resistance reduces the voltage across the actual, internal b-e junction, so
reduces the base current, which is what you don't want to do. And if you tune it
for the loaded condition and then there were no load current for some reason,
the base current might get huge.

I practice one rarely sees (ie, I've never seen) a circuit that tries to drive a
switching transistor base at a fixed voltage. There's usually enough resistance
in the driver to force the current to be pretty constant against changes in Vbe.

Mosfets are better!



--

John Larkin Highland Technology Inc
www.highlandtechnology.com jlarkin at highlandtechnology dot com

Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom timing and laser controllers
Photonics and fiberoptic TTL data links
VME analog, thermocouple, LVDT, synchro, tachometer
Multichannel arbitrary waveform generators