Transistor Operation and Operating Regions in Diode Connected BJT

I have read wikipedia and other websites regarding the diode connected BJT.

But I am not clear on where should we use a diode connected BJT or like what is the purpose.

My questions :

1. Where is a diode connected BJT used? Like, in what type of requirements do we need to use a diode connected BJT instead of a diode itself?
2. In a current mirror circuit, the diode connected BJT is used. What would be the operating regions of both the transistors in the current mirror circuit and how to select the resistor value in the current mirrored branch?

Please explain in simple terms.

"Sometimes a cigar, is just a cigar" - Freud (apocryphal)

But in junction isolated technologies any but the
simplest substrate diode, is some aggregation of
transistors and diodes. Every NPN contains a
substrate PNP as does every lateral PNP.

Anyway, sometimes you just want a diode but
are either not offered one (been there) or want
the superior log-linear range of the E-{C+B}
transdiode.

That diode is operated at the edge of saturation
and current density needs to be kept such that
you don't get internal forward bias of the order
that makes Cjc swing wide (charge storage) as
this is usually poorly modeled and variable.

Diodes that can be made, will have varying
attributes. A "40V" NPN could be configured
E-{C+B} or C-{E+B} or the E could be left
out (but sometimes you do not get to order
ala carte, only use what's offered in the foundry
design libraries, would have no model for a C-B.
The E-B diode would break down at maybe 7-8V
in a vertical NPN while C-B over 40V.

On my first high voltage chip design I meant to
use a C-B diode in an output clamp role, but
ended up with an E-B diode in its place. This
was back when schematics were in pencil and
layout on Calma stations, so there was no real
library discipline and one man's pencil drawn
diode is another man's guess, as to which.

Anyway, at FOK wafer probe I was peering
through the microscope and saw little flashes
of light from the die, and felt stings on my
forehead. Those were the E-B diodes' silicon
being launched out of their dielectric isolation
tubs after turning orange-hot. Something
about tens of watts in square microns of area.

1. Where is a diode connected BJT used?

Click to expand...

Generally it has characteristics closer to the ideal diode than just a diode.

2. In a current mirror circuit, the diode connected BJT is used. What would be the operating regions of both the transistors in the current mirror circuit and how to select the resistor value in the current mirrored branch?

Click to expand...

The idea in a current mirror is that the base-emitter voltages of the two transistors are equal when their currents are equal.
Thus the current through the diode-connected BJT will create a base-emitter voltage that ideally generates the same collector current in the other BJT (assuming they are matched transistors on the same substrate)
In practice the currents are not exactly equal due to the finite collector-emitter resistance of the BJT, so it's current will vary some with collector voltage.

The current-mirror operation is further discussed here.

A significant improvement over the two-transistor mirror is the three-transistor Wilson current-mirror.

AbhinavRajan said:

But I am not clear on where should we use a diode connected BJT or like what is the purpose.

Click to expand...

BJT diode-connected transistors are used e.g. in bandgap circuits.

I have a question: Could somebody explain why BJT diode-connected transistors are used in bandgap circuits and not just PN diodes? Is it as crutschow said:

crutschow said:

Generally it has characteristics closer to the ideal diode than just a diode.

Click to expand...

and also due to the fact that emitter in PNP BJT serves as shielding for PN collector-base diode?

Hi,

Isn't one reason that diode-connected BJTs have a near-linear response to temperature regarding delta Vbe but diodes have an exponential response? Maybe it's easier to ratio emitter widths than diode parameters on an IC. BJTs are readily available in IC design, diodes not so, IC diodes are in reality diode-connected BJTs. Please someone correct any non-factual observations.

crutschow said:

Generally it has characteristics closer to the ideal diode than just a diode.
The idea in a current mirror is that the base-emitter voltages of the two transistors are equal when their currents are equal.
Thus the current through the diode-connected BJT will create a base-emitter voltage that ideally generates the same collector current in the other BJT (assuming they are matched transistors on the same substrate)
In practice the currents are not exactly equal due to the finite collector-emitter resistance of the BJT, so it's current will vary some with collector voltage.
[/URL][/B].

Click to expand...

Thank you for the reply. Could you also provide an answer to the second part of the question, "What would be the operating regions of both the transistors in the current mirror circuit and how to select the resistor value in the current mirrored branch?" How to find the region of working of both the transistors and how to figure the maximum and minimum currents in the current mirrored branch?

- - - Updated - - -

crutschow said:

Generally it has characteristics closer to the ideal diode than just a diode.
The idea in a current mirror is that the base-emitter voltages of the two transistors are equal when their currents are equal.
[/URL][/B].

Click to expand...

Thank you for the reply. And when you mentioned "The idea in a current mirror is that the base-emitter voltages of the two transistors are equal when their currents are equal." - Are you referring the Collector current or the Base current?

Hi,

If I may... More or less:

BJTs are in active region.
Rbias (Iout) = Vsupply - Vbe. If emitter resistors are added, I assume you also need to subtract votage across Remitter (Remitter x Iout). So, (Vsupply - Vbe)/I out = Rbias
Maximum and minimum currents are a function of Vsupply min to Vsupply max, and min and max operating temperature (which will include ambient temperature and junction temperature based on BJT PD, I suppose).
Ic is same as Ib in a mirror, I think. Perfectly matched transistors are an idealized aspiration in reality. Look at a dual or quad BJT datasheet to see transistor to transistor minor variations in hFE and/or Vbe.

These are nice to read:

Current Mirror BJTs

The Current Mirror

d123 said:

Hi,

Isn't one reason that diode-connected BJTs have a near-linear response to temperature regarding delta Vbe but diodes have an exponential response? Maybe it's easier to ratio emitter widths than diode parameters on an IC. BJTs are readily available in IC design, diodes not so, IC diodes are in reality diode-connected BJTs. Please someone correct any non-factual observations.

Click to expand...

Thank you for your answer.
A diode (PN) should have a linear temperature coefficient:
http://www.bndhep.net/Lab/Derivations/Diodes.html
I think I calculated something similar when designing bandgaps, that is, I used diode equations not bipolar equations. Onced, I even designed a bandgap in LTspice using diodes (not BJT) and it worked.

Moreover, PN diodes are present, depending on the process, in IC design. Hence my question. I suspect an usage of vertical PNP BJTs may be due to noise properties (emitter around base and collector, connected to ground).

How to find the region of working of both the transistors and how to figure the maximum and minimum currents in the current mirrored branch?

Click to expand...

The collector current in the non diode-connected transistor will basically equal the current in the diode-connected transistor (assuming there is a bias voltage connected to the collector of the non diode-connected transistor).
So there is no particular minimum current and the maximum current is determined by the transistor rating.

Normally a resistor is used in series with the diode-connected transistor with value selected to give the desired collector-emitter mirror-current in the other transistor.