Group Actions
Engagement
  • Author Author: jc2048
  • Date Created: Date Created
  • Views 4455 views
  • Likes 7 likes
  • Comments 14 comments
Related
Recommended

Transistors: the Transistor as a Diode

jc2048

I thought I was going to do saturation next, but somehow I've diverted off into looking at the transistor as a diode. This came about from my

referring to Bob Pease's book, Troubleshooting Analog Circuits, to look up something else and then thinking that I'd like to measure for

myself the graph he shows of Vf against If for various diodes, including transistors connected as diodes.

 

A bipolar junction transistor contains two PN junctions, one between the base and the emitter (B-E) and one between the base and the collector (B-C),

and either can be used as a diode. When used in a circuit, the base-emitter diode normally has the collector shorted to the base (CB-E). It is

also possible to join the collector and emitter (B-CE), giving a final possible configuration (then the two diodes are in parallel).

 

 

image

 

The following graph shows the B-C B-E and CB-E forms for a  2N37042N3704 CB-E for a BD135(a medium power device intended for audio amplifiers

along with a few common diodes. I've also included a small-signal Schottky diode (BAT42) for comparison. This plot is in the usual form.

 

image

 

This is the same data with the current plotted on a log scale. This has the advantages that we can see the low current detail better and it allows us to see

and compare the forms of the curves.

 

 

image

 

All the curves are different. What I am measuring evidently isn't just a simple piece of semiconductor physics, though all the curves have a similar

form with a straight section at the lower currents and moving away from that higher up. I'm not going to sample lots of devices, but be aware that

there will be differences between different batches from the same manufacturer and differences between parts marked with the same part number

from different manufacturers.

 

The two CB-E transistor curves stand out as being almost straight lines over the five decades of current I've done the measurements for. In practice

the relationship holds much further and it points to one of the main uses of transistors in this configuration - linear to log and log to linear conversion,

often in conjunction with an op-amp.

Parents Comment Children
  • in reply to D_Hersey

    Pease mentioned that the transdiode (CB-E) is very fast. He quotes 0.1nS for a 2N3904, which is faster than a switching diode like a 1N914. I think you are right, essentially it's the transition time working as a transistor. He also points out that it has more capacitance (so may not be so good for clamping high-speed signals - because it would distort them). That makes sense because now we have the collector capacitance attached to the input (where it's slowing the input signal, rather than slowing the output in the way that it would do if we used the transistor as a switch).

     

    The major disadvantage in a real circuit is the low reverse breakdown.

     

    He also pointed out that the leakage figure can be good for some types of transistor, but not with gold doping - it sounds like at that time some 2N3904s were gold doped and some weren't. [We need to be a little cautious with some of this because transistor fabrication techniques were changing quite rapidly at the time he was writing.]

     

    I'm currently trying to understand why it gives the nice log characteristic. It seems to be taking the ropey B-E characteristic and removing or dramatically reducing or working around the effects that take it away from the theoretical relationship between V and I for an ideal PN junction. Partly, I suppose, it's because most of the current flow is via the collector and the base contribution is minimised. [It would probably help if I had paid attention in lectures all those years ago.]

  • in reply to jc2048

    Hi Jon,

     

    You have to remember that the flow of charge through the base changes the potential difference between the collector and emitter atoms.

    By tying the two together, you have a shunt effect where the base to emitter creates a lower resistance level between the collector and emitter.

     

    That change is probably what is driving the fast response time as the current has a head start through the base at lower potential levels with the collector providing the charge flow for the higher voltage levels.

     

    Don't worry too much about what you missed in class.  My new atomic theory is providing me with the insight to better explain what goes on between atoms in all areas of physics, chemistry and biology.

     

    I have just started looking into semiconductor effects.

    I have already done a good look at conductors, non-conductors and superconductors.

     

    DAB

  • in reply to jc2048

    It seems one could fab a pretty nice diode-ring mixer from a matched quad.  IIRC Gilbert was working along these lines. . .