# Threshold voltage data analysis

I want to determine the threshold voltage of a diode from the graph of iv characteristic diode. From the point of view of the data analysis I do not know how to do to find a specific value. I have read the advice of graphing the logarithm of the relationship between the current passing through the diode and the reverse saturation current. I get a straight line from which I do not know what parameter could be useful to me. I'm stuck and I do not know what to do.

• What exactly is your definition of "threshold voltage" for a diode? Is isn't a commonly used term. – Dave Tweed Oct 10 '17 at 13:23
• The voltage drop, when the diode starts to conduct. – Stefano Barone Oct 10 '17 at 13:44
• As you already show, the diode is conducting all the time. – Dave Tweed Oct 10 '17 at 14:05
• What exactly are you trying to accomplish here? Why is having an "exact" value for "threshold voltage" important? Why can't you just measure the voltage at some fixed value of current, selected for your specific application? – Dave Tweed Oct 10 '17 at 14:16
• Why such huge error bars on your voltage measurements? – The Photon Oct 10 '17 at 15:50

The threshold voltage is the minimum voltage needed to create a conducting path along the diode. As a real diode is never completely off, one can define the threshold voltage at the point where the diode starts conducting in the exponential fashion. So there are different ways to define the actual threshold. (It is also heavily dependent on temperature.)

Here is one way to get a value for the threshold voltage:

If you increase the voltage applied to the diode and measure the current flowing along the diode, you can use this data to determine the threshold voltage by plotting the sqrt of the current. I did this some time ago from simulation data for a nmos device, so actually this curve represents the diode characteristic, in which you are interested in.

In the Appendix of the book CMOS Analog Circuit Design from P. E. Allen and D. R. Holberg you find this procedure in more detail.

There is no real "threshold" because the diode starts conducting already for voltages far below the well-known magic value of 0.7 volts (as you can see in the first diagram). According to the most common specification, you could draw a tangent to the linear part of the characteristic. This line (tangent) will cross the horizontal axis at a value (app. 0.7 volts) which very often is called "threshold".

• Thanks for the answer.I am aware of the approximation of the conduction. I have already used your method but it is a bit rough because actually a current incremen with voltage there is even before 0.6.Is there a method that emphasizes the change in behavior? Maybe using the first graph.. – Stefano Barone Oct 10 '17 at 13:52
• I rather think (and I have done this) when you draw the tangent to the upper part of the curve (where we have quasi-linearity), you arrive at app. 0.7 volts. – LvW Oct 10 '17 at 14:05
• @StefanoBarone, Read his first sentence. The idea of a "threshold voltage" for a diode is inherently a very rough approximation to the diode behavior. If you want a more accurate model of the diode, read about the Shockley diode equation and be prepared to do your modeling numerically (in SPICE). – The Photon Oct 10 '17 at 15:52