Bizarre Behavior in Discrete PMOS

Click For Summary
SUMMARY

The discussion centers on the unexpected behavior of the Id vs Vsd curve for the PMOS transistor in the CD4007 MOSFET IC, which shows a delay compared to the NMOS curve. The NMOS transistor operates on electron mobility, while the PMOS relies on hole mobility, resulting in differing conductivity and curve slopes. The PMOS FET must be twice the size of the NMOS in CMOS circuits to achieve symmetry. The presence of a small resistor in series with the NMOS drain affects the measurement, leading to confusion regarding the comparison of the Id vs Vsd and Id vs Vsg plots.

PREREQUISITES
  • Understanding of MOSFET operation principles, specifically CD4007 IC.
  • Knowledge of electron and hole mobility in semiconductor physics.
  • Familiarity with CMOS circuit design and sizing of PMOS and NMOS transistors.
  • Basic grasp of electrical measurements and circuit components, such as resistors.
NEXT STEPS
  • Study the characteristics of the CD4007 MOSFET IC in detail.
  • Learn about the impact of mobility differences on transistor performance.
  • Research the principles of CMOS circuit design and the significance of transistor sizing.
  • Examine the effects of series resistors on current measurements in MOSFET circuits.
USEFUL FOR

Electrical engineers, semiconductor physicists, and students studying MOSFET behavior and CMOS circuit design will benefit from this discussion.

Apogee
Messages
45
Reaction score
1
Some of my colleagues and I captured the i_D vs V_ds/V_sd curves for the CD4007 MOSFET IC (http://www.ti.com/lit/ds/symlink/cd4007ub.pdf). We did this for the NMOS and PMOS transistors. I have attached the curves to this post. The NMOS curve is as expected. However, the PMOS curve seems to be delayed by a few volts. Does anyone happen to know why the PMOS curve would have such a delay?

NMOS curve: https://raw.githubusercontent.com/Roman-Parise/Circuits-Lab/master/Spring2017/Lab2/data/nmos.png

PMOS curve: https://raw.githubusercontent.com/Roman-Parise/Circuits-Lab/master/Spring2017/Lab2/data/pmos.png
 
Engineering news on Phys.org
The NMOS transistor operation is based on the mobility of Electrons in the Source-Drain channel.
The PMOS transistor operation is based on the mobility of Holes in the Source-Drain channel.

The mobility of Electrons is much higher than that of Holes. That translates to higher conductivity as shown in the differing slopes of the graphs you supplied.
If you want to get into the details, look at the formulae on pg 202, and the graph on pg 203 (PDF pgs 8, 9) in:
https://people.eecs.berkeley.edu/~hu/Chenming-Hu_ch6.pdf

Hope it helps.
Tom
 
  • Like
Likes   Reactions: CWatters
+1

As I recall in CMOS circuits to achieve symmetry the PMOS FET has to be twice the size of the NMOS.
 
I understand these concepts, but I'm more concerned with the fact that the Id vs Vsd curve remains zero until a certain point. For Id vs Vsg, this makes sense, since there's a threshold voltage. But this doesn't make sense to me for Id vs Vsd.
 
For PMOS you show I(mA) against Vsd.
For the NMOS you show Vr(mv) vagainst Vds.
How can we compare different plots?

It may be that since devices are operated with their P and N gates tied, there needs to be sufficient Vgs to prevent heavy conduction by both P and N at the same time.
 
Baluncore said:
For PMOS you show I(mA) against Vsd.
For the NMOS you show Vr(mv) vagainst Vds.
How can we compare different plots?

It may be that since devices are operated with their P and N gates tied, there needs to be sufficient Vgs to prevent heavy conduction by both P and N at the same time.

I apologize for the confusion. A small resistor was placed in series with drain (about 100 οhms) when measuring the NMOSs drain current. Vr is the voltage drop over that resistor. So id is essentially proportional to that curve.
 
Apogee said:
A small resistor was placed in series with drain (about 100 οhms) when measuring the NMOSs drain current. Vr is the voltage drop over that resistor. So id is essentially proportional to that curve.

Might not be related to your question but that means the current range for the NMOS FET was roughly 0 to 0.9mA and for the PMOS it was 0 to 350mA?
 
Apogee said:
I'm more concerned with the fact that the Id vs Vsd curve remains zero until a certain point.

Almost as if there is a diode in the path somewhere.