Question of a test circuit setup in temperate sweep for subthreshold operating MOS

Hi all, I am trying to use a simple circuit to probe the VGS behavior of a subthreshold operating MOSFET in temperature domain to validate the subthreshold models.

I am trying to use either floating gate or gate-source connected MOSFETs to operate in subthreshold region in the entire temperature range (say 0C to 100C). I want to use the simulated VGS(T) waveform to compare with the model provided in page 10 of https://www.cppsim.com/CircuitLectures/Lecture16.pdf.

This is my setup and the equation in the pdf file above:


I want to use a temperature-dependent bias current ID (for example, PTAT) to probe VGS behavior in the interested temperature range, by making sure the effect of VDS in this equation can be ignored. However, I found that in using the floating gate test circuit, VDS can vary a lot and in high temperature range, VDS may get closer to VT. If I use the gate-drain connected structure, the VDS always equals to VGS, and thus the VDS>>VT condition cannot be met.

Would anyone give me the suggestion how to correctly setup this test (change the connection, adding the resistor or anything else), in achieving (1) subthreshold operating MOSFETs, by using the drain current to control VGS (2) the drain current can be defined by the user, to be either temperature independent or dependent (3) the impact to VGS from VDS is minimized.

Thank you!

You could use feedback. Drive the gate through an ideal VCVS with a high gain. The VCVS compares the VDS of your device and the VDS which you want to set.

deba_fire said:

You could use feedback. Drive the gate through an ideal VCVS with a high gain. The VCVS compares the VDS of your device and the VDS which you want to set.

Click to expand...

Hi deba_fire, thank you for your advise! I tried the following circuit but cannot make VGS working in a proper range. For example, I expect Vgs to vary between 0.2V to 0.05V in the temperature range (VGS needs to be smaller than Vth), but what I got is a very high Vgs in the target temperature range. Do you know what changes I should make to let it work? Thank you!

You just need to reduce the current for a given W/L to enter subthreshold.

deba_fire said:

You just need to reduce the current for a given W/L to enter subthreshold.

Click to expand...

Hi deba_fire, thank you for your help. What I want exactly is to use the sim data to proof the expression between Id and Vgs in page 10 of https://www.cppsim.com/CircuitLectures/Lecture16.pdf in different temperatures. However, I am not able to get matched data. Do you know how can I find out the equation the simulator/PDF is using in relating Id and Vgs? I am using Cadence virtuoso.