Mosfet Threshold Voltage

from MOS theory we know that, in case of a P substrate, a Gate-Bulk voltage higher than a certain threshold value creates an inversion layer, in this case made of negative charges. This is also described in wikipedia (https://en.m.wikipedia.org/wiki/MOSFET).

But we always consider as threshold voltage that between Gate and Source, and this seems in contrast with the previous description. The answer cannot be (as someone told me) "Source and Bulk terminals are often connected together", there is a deeper analysis.

For instance, let's consider a pass transistor circuit made with a N-channel MOSFET (enter link description here).

The input signal is sent on the drain and the output signal is taken at the Source. Its maximum value is VDD - Vthreshold since from that value, the Vgs voltage would be lower than Vthreshold, as we know from theory. From this analysis we understand that the voltage that switches on a NMOSFET is that between Gate and Source. And here the Source is not connected to Bulk, which is at GND. This does not agree with MOS theory.

So, which is the voltage that switches on a NMOSFET?

You should know that the electrons for the inversion layer, formed under the gate, are coming from the source terminal of the MOSFET consequently, it is the $$\V_{GS}\$$ that turns an n-MOS on. But the potential barrier between the source and the bulk depends on the bulk potential. Thus the threshold voltage itself depends on it (bulk potential), known as body effect.

• But why in Metal-Oxide-Semiconductor theory is it written that the inversion layer in the substrate is due to Gate-bulk voltage? Jul 28 '19 at 7:41
• Moreover, why do electrons come from source and not from drain? Jul 28 '19 at 7:42
• @Kinka-Byo Can you point any source which says inversion layer is formed due to Gate-bulk voltage? Jul 28 '19 at 9:00
• Yes, for instance the wikipedia link I inserted in the question, that is this (en.m.wikipedia.org/wiki/MOSFET), in the section "Operation". Jul 28 '19 at 10:12
• @Kinka-Byo That section is for MOS-Capacitor where there is no source (and drain) so indeed the electrons need to come from bulk. But the operation is different for MOS-FET. Jul 28 '19 at 10:57

Some FlipFlops use PassGates to switch between Feedback (store a bit) and Update (change the bit).

The PassGates look like this

simulate this circuit – Schematic created using CircuitLab

Notice the FETs are 4-terminal. As the signal varies, the gate-channel voltage varies, and the channel resistance varies, and the FF setup/hold times will vary.

The input signal is sent on the drain and the output signal is taken at the Source. Its maximum value is VDD - Vthreshold since from that value, the Vgs voltage would be lower than Vthreshold, as we know from theory. From this analysis we understand that the voltage that switches on a NMOSFET is that between Gate and Source. And here the Source is not connected to Bulk, which is at GND. This does not agree with MOS theory. So, which is the voltage that switches on a NMOSFET?

I guess you are forgetting or no one told you, that mosfets are in a state of conduction. There are two types enhancement mode (at 0 conduction) and depletion mode (at 100% conduction). The gate increases current on enhancement mode types (turning on), while the gate decreases current on the depletion (turning off). Gate voltage bias operation depends on the gate channel type and operating mode it is made to operate. For an N-type , enhancement-mode devices have positive thresholds, and depletion-mode devices have negative thresholds; for a P-type , enhancement-mode negative, depletion-mode positive.

The gate state is the gate state and if your example used N channel depletion mode mosfets, the gate would sit there floating just above threshold and allow current to flow till the gate is grounded or a negative dc voltage applied.

• I think the OP understands this just fine. They stated that "voltage higher than a certain threshold value creates an inversion layer" which clearly indicates that we are talking about an enhancement MOSFET. On the other hand, what does "the gate state is the gate state" mean? It's not the gate that determines whether a device is enhancement or depletion, it's the doping of the silicon underneath. Mar 28 at 1:02
• Gate has no bearing on this. As what I think he thinks all of them behave like enhancement mosfets, but to even confuse his question even more, the example didn't state that either. I just think the OP doesn't know that there is "always on" mosfets as well as "always off" types. [Edited by a moderator] Mar 28 at 1:15