3
\$\begingroup\$

Gray, Hurst, Lewis, and Meyer give the following discussion around PMOS transistors which I followed except for the sentence which reads:

Good use can be made of this fact in analog circuits to alleviate the impact of the high body effect in these devices.

Now I understand that (assuming a single n-well process) having access to the body terminal in the case of PMOS built in n-wells lets us play games with \$V_{SB}\$ which we cannot play with \$V_{SB}\$ for the NMOS which are built right into the substrate. However, I am confused by an allusion specifically to a "high body effect in these [PMOS] devices." I am reading this as the suggestion that, ceteris paribus, PMOS suffer a worse body effect and, therefore, we use an n-well process so that we can control this more drastic body effect, rather than p-well in n-substrate so that we could control the bodies of our NMOS. Is this a correct reading? If so, why do PMOS indeed suffer a worse body effect than NMOS, ceteris paribus? There's nothing in the body parameter \$\gamma\$ which makes this obvious to me.

enter image description here

Source: Analysis and Design of Analog Integrated Circuits by Gray, Hurst, Lewis, and Meyer - 5th Edition, Wiley, 2009, page 144

\$\endgroup\$
0

1 Answer 1

-1
\$\begingroup\$

One suspicion I have is that the "high[er] body effect" of pMOS (above nMOS) comes from the higher doping of the body (since the doping of the n-well must overwhelm that of the underlying wsubstrate) in an n-well as contrasted with the body (the substrate) of an nMOS built in this single well process. Since the body effect parameter \$\gamma\$ goes as the square root of the body doping, we obtain the conclusion from GHLM.

The reason I am not entirely convinced by this, however, is that we generally have a threshold implant in order to precisely adjust the doping. Given that, it's not clear that there are actually more net dopants in the body for pMOS than nMOS.

In the end, though, I suspect that the phrase "high body effect" is not meant to pit pMOS versus nMOS but is rather just a comment that at the relatively high doping levels used for the body in either case, one has a significant body effect.

\$\endgroup\$

Not the answer you're looking for? Browse other questions tagged or ask your own question.