ldnmos bulk must connect to gnd？

[fengluan]

the bulk of ldnmos and substrate has isolated，why bulk still must need to connect gnd？

 

[erikl]

May be the P-Epi layer is continuous over the full die/wafer? See your PDK docu!

 

[dick_freebird]

That doesn't look very isolated to me.

The LDMOS body must be tied stiffly to source or it's
very likely to turn on its parasitic BJT, or fail to properly
push the depletion into the drift region instead of the
channel, punching out the gate. Either way, you lose.

 

[ZekeR]

Not an expert in processing here, but it seems if the distance f is insufficient, PBD can out-diffuse to P-Epi and create a soft-short; tying bulk to a voltage other than 0 would create a large current to substrate. If f is sufficiently large, however, I don't see any reason to require tying the MOSFET's bulk to substrate. Tying the bulk to source, yes that seems necessary. But not bulk to substrate.