is there any problem if I connect S,G,D of dummy PMOS to GND?

Status
Not open for further replies.

chaojixin

Member level 1
Joined
Jul 8, 2011
Messages
38
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,286
Activity points
1,529
Dear all:
It seem ridiculous but I can't tell where is wrong...what effect does it have if I do so?
 

Hi,
In theory it may not have any problem, but in reality you are creating extra parasitic RC in series.
Your N-well is acts as high resistance in series with active capacitor. (very prone to substrate noise). That is why, this kind of dummy's are avoided where critical signals are routed.
 
... in reality you are creating extra parasitic RC in series. Your N-well is acts as high resistance in series with active capacitor.
I'd think the additional parasitic cap due to the grounded PMOS is small/negligible compared to the n-well-substrate parasitic cap. So it shouldn't change a lot.
 

... such dummys are avoided with critical low noise signal path. What may be the reason?
Where did you come to know that? No reason given? I can't imagine any. I'd think more capacitance between VDD & GND should help decoupling noise?
 

Where did you come to know that? No reason given? I can't imagine any. I'd think more capacitance between VDD & GND should help decoupling noise?
sorry,I may not give reference for this...but, still, i'll tell, PMOS dummy's never used in the lower rail. ( here S,D are connected to GND)( capacitive benifit mentioned may not available here).But it will be vulnerable to breakdown and leakage current may increase. You may find this in any of layout designed in <180nm.
 

thanks to all of you!
As S,D and G are all connected to gnd, the dummy PMOS is worked as a decouple capacitor and will not affect the signal path, and the only disadvantage is the leakage current, am I right?
 

thanks to all of you!
As S,D and G are all connected to gnd, the dummy PMOS is worked as a decouple capacitor and will not affect the signal path, and the only disadvantage is the leakage current, am I right?

Hi Chao,
One thing need to understand that, dummys should not be used in active state, (dummy PMOS gate should not be at lower potential and Dummy NMOS gate in higher potential.)
Next PMOS dummys are used in the upper rail voltage ( D, S are shorted to VDD) while NMOS dummys are used in the lower rail. This is the best way to use dummys.
 

... dummys should not be used in active state
This dummy PMOS isn't in an active state: Vgs=0 !

(dummy PMOS gate should not be at lower potential and Dummy NMOS gate in higher potential.)
... and what about source followers?

PMOS dummys are used in the upper rail voltage ( D, S are shorted to VDD) while NMOS dummys are used in the lower rail. This is the best way to use dummys.
Right; but Chao's application is quite ok -- apart from a possible higher leakage current -- as stated by Chao above.
 

... and what about source followers?

Source followers are different case, (where Source of the dummy is connected to signal level while others follow the same idea.)
Aprt from all that , You don't see , where actually we place PMOS dummy, and what is the benefit for using GND.
PMOS dummy may be used beside PMOS active transitors, ( and I don't see much application where you need to connect PMOS directly to ground.)
 

Source followers are different case
Yes, you're right, those aren't dummies, sorry!

I don't see much application where you need to connect PMOS directly to ground.
E.g. a PMOS cap in inversion state -- a very effective (high cap/area) capacitor for rail decoupling & LF filter applications, s. below. Shouldn't be labeled as dummy, either.
 
The overwhelming reason why dummy PMOS is tied to supply rather than gnd is simply because the routing would be more convenient from the well tapping of the n-well. There is no convincing reasons that tieing it to gnd will give any issues.
 
Status
Not open for further replies.
Cookies are required to use this site. You must accept them to continue using the site. Learn more…