How to compensate between PMOS & NMOS?
My question is how can i estimate this in the simulation? as the DC simulation just gives me one value of current.
Secondly, if i want to avoid this mismatch, how large should the PMOS be as compared to NMOS to compensate for the lower mobility?
Looking forward for some quick answers..
Regards,
simply for frist order approximation the get the ratio between the mobilty of both from the model card of the technology
or for more accurate . jus make DC analysis for both and get the W/l rario to get the same current
khouly
When i do the DC analysis of PMOS and NMOS separately, should i use the same node voltages which i get in the VCO DC analysis or other voltages can be used to bias the devices ?
Regards
it should be the same
about the LC VCO the point between the PMOS and NMOS should be the VDD/2
this holds that both NMOS and PMOS should have the same bias point
khouly
The mid-point voltage is close to VDD/2. It is 588 mV instead of 600mV. I am attaching the pic which shows the setup and the results.
The strange thing is that the dimensions of PMOS and NMOS are exactly same (39um/65nm) and still the current is quite closely matched...any explanations ?
Isnt this misleading ?
Regards[/img]
check the mobilty in the model card
khouly
I checked the results browser,
Mobility (μo) NMOS -- 10.67 E-3 cm2/V.s
Mobility (μo) PMOS -- 5.377 E-3 cm2/V.s
So, the ratio is 1.98...however when i use the PMOS with a Width, 1.98 times the NMOS....the resulting drain currents are different...(NMOS: 1.54mA and PMOS: 2.93mA)
more ambiguity...!
wot am i doing wrong...
Regards,
yeah it is very strange
can u sned me the model card , and i will try to simulate
khouly
I checked the threshold voltages of PMOS and NMOS, and they are different in this technology...is this the reason that the drain current for both is same(with same W/L but different mobility) because the overdrive voltage is different...?
Can somebody comment ?
Thanks..
also something u need to check the mobility of both trainsistor , it should be different
if the mobility is the same , this may show how the currents are the same
khouly
As i mentioned in one of my previous posts....the mobilities are different...given again below:
Mobility (μo) NMOS -- 10.67 E-3 cm2/V.s
Mobility (μo) PMOS -- 5.377 E-3 cm2/V.s
The only reason i see is that the overdrive voltages are different and that makes the currents equal...
Regards
i see that the Nmos is twice the PMOS , is ur override voltage is the same ratio between Nmos and PMOS
khouly
Yes, the overdrive almost has the same ratio but not exactly....
Overdrive (NMOS): 92.65mV
Overdrive (PMOS): 195.44mV
Also, the overdrive ratio cannot exactly be compared to mobility because it is squared (Vgs-Vth)2 whereas mobilities are not..!
Regards
don't forget ur using the lowest channel length , this means also the channel length modulation factor will play a big role in current ratio
u can do the same setup with large L , this can show some results
khouly
I checked with different lengths of the both PMOS and NMOS and noticed that as the length increases, the difference in current is more as compared to smaller lengths. I am writing the information below. Let me know your observations about it...
W in all cases is 39um
L= 60 nm (Min length)
NMOS Id= 1.541 mA
NMOS overdrive= 92.65 mV
PMOS Id= 1.538 mA
PMOS overdrive= 195.44 mV
L= 120 nm
NMOS Id= 1.05 mA
NMOS overdrive= 118.35 mV
PMOS Id= 1.15 mA
PMOS overdrive= 235.46 mV
L= 180 nm
NMOS Id= 0.827 mA
NMOS overdrive= 139.01 mV
PMOS Id = 0.914 mA
PMOS overdrive= 266.26 mV
L= 240 nm
NMOS Id= 0.673 mA
NMOS overdrive= 152.79 mV
PMOS Id = 0.782 mA
PMOS overdrive= 280.4 mV
Regards,
don't forget that in this tech it is very short channel and the MOS devices become almost linear not a quadratic relation
and this show y the ratio is linear between the current and the override voltage
khouly