A very simple Hspice question

[sgharagoz]

Hi guys,

I have got a very simple netlist that I trying to simulate in Hspice:


MN1 VCC Gat 0 0 CMOSN L=360n W=360n
VDD VCC 0 DC 1.8
Vgs Gat 0 DC 1.43

.OP 1us

.END

The model used is 0.18micron TSMC. In the model K' is defined as (µCox/2) = 177.2uA/v² and Vth =0.43v.
Base on my circuit setting, I expect the MOSFET to be in saturation region and the drain current to about 177.2uA.

But operating analysis indicates that NMOS current is about 94.3uA
---------------------------
| Element | MN1 |
|------------+-------------|
| Model Name | CMOSN |
| region | Saturation |
| id | 9.4313e-005 |
| ibs | 0.0000e+000 |
| ibd | 0.0000e+000 |
| vgs | 1.4300e+000 |
| vds | 1.8000e+000 |
| vbs | 0.0000e+000 |
| vth | 4.3606e-001 |
| vdsat | 5.5645e-001 |
| beta | 3.4911e-004 |
| gam eff | 5.5635e-001 |
| gm | 1.4045e-004 |
| gds | 2.7543e-006 |
| gmb | 3.6982e-005 |
| cdtot | 2.8498e-016 |
| cgtot | 1.3682e-015 |
| cstot | 6.8866e-016 |
| cbtot | 2.1474e-016 |
| cgs | 1.0430e-015 |
| cgd | 2.8096e-016 |
|--------------------------|

Can anyone tell me why there is difference in my hand calculations and the actual value?

Thanks

 

[erikl]

Your hand calculation probably didn't consider

• channel-length modulation (CLM)
• carrier velocity saturation (mobility reduction due to the longitudinal field)
• vertical field mobility reduction (VFMR)

 

[sgharagoz]

Hi Erikl,

I don't think adding CLM would help me. In fact if I consider that, it will make the drain current even more than I calculated.
The velocity saturation is more plausible but I have used this model card before and I am very sure that hand calculation should be very close to the actual numbers.

Any other reason that could explain this?

 

[Dominik Przyborowski]

By considering only a velocity saturation, the calculated drain current is equal to 101.7 µA. Hand calculation are very often close to reality but only if You are using a proper analytical model, not a square law long channel equation.