EE 421L Digital Integrated Circuit Design - Lab 5

Design, layout, and simulation of a CMOS inverter

 

Shadden Abdalla

Abdals1@unlv.nevada.edu

September 20, 2019

Lab5.zip

 

PRELAB:

Finish tutorial 3.

 

Above marks the end of tutorial 3.

 

About this lab

 

This lab has 3 parts:

 

Part 1. Draft two schematics of two different inverters, a 12u/6u and a 48u/24u inverter and the create symbols for each.

 

Part 2. Create layout for both of those inverters, DRC, LVS them versus the schematics created earlier.

 

Part 3. Create schematics using the previously made inverter symbols and simulate them over four different capacitor values.

 

Part 1: Inverter Schematics and Symbols

https://lh5.googleusercontent.com/jtIMmKPFizgQwYAF7L9COI2HKDeNoVy_kGP-OFFPZs_ZAnG5KSfCwnl714axt9awZEn27R5L7X8EOqu4giKyMaH0vE98VScILEJZ61zRazdC17T_vr2JuRfLWI84oTJ3MGQtmANxhttps://lh6.googleusercontent.com/HlU4EEINU1EZ7tirrsm56MI61SvNeGbTyiMQO_ppU39Xrl8JRPhwjwEIJPWEi1wHh7Z-dnnueBTB6ny8Ht4aJdx0O1YVSXYak9gmJJ8yyOLGV2VrjS31PpYf73hCyRrFawxtIpXq

The left schematic is of an inverter using pmos4 and nmos 4, each as a multiple of one. The right schematic is of an inverter also using the same components but instead of just a multiple of one they are of multiple four.

 

 

This is the symbol view that I created for inverter of multiple one (m=1).

https://lh5.googleusercontent.com/kzTzrDWnoRz--mtoGQu4iOH4sxWLlUGwuToQ3b1rLUW1cQ0CuOg9PmY216O6vbOGf4ujGjRXEwfbdsa0WobRyckhcuuOu7REmBbrmzxls_APEk-8Kzf2dBXWmOqfplVj-J3kLT6B

Below is the symbol view I created for inverter of multiple four (m=4).

 

https://lh3.googleusercontent.com/_t5oSffDCS0uWPufvh8DkLwYJE0Y3kNs_8LVW2QOEnHhcDR957BKXCOHghQDOPtqN-6iEaN7BuQCA1J5Aw5Mzkxic-7kG2YtXTqIKBg7pOVb-WQZrAKeKP4zmdPJDIlgFRaHaZeX

 

Part 2: Inverter Layout and LVS

 

 

Below are the layout and the extracted view of the layout for the inverter of multiple one.

The layout shows the PMOS connected to an ntap with a vdd! Pin using metal1. Then the nmos on the bottom is connected to a ptap with a gnd! Pin using metal1. The input A that goes into the gate is connected to poly which is what connects to the gates. The output pin is connected to the Drains of the NMOS and PMOS.

https://lh4.googleusercontent.com/3TguWZUX4aw4PLUwjher9xYUBW4nWuT7PYcXJgnfcFdLjBS04dnvQKmlganlW95LirKkwwJXo4M-UBupT9G-QHN8-5tujsi6UDNGGKcXcdU5SDFDKFd-MSOeM8rVd9L34K0OfzdW  https://lh6.googleusercontent.com/rXdpVHfgA3jVj6eiRyM9PJi2bKu23PAvz3ci4X5Mf-WoWiBoLmZkRzfKNTblm9DrGzq0Dj2wE549IwAftRYvxPTFeRZ3mdX8yTxU2dQeSg2SFDK3yVssnMFDYjfGc9nvLoEPCYNB

 

I DRC ed the layout  and it had no errors.

 

 

 

I LVS’d the schematic view of the inverter of multiple one versus the extracted view of the layout and saw that the net-lists matched, showing that my design was accurate.

https://lh6.googleusercontent.com/NqWYwx4TPKWCmtV8WToqI9llYq5camATYc0CF1VNwQvUdCzzyp7ZI7Y2Bz2cotu4U8rgPPYN2VHhT1WD3spsBkOrbv5Pg_LalTSHt6OJpAW8LV4XsC8Zn5wHOcLJS3jiQvR8zdW7https://lh3.googleusercontent.com/6eGdtCH3uxzcyCx_CToI9UMYOjnlXYqgHMS3qhwPdyVKqFJBkGoPyhAzdpvByLcbkMM6QD8XP9wE_LlJ9VsJbw-U5nYUzP9ly3HCOcbE3sau8AlWwszsg9bpoNcIt6ZOFqDK2A0ahttps://lh4.googleusercontent.com/URUsmVJzMqjp_PzeFiELpv4fp_ii7BTdtI0vT6S6FF6_zQQprjWNLocP-HCJiFem0xXyq9HJ2zpQFa-e4pRiDnqEYuD8OBbS9L5ZpY7DYlk3tyV7qSFDLBWibcx5p3lt_9jIVEdL



Below is the layout and extracted view of the inverter of multiple four. I connected the PMOS to the ntap with a vdd! Metal1 pin and the NMOS to the ptap with a gnd! Metal1 pin. I also connected the input A to a m1_poly since its connected to the gate. The output pin is connected in between two terminals of the source and drain.

https://lh6.googleusercontent.com/-twi3tWBmiu5RrDLnfFz1b0ap7TiutYlzXtKz8ES-newsN_fitxt8n-uXsDIydIqzSva7ghYfeT3D4DljEJXcPVktXrLWVhjjWjzx80X6-jVBwbedZBWLJ6qd3GoSlodB8Yb7AeVhttps://lh4.googleusercontent.com/NfVm8juKrXKX_vypHQ141Pb3oX5ZRtvDjEPY-swRLUhqmBhykngIu55zUxe863Y6DL7I1I0jyIC2fiVsUxgUbUgb0kMcC2vIuQJO9zAIObZiKppMtop2-lMH_6zPl640OmxBwIAo

 

I DRCed the layout and it had no errors.

 

 

I then performed an LVS of the schematic view of the inverter with m=4 versus the extracted view of the layout and saw that they corresponded with matching nets.


https://lh3.googleusercontent.com/xPajvCed2eh35zd6UW1es83K4kBzOwkNuS5ypZUUI7oiPW2kJ_xBatoUgQzgx0CQMXEsqnuEE-v7-Xp-8A6Pmw7eCZz05y8uKH1uRpnhXNKEPyPjWUWMoF-vRnxlu6HPXhg5U3bL  https://lh4.googleusercontent.com/25NdaoaPwgkCcA36928rt-wlHt42Dhgd7Qq50JMao6MOAWt2RRiqpJUxTfeiQMaLlGHEofdYLMgQy_r3BLQ0Cz5q_2Wbzd3pKN6TTcjd0CANnhDHerjzHJjh1FzMQ9AEZYULdlts

 

 

Part 3: Simulations, Spectre and UltraSim

 

Simulation Methodology:

I used a variable for the capacitor size to maximize efficiency and save time. I simulated the schematic using parametric analysis of a decade and a step size of 1, driving loads of 100fF, 1pF, 10pF, and 100pF. Using a decade analysis with a step size of 1 simply measured increments of a decade starting from 100fF and resulted in measurements of all four capacitor values on the same graph. I created the variable, used the model libraries, choose the nets on the schematic and ran the parametric analysis.

 

I used the same technique for the UltraSim measurement of this inverter, as well as the other two simulations using the 48u/24u inverter. Ultra sim is used for simulating large circuits since spectre is very slow when simulating large circuits. Ultrasim simulations are usually not as accurate as the ones in spectre. Here there is not much a difference since the simulation is not very large.

 

*******************

For the 12u/6u inverter:

https://lh6.googleusercontent.com/V_BjBgM9GgNxrCxvUV4XpTMRWTDLe9noF3qFwYbFTd2ySgAtalB8dn4dEP48zB914kfaZxRSmDiaCD07eq30cuKcvkIYve2tFmv7ppEjruFYZGgUVFew11mv-9hanr9PUdXV4fBe

 

Below is the regular simulation using spectre_state. It shows net1 as the input which his the black line and all of the different outputs that come with the pulsing voltage source.

https://lh6.googleusercontent.com/sBRT8kH830ze_YgshN7r0DnOUA5eOG5g4xP477Zv3xEZbvd5d-ek5fCTO0SCRQ8x8mpXqfKotVhHhEoduKwffdqEvsvBoEgaKuIE5cLLemhFekNCUrQTwdUaCnHiXRtYP94Mwi_J

 

 

Below is the UltraSim simulation. Both are generally the same, however the UltraSim simulation looks a little cleaner.

https://lh3.googleusercontent.com/zu8bGm3Ftdnb2-OJEMIaqjerhDGBJBBxz0QRKqiGeiGUFVMY6KQOPbQoSum-HQ3e9ee_77eKKDcw7UcmrAuXeRRMBzx9rqQto39QDUkLSO4icypjyRow5bveZIV9vOPNYNq1Y4rZ

 

For the 48u/24u inverter:

 

https://lh3.googleusercontent.com/K1DF5oj3o3JZ7AFlnKMBHMGXy8mU2Q3aglWBfivAYyh_COGoGIp5hqThoPJgteVUXHutiZ8UeCMu4WkzDsbu94I_qFsETRwBAoAypD8Bvrl-igoSkbQppCFU0hWNXUpeOdN5VZl0

Below is the simulation using spectre_state and the same parametric analysis described earlier.

The results are the same as seen above. Net 1 is the input and net 2 is the output.

https://lh4.googleusercontent.com/AJ3FJvhU5MXc3woZuR85m6aptPkSKZ40JMwtzbB6-NI18saQgGfR6bZb6CbT_F8dXuW2NHfcuIdrq_jqlBvgTKKsN3QCAQ95hZjja_yI-HqSV3VwE3Ft4xCt_K8AhT7tbt38kbpm

 

Below is the same schematic simulated using UltraSim.

https://lh3.googleusercontent.com/4jvrYewNGK5cLNLCRMp7-6R6yFzV9L1BJvVgRVioA1b-PXnWtuT7oeboxNpcUPeAzBsxgzqRO7fsXCTL5Altsed8PO__H5VlEuPCKU3p1cx-10BdbY5b9nUNpkjgXmLauYfGD5cD

 

I zipped up my Cadence files and put them in my Google drive.

 

Return to Shadden’s labs