Logic Design for Array-Based Circuits

by Donnamaie E. White

Copyright © 1996, 2001, 2002, 2008, 2016 Donnamaie E. White , WhitePubs Enterprises, Inc.

 

Simulation

Last Edit July 22, 2001


The Simulations

Control files vary from workstation to workstation, from simulator to simulator, from product grade to product grade, from array to array and from simulation to simulation. They are too specific to be shown here.

The outputs of the simulators are for the most part non-compact and even difficult to read. Only formatted outputs will be shown here. First, they are compact, a requirement since there are 107 I/O signals and 4 internal enable signals that the vendor requires be listed. Second, they look the same regardless of the workstation used, Dazix, Mentor or Valid. The simula-tions were run using Mentor.

All simulations shown are the result of military worst-case maximum timing. Minimum simulations are often also required.

Functional Simulation

The wafer-sort and packaged-part sort test vectors are derived from the functional simulation. (The name is vendor-specific.)

This simulation is done following vendor-specific rules. These include high-coverage, no more than 16 outputs changing in any one vector, the high limit allowed because parametric vectors and a parametric gate tree are used.

All signal transitions must be included - 0-1, 1-0 for standard macros, 0-1, 1-0, 0-Z, Z-0, 1-Z, Z-1 for 3-state macros. To prevent the vector checker from complaining, the PARAM signal from the gate tree is toggled at the end of the vector set.

There will be one error message - the initial reset will cause 64 signals to change state. This cannot be avoided. The circuit must be brought up exactly as shown, with the reset "disabled", and then the reset activated.

A vector set for a 16:1 MUX is shown in Figure 8-1. Checking of this vector set shows 100% coverage of the internal nets and primary I/O, excluding a gate tree. It passes the vector check software with the allowed exception of the reset error message.

Figure 8-1 Functional Simulation - 16:1 Mux


   MINIMAL TEST SEQUENCE FOR 16:1 MUX
   TEST CASE ON A Q20010E
   ________________________________________________________________________
   1***CIRCUIT IDENTIFICATION =
   EESSSSDDDDDDDDDDDDDDDDYP
   XXEEEEAAAAAAAAAAAAAAAAOA
   TTLLLLTTTTTTTTTTTTTTTTUR
   CRCCCC0123456789111111TA
   LSTTTT 012345PM
   KT3210 T 
   
   TIME
   99.990 010000100101100110100101
   199.990 110000100101100110100101 
   299.990 000000100101100110100101
   399.990 100000100101100110100111
   499.990 000000000101100110100111
   599.990 100000000101100110100101
   699.990 000000100101100110100101
   799.990 100000100101100110100111
   899.990 000100100101100110100111
   999.990 100100100101100110100101
   1099.990 000100100111100110100101o o o o o o o o
   


Full File Listing - Functional, Sampled Simultation 16-Bit Register with Mux Output Sample Circuit

A partial vector set for a 32-bit register similar to that in the schematics is shown in Figure 8-2. The sample step is 100 ns and the sample is taken first at 99.99 ns. The simulator output is integer - place the decimal two places from the right. The signals are listed in vendor-specified order, all inputs, all outputs, and all 3-state enables listed last. Only sampled func-tional simulations are submitted.

Figure 8-2 Functional Simulation - 32-Bit Register (partial)

Figure 8-2 Functional Simulation - 32-Bit Register - Full Listing

Exercise

Create a complete functional vector set for the schematics shown in the Appendix of Chapter 3.

 

Copyright © 1996, 2001, 2002, 2008, 2016 Donnamaie E. White , WhitePubs Enterprises, Inc.
For problems or questions on these pages, contact donnamaie@-no-spam-sbcglobal.net