Embedded systems directory and blog

Verification of Mixed-Signal Systems

Benchmark Results

Environment set up
The engineering time to set the environment is considered for each benchmark in two conditions:

» Set up of the environment the first time

» Re-use of the environment to verify another DUT of the same type

	BAND GAP BLOCK (BG1)	CAN TRANSCEIVER (CAN_TR1)	CAN NODE (CAN_N1) (***)
Standard Approach	30 min	4 days	5 days
Using MSV-kit	15 min (*)	1 day (**)	2 day

Note that the evaluated set up effort is related to the verification items described in the previous section for the three cases. In particular, only one verification item is considered for the CAN transceiver and the CAN node. Based on the capability of the Specman-based verification environment, it is possible to give a first estimation about the increase in effort including other verification items as reported in the following plot.

	BG1 -> BG2	CAN_TR1 -> CAN_TR2
Standard Approach	20 min	2,5 days
Using MSV-kit	10 min (*)	<0,5 day (**)

(*) no need of Specman expertise
(**) man-time of a verification engineer with also CAN protocol expertise
(***) starting from MSV Kit, not from CAN transceiver verification

Simulations
For the band gap block, in the proposed approach based on MSV-Kit:

• The complete simulation can be automatically split in parallel run using regression feature already available in the kit itself.
• A close loop feature is implemented (selected by the user) in order to discard a defined trimmer word as soon as it fails for a certain temperature.

Number of simulations to cover the full process parameters range (NC models cards) and the temperature range (NT number of points) for a band gap with N bit trimmer word:

number of simulation = NC * NT * 2^N = 9*5*2^5 = 1440

# parallel run	1	2	3	6
Standard Approach	1440	parallel runs set up to be done on purpose
Using MSV-kit (*)	432	216	144	72

Run time based on the same hardware equipment for both approaches.

# parallel run	1	2	3	6
Standard Approach	30 min	parallel runs set up to be done on purpose
Using MSV-kit (*)	15 min	10 min	8 min	5 min

(*) close loop feature enabled; results extracted in average using three different cells

For the CAN transceiver, in the proposed approach based on MSV-Kit:

• The complete simulation can be automatically split in parallel run using regression feature
• A close loop feature is implemented (selected by the user) in order to discard a fault as soon as the transceiver fails twice in detecting/recovering it

In the following, 6 possible faults are considered with 720 possible combinations; one of them is not properly detected.

# parallel run	1	2	3	4
Standard Approach	40 min	parallel runs set up to be done on purpose
Using MSV-kit (*)	15 min	9 min	6 min	4 min

(*) close loop feature enabled

For the CAN Node the verification environment for the CAN Transceiver is completely re-used with the same test and the same conditions.

# parallel run	1	2	3	4
Standard Approach	45 min	parallel runs set up to be done on purpose
Using MSV-kit (*)	18 min	10 min	8 min	6 min

(*) close loop feature enabled

Results collection and analysis
In the proposed approach based on MSV-Kit, there is a mechanism to compare the current result with the expected one (scoreboard) and automatically correlate it with the stimuli applied to the DUT (cross-coverage). The graphical interface for coverage report available in Specman Elite allows a clear visualization of results.

BAND GAP BLOCK

Standard Approach	Number of results set	9
	Time needed to analyze a set	5 min
	Total time	45 min
Using MSV-kit	Number of results set	9
	Time needed to analyze a set	-> 0
	Total time	-> 0

CAN TRANSCEIVER

Standard Approach	Analysis time	60 min
	Coverage	manual calculation
Using MSV-kit	Analysis time	few minutes in loading data
	Coverage	100%

Previous Page | Next Page
1 | 2 | 3 | 4 | 5 | 6

If you found this page useful, bookmark and share it on:

Enter your email address to sign up for our free newsletter:   

If you are familiar with RSS feeds, you can also sign up for our free blog feed. Our RSS feed is updated in real-time while our newsletter is updated daily.