NX24X7: Challenges and Considerations when Designing a Life Cycle Tester

Katya Prince

Prince Consulting, LLC

Durham, NC

katya@princeconsultingllc.com

July, 2024

​

Introduction

• Definition:
• System that repeatedly cycles a product over an extended period of time often to device failure.
• Purpose:
• R&D and product development
• Industry Qualification
• Design Consideration:
• 24-7 Operation
• Multi-DUT Testers (Parallel or Batch)

System Design Assumption

• LabVIEW Application
• PC with Windows
• Communication with NI DAQ, cRIO or Stand-alone equipment

Example Hardware

• DUT:
• Thermoelectric (Peltier) Device
• Apply voltage/current and will create temperature differential across the device
• Used for heating or cooling applications

​

Example: R&D or Product Development

• When Used
• New product development
• Design change qualification
• Goals
• Parallel Execution
• Flexibility to start-stop each DUT
• Flexibility to change testing parameters for each DUT
• Detailed information on each DUT over time

​

Example: R&D Hardware Layout

DUT (i)

TC 1 (i)

TC 2 (i)

DAQ (shared)

Power Supply (i)

Volt Meter (i)

Temp Controlled Base Plate (i)

PID Temp Control (i)

Chiller (shared)

Example: Qualification

• When Used
• Devices are ready for production, but may need to pass industry (e.g. automotive or aerospace) qualifications for multiple parts.
• Goals
• Lot testing for identical devices.
• Statistical data collection on the entire batch of DUTs.

​

Example: Qualification Hardware Layout

DUT (i)

TC 1

TC 2

DAQ

Power Supply

Volt Meter

Temp Controlled Base Plate

PID Temp Control

Chiller

DUT (i)

Bypass

Bypass

Relay Controller

…

Design consideration 24X7

• Must run with minimal user intervention
• Will be active the majority of the time
• Example of other 24-7 apps
• Production Line Testing
• Process Control

24X7: Build for WHEN not IF (pt1)

• Power Interruption or Loss
• Mitigate with UPS on PC and AutoPause/Restart
• Unsafe/critical situations
• Over Temperature
• Air Pressure or Water Pressure Loss
• External Equipment Failure
• Mitigate with Alarming – Pause Testing with user intervention to restart

​

24X7: Build for WHEN not IF (pt2)

• Communication Loss with Equipment
• Mitigate with automatic retries
• Mitigate with AutoPause/Restart
• Network Communication Loss or Server Reboot (for Files or Dbase)
• Mitigate with Local Storage and Retries
• PC or Application Crash/Reboot. Hard Drive Failure.
• User intervention – consider system action logging.
• Create reinstall files and installation instructions.
• Use cRIO for mission critical to safely stop test if PC loses coms.
• Watchdog timers and Master Control Relays.

​

24X7: Hardware Communication Tips

• AVOID USB
• Build in retries.
• Some equipment will go into a sleep mode and may need to wake up before responding correctly
• When changing settings, read back values to verify actions
• This may be a bigger deal with serial drivers that do not reply

​

​

24X7: Memory Management Tips (pt1)

• Try to avoid building/resizing array.
• Tip for array update using inplaceness:

​

​

​

​

​

• Truncate status display strings

​

​

​

24X7: Memory Management Tips (pt2)

• Close References

https://www.ni.com/en/support/documentation/supplemental/13/closing-references-in-labview.html

• Tools for tracking issues
• Tools-> Profile

​

​

24X7: File/Data Tips

• Save as you go. Do not wait for testing to be stopped
• Break up files based on size or time
• If networking issues, may want to save local and move to remote
• During design calculate how much data you will be creating
• May want to separate file into year or month folders
• May need to reduce data collection either by decreasing scan rate or increasing ‘detailed’ data collection interval
• Consider database size
• Don’t keep all the data in memory
• Collect, summarize, log and toss the raw data.

24X7: PC Settings

• Disable Auto Updates and Restarts
• Disable File Indexing – especially if doing a lot of file IO
• Disable automatic defrag

​

​

​

24X7: LabVIEW Licenses

• ALWAYS use EXE
• ALWAYS use deployment licenses (e.g., Vision, TestStand, OPC)

​

​

24X7: Operation Tips

• Build in Pause/Resume
• Update the UI with Current status –
• for long delays, update a timer
• No ms wait longer than 1 or 2 seconds – long delays decrease application responsiveness => Frustrates user
• Update Test Results as you go
• Periodically save status to file/database

N: Introduction

• In R&D, each DUT are tested in parallel
• In Qualification, DUTs are batch tested.

​

N: Qualification Batch

UI Ctl

TC Reader

Chiller Ctl

Database

Handler

Main Level

Email Sender

Error Log

Test Step Engine

PS

Volt Meter

Logging

Relay Control (DUT 1-N)

UI Ctl

Database

Handler

Email Sender

Error Log

Test Step Engine

​

​

​

​

​

​

​

​

​

​

​

​

PS

Volt Meter

Logging

Relay Control (DUT 1-N)

TC Reader

Chiller Ctl

Main Level

N: R&D Parallel DUT Tester

Main UI Ctl

TC Reader

Chiller Ctl

Database

Handler

Main Level

Email Sender

Dynamic Launch N

Error Log

N: Determine Shared vs Individual Resources

• Shared Devices
• 32xTC DAQ board
• Chiller
• Other possible shared Resources
• Communication bus – e.g. GPIB, RS485
• Error Log Files
• System Log Files
• Configuration Files

​

​

N: Re-entrancy (pt1)

• Definition
• Use Re-entrancy and dynamic launch to create duplicate ‘Test Stages’

​

N: Re-entrancy – Example Launcher

N: Re-entrancy (pt2)

• VIs should be Reentrant ‘all the way down’
• Non-reentrant SubVIs could block action of reentrant caller
• Most device drivers are NOT reentrant. You may need to re-write them
• VI Analyzer has a reentrancy test
• May not want reentrancy in things like Test Start Window

​

Nx24x7: SW Packages

• DQMH
• JKI State Machine
• Actor Framework
• TestStand (maybe?)
• Roll your own

Nx24x7: UI Suggestions

• Allow for summary information and then detailed information
• Use ‘pause all’/’resume all’/’stop all’
• Build with extension in mind. Don’t let the software be the limiting factor.

​

​

Example UI

NX24X7: Troubleshooting

• Behavior: Application continues to grow
• Possible Reasons: Unbounded array or string. Unclosed references
• Behavior: UI updates and data collection gets slower as more stages are activated.
• Possible Reasons: Something is non-reentrant.

Summary

Life cycle testing is needed by industry, and knowing the core considerations will help you address the myriad of challenges when building such a system.