HALT 2: Preparation is Everything



An experienced embedded solutions manufacturer shares time- and money-saving information

In the first article in this series it was determined what Highly Accelerated Life Test (HALT) was, and what it wasn’t. This section will cover tips, trips and lessons learned as Technologic Systems prepared for and underwent HALT.

Time is money. When scheduling HALT, typically the lab is secured on a per day basis. Being well prepared beforehand will save time, frustration, and money upon arrival. HALT is not quick, so expect to spend a minimum of three to four days at the lab, if observing testing.

Figure 1: (Left) Colored CAT5 cables for quick visual identification once in the chamber. (Right) Technologic Systems’ Jeff Palmer verifies connectivity.

Figure 1: (Left) Colored CAT5 cables for quick visual identification once in the chamber. (Right) Technologic Systems’ Jeff Palmer verifies connectivity.

Set Up
The most important thing to set up and prepare will be the systems that are being evaluated by the HALT testing. Consider taking multiple products to the lab together, maximizing the lab time across multiple different systems. Make sure that you prepare the latest revisions of the hardware for testing, including ensuring the appropriate software is loaded. Appropriate software may be the product’s standard software installation, or else it could include product testing functions, data logging for collecting test results, and other special installations specific to your HALT needs. Set up and run the full test configuration in the office before packing up to head to the HALT lab, and while packing up make sure you include everything that was in contact with your test configuration. It will be much easier to identify and source missing components or update software at the office than at the lab. Setting up the test configuration prior to leaving the office will also identify potential bad support equipment.

Bring backups of everything. Plan on bringing multiple units of each system, and instead of individual units being allocated for each test run, use multiple copies of the same unit through all of the testing. If there is a manufacturing error or other defect, it will be more apparent if the same failure is observed on multiple boards, or perhaps just one board performs poorly and the others have no issues. The more data collected, the better.

Identify the Systems
Make sure the hardware can be identified once it is in the chamber. The chambers typically have viewing portals, but a quick visual indicator is very helpful when the device is being shaken at 35g. Identification of a system can be as simple as different colored Ethernet or other communication cables attached to each device, so it is readily apparent which device is positioned where in the chamber, as well as which port it is attached to on the outside. Power supplies should be well marked, so if need be, just one device can be power cycled. Have a unique identifier for each unit. We used the last four characters of the MAC address. Equip each unit with a label that is clearly visible from outside the chamber. Also consider equipping the cables attached to the unit with the same label as they come out of the chamber. These identifications help in the documentation stage and essentially provide a nickname to refer to when taking notes.

Figure 2: (Left) Power and CAT5 cables going in to the chamber. (Right) Liquid Nitrogen is used to cool the chamber.

Figure 2: (Left) Power and CAT5 cables going into the chamber. (Right) Liquid Nitrogen is used to cool the chamber.

Bring Everything
The lab may be able to provide a lot of miscellaneous supplies, but don’t waste lab time running to stores. Every application may require different supplies, so review your systems’ requirements carefully. However, some things are universal. Make sure there are enough power strips, especially if dealing with wall warts, to allow for only being able to get two or three devices per power strip. Bring a backup for any communications systems or special software that your systems under test interact with. Bring various kinds of tape, labels and different colored markers—these are helpful when additional identification for systems, cords, components, and plugs is necessary. Bring an extra laptop or notebook to jot down notes about reconfigurations, or test details, and if you use a laptop as part of your test configuration make sure it could act as a backup to this device in a pinch. Plan on taking the notes on a device that is not involved in the testing in any way, thus avoiding constantly toggling between windows or missing new information while documenting. Also keep a pen and paper handy; old fashioned note taking never goes out of style and rarely fails.

Don’t Rely Solely on the Lab for Information
Send someone to the lab with the equipment and plan to have them document separately. The lab will document and provide a comprehensive report after testing, but it will solely be based on the readings from their equipment. Anecdotal notes could help fill in the gaps. For instance, did a device really go off line during vibration testing, or did the Ethernet merely come loose?  Things to note are sporadic behavior, failures, and malfunctions. It’s important to note if the failure happened immediately in the cycle or towards the end. Also note what is required to recover a device if a failure does occur. Did it recover on its own?  Was a particular manual mechanism required? Did it start up correctly the first time?

Communication is Key
Ensure that there are multiple communication paths to the systems during testing. Get creative with how to get the equipment to communicate in as many ways as possible about what it is experiencing. In addition to your system’s standard communication interfaces, an external I/O signal, LED communication, and various serial interfaces all can be helpful communication methods. Additionally, try to set the system up to do some local logging of data so that data is available after the fact, but don’t rely on this for immediate feedback. If the device has an onboard clock, set the time close to the lab time for a reference point when reviewing the logs to determine which testing was running when something of note was recorded in the lab’s notes.

Think Visually
Bring a camera. Different labs may have different policies about taking photos, so ask ahead of time, but having a camera to take a visual record of testing is very helpful. Once a series of tests is completed, take a close up of each unit. There may have been damage that isn’t noticed in the moment that will be able to be identified in the photographs later. If the lab allows, take photos of the test setup; they say a picture is worth a thousand words for a reason.

Figure 3: (Left) The Rigid 5 outlet power hub in handy when dealing with multiple wall wart power supplies. (Right) Single Board Computers in chamber mounted, connected, and identified.

Figure 3: (Left) The Rigid 5 outlet power hub is handy when dealing with multiple wall wart power supplies. (Right) Single Board Computers in chamber mounted, connected, and identified.

 

Lab Day Preparations
Once in the lab it’s time to set up, verify that each unit is operating properly, and put the devices in the chamber. The lab technicians will advise about the best setup for their chamber, but be prepared to explain to the technician what access you need to the units and how they will interact with each other or your test equipment outside of the chamber. Place anything that requires manipulation close to the portal. If there are multiple units needing manipulation they will need to be arranged so that they are all reachable. Make sure all of the careful labeling is readable. Verify that all of the cables are attached such that they won’t interfere with each other or fly off during testing.

The more prepared you are the more efficient you will be in the lab and the more you will get out of your testing. Collecting as much usable data as possible is the goal so make sure you don’t impede your ability to do so, and more importantly, make sure you get your money’s worth. In the conclusion to this series we will go through an explanation of each of the steps in HALT and provide examples of real world data from our testing.


ABrownAlan E Brown is a Marketing Communications Manager at Technologic Systems.

Share and Enjoy:
  • Digg
  • Sphinn
  • del.icio.us
  • Facebook
  • Mixx
  • Google
  • TwitThis