0

How does one go about picking the limits for a HALT test? Voltages, temperature, ramp rate, how much vibration and at what frequency? Is there a science to it or some standards? I remember one lab asking me if I had to do HALT style vibration or just vibration and I didn't know the difference.

I'm working on a piece of commercial outdoor telecom equipment (smaller than a breadbox). I've done halt before but always just handed it off to the test lab.

After that how do I take the data I get back and come up with a screen? That might be too broad of a question though so just the general idea would be OK.

confused
  • 1,806
  • 1
  • 20
  • 34

1 Answers1

1

I would call it an art rather than a science, in my experience you start developing the screen with very little heat and vibration and gradually increase temp and vibration independently until failure occurs or you hit predetermined or reasonable limits. You combine the two (vibe and heat) to run a sequence of cycles at the maximum determined limits (say 2 or 3 Hot/Cold cycles with vibration) as a HALT

The HASS profile is determined from the results of the HALT. Typically you pull back the profile across the board and run a Proof of Screen to verify that the HASS is not significantly damaging or impacting the lifetime of the unit. 10% on heat and 50% of vibration is the rule of thumb that I have used in the past.

The difference between HALT and HASS is important to understand. In HALT you are testing the lifetime of the device under accelerated conditions, the goal is to find that limit so you are expecting units to fall out. For HASS the intention is to screen for premature failures and defective units. Say that 95% of your failures happen in the first 5% of the lifetime (or warranty period :P ), a good HASS will accelerate that first 5% in order to catch those premature failures before they leave the door. The art is in determining based on the HALT (which will run through a larger portion of the lifetime) the proper profile so that the non-defective units will function normally for the remainder of the lifetime.

There are different philosophies towards testing, but this is the one I was taught by a lab that does testing for tesla, google, apple, etc. products.

To summarize:

HALT: Push known non-defective devices to failure limits (doesn't need to explode, a recoverable failure to function is enough)

Proof of Screen: Reduce the profile and verify that non-defective units do not acquire damage from the screen. Known defective units should be dropping out at a reliable rate

HASS: Run your production units (or a portion of them), with confidence that you will catch a majority of deffective premature failures and that you are not damaging properly assembled units

crasic
  • 5,815
  • 1
  • 21
  • 43
  • @confused This answer is from my experience running a HALT with the goal of designing a screen that would go in production. The process would be different if e.g. you wanted to estimate the expected lifetime in real years. You should probably consult with your customers about the type of testing they expect to do, you will still likely need an external lab, but you can be as intimately involved with testing as you wish to be – crasic Jan 28 '16 at 07:31