Advertisement

Blog

Prevent Your Next High-Flying Product Launch From Being Grounded

The widening gap between the 24V-to-28V intermediate bus commonly found in industrial, aerospace, and defense systems and the input supply voltages of modern digital processors presents design risk that could easily result in system failure, noxious smoke, or, even worse, a fire.

A relatively higher input voltage makes it increasingly difficult to maintain power supply voltages within the safe limits of the processor and has direct implications for the product in terms of size, operating costs, safety, and reliability. These risks of overvoltage include, but are not limited to:

  • Input surge events on the intermediate power bus
  • Slight inaccuracies in the PFM/PWM timing of the switching regulator
  • Erroneous, gray market, or counterfeit capacitors introduced during manufacturing

While most firms deny that illegitimate electronic components could ever enter their assembly lines, customer support calls we’ve received at Linear Technology, discussions with our peers at other IC firms, and even a US Senate Committee report released in May 2012 unfortunately indicate black market and counterfeit electronic components are a regular occurrence, even at the most reputable companies and in the most secure applications.

Any one of the three events listed above could cause output voltage excursions exceeding the load’s ratings, potentially causing the costly FPGA, ASIC, or microprocessor to be permanently damaged and, in some extreme cases, ignite. Depending on the extent of the damage, the root cause may be quite challenging to determine, and the resulting high repair costs, lost customer productivity, and harm to your reputation can be extremely frustrating.

If your system employs an intermediate voltage bus, a risk mitigation plan deserves your consideration to minimize cost and inconvenience to customers. Traditional overvoltage protection schemes involving a fuse are not suitable for protecting modern FPGAs, ASICs, and microprocessors, particularly when the upstream voltage rail is 24V or 28V nominal. The response time is highly variable and too inaccurate to guarantee protection in such a high input-to-output voltage ratio application. Moreover, even if the digital logic device is successfully protected, the recovery process is lengthy and the resulting downtime irritating to clients, as human intervention is necessary to replace the fuse before attempting a system restart.

A new solution has been created, combining a 38V-rated, 10A DC/DC switching regulator with circuitry to defend against many faults, including output overvoltage to protect high-value FPGAs, ASICs, or microprocessors.

In case of an overvoltage event at the load, every regulator is tested at the factory and guaranteed to engage protection within 500 ns before it ever reaches a customer. Furthermore, recovery is fast and simple. Simply toggle a logic level control pin to resume normal operation, assuming the fault has cleared — otherwise the protection will immediately engage again, indicating a more serious problem. Power and protection for today’s most advanced digital logic devices are now available in one compact surface-mount device.

Click on the photo below for a two-minute Techclip video demonstration:

3 comments on “Prevent Your Next High-Flying Product Launch From Being Grounded

  1. eafpres
    May 16, 2013

    Hi Willie–interesting post and situation created by the evolution of voltage standards within industries.

    Question for you: Over on The Connecting Edge (sister site) there has been a 2-part tutorial on signal spectra basics.  You can find Part 2 here:

    Basics of Signal Spectra, Part 2.

    The point has been made clearly that for digital pulses, there are high frequency components of significant magnitude.

    Just doing a back of the envelope calculation, is there a potential that in an RF system the 500 nS protection time might be insufficient?  I have to admit I don't really know what to expect in terms of voltage transients you are referring to.

  2. WKetel
    May 23, 2013

    I suppose that the constantly dropping supply voltages do make greater demands on the secondary regulators from those intermediate power busses. And for all of those systems designed with minimum total cost as the primary design objective, as well as the secondary target, I have no sympathy at all. A bit of filtering at the input to the regulator will not only slow many of those spikes down to a much more manageable speed, and much lower amplitude, the same filter will prevent noise from the switching regulator from traveling upstream and causing problems elsewhere. Just because the internal regulator does not need to meet the same requirements as the AC line input regulator module does not mean that designing unfiltered junk should be the rule.

  3. Willie Chan
    May 24, 2013

    @eafpres: Thanks for the comment. In the scenario I describe, I focus on the intermediate power bus where voltage spikes in the several or tens of volts are caused by other abrupt loads such as solenoids, motors, fans, hot plugged daughter cards or on the upstream side a brownout from the utility or switching from one power supply to another i.e. auxiliary power or battery power to main power.

    While there could be some noise coupling of high frequency components from high speed data lines, their energy is relatively low such that they would easily be absorbed by the bypass capacitance at the output of the DC/DC regulator. Moreover, good PCB layout practice would dictate a short trace length between the output of the DC/DC regulator and the digital processor to minimize any energy coupling into the power rail. I hope I answered your question.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.