Search

Choosing an Uninterruptable Power Supply for your ATE

Choosing an Uninterruptable Power Supply for your ATE

Why do you want an Uninterruptable Power Supply (UPS) for your ATE (automatic test equipment) in the first place? There are many reasons to have a UPS. UPSs are useful in many industries. I will limit the following brief discussion to UPSs inside your ATE and those using battery back-up. I will not be discussing redundancy features of some UPSs.

Here are some common reasons for including a UPS into your ATE design:
  • Prevent data loss.
  • Prevent damage to UUT.
  • Prevent damage to people.
  • Keep running if primary power fails (prevent downtime).
  • Gracefully shutdown your equipment.
  • Preserve data previously collected.
  • Condition primary power where it is noisy or unstable.
  • Protect expensive measurement equipment from primary power faults such as lightning, overvoltage, and brownouts.
  • Notify thru email of problems with primary power so someone can take action.

Before we discuss these reasons for having a UPS in your ATE, we are going to discuss three (3) main UPS architectures and look at their benefits and drawbacks.

blank 
blankblank  
blank
Now, let’s revisit some of those common reasons for having a UPS in your ATE using some slightly fictional scenarios.
  • Prevent data loss – You are happily making measurements when all of a sudden someone turns on a hairdryer and the circuit breaker trips. Your ATE immediately powers down and any data that was not written to non-volatile memory is gone. If you had a UPS, that data that was lost could be written to non-volatile memory (hard drive perhaps) and you could then pickup when primary power is restored.
  • Prevent damage to UUT – Your ATE is controlling some mechanical behemoth and lightning strikes the transformer outside your factory. A surge of high voltage courses thru the factory wiring. Goes thru your ATE and fries the electronics in your behemoth. Additionally, as the behemoth was moving at the time of the lightning strike, and control from the ATE became non-existent, it keeps moving right thru and punctures that expensive structure supporting it. If you had a UPS, your ATE control might have never ceased and none of your behemoth circuits would be fried. Your mechanism would have simply parked itself in a safe home position and waited for primary power restoration.
  • Prevent damage to people – Your ATE has several safety interlocks that prevent high-voltage generated by the UUT from electrocuting test technicians. Some of the safety interlocks are mechanical which are not affected by power outages, but several are electrical in nature. A car driving erratically down the road careens into a light pole jarring it just enough to pull the power line too tight, which cause a mechanical disconnect to occur, shutting down power to your company. Your ATE power fails and testing stops. Your test technician reaches over and flips open the protective cover on the UUT to remove it and put it away. She then gets a strong jolt of electrical charge because the electrical interlock has failed. If you had a UPS, all protective interlocks would have remained viable, and your technician would merrily resume testing when the power was re-routed to your building.
  • Keep running if primary power fails – Your ATE is buzzing along testing UUTs at a profitable pace when brown-out condition occurs in your area of the city. Some of the instruments keep right on working while other instruments enter a useless cycle of resetting themselves. Profits plummet. If you had a UPS, your ATE would have kept operating just as if the brownout had never occurred and your stock prices would go up.
  • Gracefully shutdown your equipment – Some knucklehead at the power company causes the voltage on the distribution system to go higher than normal. Some of the equipment in your ATE just gives up and powers-off. You look around wondering what happened. If you had a UPS, your ATE would have been protected from the higher than normal wall voltage. The computer controlling your ATE would have received an alert and your well-crafted test application would have notified the operator of the event, saved all outstanding data, put all equipment and UUT into a safe operating position, and patiently waited for the wall voltage to get back within reason.
  • Preserve data previously collected – Your ATE has just finished a long-term temperature profile while testing a microwave power detector. It has taken you several days to collect all the measurements. A squirrel adds the last bit of material to her new nest and manages to short the transformer feeding your building. All data from the last several days is gone as darkness surrounds the ATE. If you had a UPS, your test application would have had time to preserve all data before the test controller turned off.
  • Condition primary power where it is noisy or unstable – Your brand new shiny ATE is taking measurements of some sensitive audio equipment. The UUT appears to be noisier than the spec would indicate. Unbeknownst to you, the wall power supplied to your ATE is very erratic and full of glitches because of that car battery recycling plant next door. If you had a UPS, your ATE might have been able to measure the noise floor of the audio amplifier very reliably without issues.
  • Protect expensive measurement equipment from primary power faults such as lightning, overvoltage, and brownouts – You just added the world’s greatest gas chromatograph analyzer to your ATE when lightning strikes a power pole several cities away from your lab. The high voltage surges thru the power distribution system only attenuated slightly by protection devices. It reaches your university lab and your shiny analyzer has become the world’s most expensive and hard to use paper weight. If you had a UPS, the surge protection devices successfully clamp the destructive spike and your research into the effects of perfume on male adolescents continues unabated.
  • Notify thru email of problems with primary power – Your new ATE is making great data churning thru its test sequences and you decide it is time to get to know your family again and go on vacation. You are training your children in fly fishing when your smart phone dings at you. Your co-worker tells you your ATE has crashed, but not to worry until you get back from your vacation. Just enjoy the flies. If you had a UPS, your phone would have ding at you informing you the power on your ATE had just glitched and then been restored. Minutes later, your coworker, having checked on the ATE and noticing nothing was out of the ordinary, would have returned to his regularly schedule work. You are free to fish for more flies.

Choosing a UPS requires considering a rather long list of engineering tradeoffs. Obviously, the more you know about your ATE, the easier it is to choose an appropriate UPS that is economically viable. Some engineering questions (not all UPS selection questions are represented here, this just a brief list of common questions) to ponder when choosing a UPS for your ATE are listed below:

  • What style or architecture do I need for my UPS?
  • Do I need a pure sine wave output for sensitive equipment or those with active power factor-controlled power supplies?
  • What voltage and current are available as my primary power input?
  • What type of receptacle exists for the primary power outlet?
  • Should I have a dedicated circuit wired-in for my primary power outlet?
  • What voltage and current would you like to supply to your ATE?
  • What type(s) and quantities of receptacles should the UPS offer for your protected equipment?
  • For how long should the UPS hold up at the specified voltage and current output?
  • Do you want to power the entire ATE, or just a portion of it such as the PC and monitor?
  • Does the UPS offer surge protection? Almost all modern UPSs have this in varying degrees.
  • Does the UPS offer under-voltage and over-voltage protection without battery draw?
  • How large and heavy is this thing? Affects center of gravity and floor loading.
  • How is the UPS to be mounted? Rack mounted or will it sit on a flat surface?
  • Can the UPS be monitored by a computer/network/webpage?
  • Can the UPS send notifications when predefined primary power events occur?
  • How will this UPS connect to my ATE controller? Ethernet, USB, Serial?
  • Can this UPS be expanded with additional batteries if conditions change?
  • What is the environment? Hot temperatures can degrade battery life.
  • What is the cost?
Let’s go through a fictional design challenge as an example.

My customer asked me to choose a UPS to power up three (3) PXI chassis for 30 minutes and send out an email to any number of recipients in case of a power outage. I asked what primary power was available and the answer was a standard wall outlet. Being in the US, this meant a NEMA 5-15R (single-phase 120 volts @15amps). Some quick research at ni.com indicated that our PXIe-1085 chassis could be loaded with almost 800 watts worth of cards. Adding the power consumed in the PXIe chassis itself brings the total up to roughly 1000 W or 1kW. This means that my customer wanted the UPS to supply about 3kW for 30 minutes, and be rack mounted.

Several prominent UPS manufacturers have helpers or wizards to aid you in choosing your UPS from them. A quick check showed that none of the manufacturers offered this amount of capacity in a single-phase primary input of 120VAC. There are, however, some offerings at higher voltages.

An engineering compromise must be attempted! Practically speaking, PXI chassis are rarely populated fully, and when populated fully, rarely need all the power capable of being dissipated. Also, the equipment being protected is quite sensitive. Low noise operation would be a plus. This necessitates an “online” or “double conversion” architecture.

So, I decided to try half as much power and check in with the UPS wizards. 1.5 kW for 30 minutes in an online or double conversion architecture with a NEMA 5-15P input plug. Still no choices exist with these criteria from one major manufacturer. Yet, another manufacturer offers some hope. I click on an offering and peruse the specifications of the device. What! At full load, this UPS will last only 4.5 minutes! The wizard has let me down. I feel another engineering compromise coming on.

Okay, let’s require the customer to rewire their facilities to provide 30 amps at each outlet. Let’s require an L5-30R (twist lock, 120V, 30amp) receptacle. Now we are getting somewhere. Several vendors offer 19” rack mount UPSs with extended battery banks that will meet these requirements. Each UPS needs an external battery bank to be able to hold up for 30 minutes. Each UPS comes with a wired L5-30P connector on a 8-10 foot cord and multiple 5-15R/5-20R receptacles. These UPSs also boost 2.1-2.4 kW of output power. This is actually more than I specified when asking the wizards. These models also offer a network interface card that promises to email the customer if predefined primary power events occur. Surprisingly, this UPS + extended battery combination only weighs a little over 200 lbs (or 92 kilos).

Warning! This example design search required me to make some assumptions that were not based on reality! I.e., I don’t know yet how many protected outlets my customer really needs! I don’t know the actual load required by the ATE or which equipment should be protected. I don’t know if the customer can re-wire their facility. After presenting these parameters to my customer, I may require another session with the UPS wizards.

There is a better way to select the energy required from you UPS. You should calculate the actual watts required to be supported and then select your UPS accordingly. If the hardware is available to power down and interrupt, you can connect a wattmeter in between the primary supply and your ATE, and get exact current draw for your ATE under various conditions. Some ATE conditions are going to draw more current than other conditions. i.e. power up and dynamic operation.

Suppose your hardware is not assembled yet. You can take the absolute maximum of every supported instrument, sum them up, and use that as the load to select your UPS. Spreadsheets are great for this. This is frequently the game plan when your ATE will be built in the future and you want to be 100% sure of the supported backup time. Or you could even take some percentage less than 100% maximum load, as I did in the fictitious example above.

Conclusion:

UPSs are very useful additions to your ATE and can provide a multitude of benefits to your ATE system. However, choosing the proper model involves finding answers to more than a few questions. Don’t expect all these answers to come easily. Choosing a UPS is not an easy task. However, it is reasonably straight-forward. Don’t be surprised if you must compromise your design goals to match what is available by UPS manufacturers. And don’t be surprised if your customer does not know what they want or what they can afford.

Latest Posts

blank

Automated Test System for SunPower

G Systems, known for their expertise in test system development, was granted the task to help SunPower develop a solar power automated test system to test new printed circuit boards (PCBs). Through this project, G Systems aided in the increase of production line accuracy, reliability, and efficiency, all while decreasing costs to SunPower.

Categories