How to Test Protective Relays Correctly

Usually I try to keep my posts as simple and practical as possible.  This post is a little different because I will discuss how I approach all relay testing.

Skip this one if you are looking for relay testing steps for a specific element, or skip ahead to the “How Should You Test Protective Relays Treatise” section if you don’t want the background.

Why I Wrote This Post

I was writing the next impedance relay testing post when I realized that a large part of it was going to be about the “right” way to test relays. I’ll just end up writing it over and over again in every post because these principles apply to all relays, so I figured that I should just write it down once to get it out of my system… and then link to it in all future posts.


I didn’t start my career as a relay tester.  I started in a full-service, high-voltage testing company. My first job was running the ductor leads for a switchgear contact resistance test. I was very lucky because the company I worked for didn’t just teach the right button sequence to get a successful test; they wanted me to understand the principles behind contact resistance before we looked at the test-set controls, including:

  • Why I was performing the test.
  • What the test was measuring.
  • How the test-set was measuring it.
  • How to connect my leads to get accurate results.
  • What a good test result looks like.
  • How to troubleshoot problems.

I don’t know if this was on purpose, but understanding these basic principles about a contact resistance test meant that it didn’t matter which test-set they sent me out with; I knew what I wanted and it was just a matter of figuring out which buttons to push to produce and measure the signals I needed. If the test-set broke down, I could find anything that produced DC current with a meter that could measure DC voltage, and I was back in business.

One of the most common complaints I get from students is something like, “We don’t test relays every day, so it’s hard…” I find this one difficult to understand because you could put me in any high-voltage electrical testing environment today and I could get the job done; even though I haven’t tested anything other than relays in nearly a decade.

This isn’t bragging.  I don’t think I’m special. I know plenty of test managers who are in the same boat as I am. I think it has more to do with my approach to testing.

Our modern over-reliance on automation is the biggest problem. I wrote a rant in The Relay Testing Handbook: Principles and Practice against automation in all forms, but I’ve changed my thinking quite a bit since then. I think I got lucky again with that first testing company.  They were too cheap to buy the latest and greatest equipment, but forward thinking enough not to rely on carbon piles, lantern batteries, and multi-meters. We had to:

  • make the right connections,
  • run the test manually,
  • understand what the results meant, and
  • make sure they were correct.

If you understand the test and perform it manually until it becomes boring, those fundamentals will stick with you forever. You aren’t going to learn anything if you simply follow the test-set instructions for connections, push a button, and then start checking your email while the test runs.

Here’s a quick test: How does your EZCT-2000 (or any other CT tester that you can carry to the test location) test the CT ratio? If you think that your CT tester uses current to test the CT ratio and polarity, you shouldn’t be testing that CT.  That automated test-set is basically a slot machine if you don’t know what you’re doing.  It looks like you’re winning because you keep getting passes, but your luck will run out, eventually, and you’ll be stuck trying to figure out how to get out of the hole you’ve dug for yourself.

I encourage you to avoid the automation trap and look for test-sets made in the 80’s or 90’s (or run your modern test-sets manually) to do your testing. Learn the principles with manual testing until you are bored to tears.  Then start playing with automation. A confident electrical tester who knows their stuff can almost write their own ticket in this industry. Every testing company in the world is actively looking for (and poaching) skilled electrical testers.

How Should You Test Protective Relays Treatise

Electro-mechanical Relay Testing

Electro-mechanical relays had multiple failure points that could affect the relay pickup and timing characteristics. In fact, you can make one characteristic pass by changing another characteristic. For example, one of my students found the relay timing out of calibration. He temporarily forgot how to adjust the time delay and started adjusting the pickup characteristic instead.  He managed to get all the timing tests to pass, but the student who checked the relay after him found that the pickup was nowhere near the calibration tolerances.

The acceptance, commissioning, maintenance, and troubleshooting tests in electro-mechanical relays were all identical because the failure points in the relay didn’t change.  You could calibrate and clean a relay to perfection today, and the electromagnetic and mechanical nature of the relay would cause the pickup, timing, and contact characteristics to change over time.

Digital Relay Failure Points

Those electromechanical relay problems are almost impossible with digital relays. The functions either work, or they don’t.  There is no reason to find the exact pickup points in a digital relay because there would be nothing to tweak to bring it into tolerance. The real failure points of digital relays are:

  1. Analog/digital converters
  2. Power supplies
  3. Electronic components
  4. Digital inputs
  5. Digital outputs
  6. Relay algorithms
  7. Relay settings

The first five failure points can be tested with simple test procedures and don’t need fancy test-sets. The relay algorithms should have been tested before the owner purchased the relay and will not change over time. If you set and measure a pickup and timing test today, it will be the same forever as long as the first five components remain functional.

Digital relays can have different test procedures for the different stages of a digital relay’s lifespan. In fact, the first four failure points could be, and should be, continually tested while the relay is in-service and online. Digital outputs can be tested in a few seconds during periodic maintenance intervals. The algorithms shouldn’t change, so you shouldn’t need to keep testing the relay pickup and timing during every maintenance interval.

The question you should ask before you perform a maintenance test is: Do I trust the person who commissioned this relay?

I’ve found too many in-service relays that were incorrectly set to fully trust that any relay has been correctly commissioned.  Many of the setting problems ended up disabling one, or all, of the protective elements in the relay; but they had a tested sticker and test report. I’ve also found problems with relays that I’ve tested over multiple maintenance intervals as my testing evolved, so I always perform a full battery of tests during maintenance intervals because it doesn’t take that much longer if you perform dynamic tests.

Relay settings problems are THE most common failure points with digital relays. Some of these setting errors can include:

  1. Relay setting engineers misunderstand how the relay functions.
  2. Relay setting engineers make a typo, or copy and paste error, when creating the settings.
  3. Relay testers convert incompatible setting files in the field.
  4. Relay testers incorrectly upload the relay settings.
  5. Relay setting engineers base their calculations on incorrect information.

If you base your test plan on the settings inside the relay, then build a test plan based on those settings that changes values to make the relay test simpler; you will not be able to find any of these issues unless they are incredibly obvious.

You can find the first four problems by knowing what the relay is supposed to do (from a description of operation from the design engineer, or experience) and then applying realistic fault scenarios to prove the relay operates as it was intended. You won’t find any of these problems if you ask the relay what it is programmed to do and isolate each element for testing.

You can find the #5 problems on the list by reviewing all the site documentation and connections to make sure everything matches. There are often many, many revisions on a project and you are the last line of defense. You should use the final drawing revisions to make sure the electricians and engineers were all on the same page.

A Relay Testing Analogy

Let’s simplify this by replacing the relay with a bank machine, or ATM, that distributes three kinds of bills ($5, $10, $20) stored in three separate hoppers.

A smart and simple way of testing this bank machine is to install it, make sure it powers up, load it full of money using the standard procedure, and then walk through the withdrawal process as if you were a customer. If you ask for five dollars, you should get five dollars.  Repeat the process for some basic combinations and head home after a couple of minutes.

The procedure I just described is called dynamic, or system, testing in the relay testing world. If I have a relay that is supposed to protect something, I will:

  • find out which conditions should make the relay operate, and then
  • create some realistic scenarios that are just within the relay’s operating tolerances, and then
  • create some realistic scenarios that are just outside the relay’s operating tolerance, and then
  • run the tests to make sure the relay operates when the test is within the tolerance and doesn’t operate when the test is outside the tolerance.

When I use this testing philosophy, I can walk away from the relay fairly confident that the relay will operate correctly when in service.

Most relay testers create some kind of automated procedure to test their relays that would look something like this using the ATM machine analogy:

  • Disconnect the keypad and replace it with a connection from the test-set that will virtually press buttons. (This would be similar to connecting your test-set voltage and current channels to the back of the relay the way the relay manufacturer shows the connections instead of looking at the drawing and trying to replace the CTs and PTs with test-set channels, using the site’s three-line drawings.)
  • Disconnect the hopper that dispenses the bills and replace it with the test-set digital input feedback inputs. (This would be the same as mapping spare inputs for your relay testing.)
  • Download the program from the machine, analyze the program, and create simple commands that will simulate the keypad presses required to get hopper #1 to operate.
  • Repeat for hopper #2 and hopper #3.
  • Run the three tests and make sure the correct hopper operates for each simulation.
  • Disconnect the test-set from the ATM and reconnect the keypad and hoppers.
  • Print your report.

This test procedure is much more involved and makes it look like you are really working hard on your test, but what did you actually prove?

Have you ever noticed that the keys on a telephone keypad are opposite those on a computer keypad?

Which one is the right style for the ATM? Did you program the correct style when you built your test plan? You could go down a rabbit hole of research to press the correct virtual button with your automated test, but these questions are unnecessary when performing a dynamic test.  If the right amount of money comes out when I pretend to be a customer with the ATM test (or the correct responses occur when I simulate a fault on either side of a relay pickup), I don’t need to learn the machine language responsible to prove that function is correct.

You can get passes on your test sheets in the relay testing world by looking at the relay terminals for your test-set channels, but the relay may never operate if the CTs or PTs are incorrectly  connected to the relay. You can get passes on your test sheets AND find out if the CTs and PTs are connected to the relay correctly by connecting your test-set to the relay, using the site drawings. ALWAYS simulate real-world conditions when you can!

Downloading the program (relay settings), figuring out the ATM programming style (relay element names and references), and reviewing the ATM logic (relay logic) is something that an experienced tester or engineer can do after years of experience, but is it necessary to test the ATM (relay)? You can reverse engineer all of the settings to create a test plan, but you’ll only prove that the machine will do exactly what it is programmed to do, not that it is appropriate for the application. Garbage in usually equals garbage out.

For example, what happens to the numbering of the hoppers if we looked at the back of the ATM machine?

You could start a rabbit-hole of research to determine whether the numbers #1-2-3 start from left to right from the front or rear view, and build a test plan to test that; but you would only prove the machine does what is programmed to do. If the machine gets loaded from the back and the labels show #1-2-3 from left to right, the machine will be loaded with the $5 and $20 bills in the wrong location.  The machine will dutifully spit out $20 to people who ask for $5 and $5 to people who ask for $20, even after the complicated test procedure gives you all passes.

An automated test makes it look like you are winning with your good-looking test report ($20 instead of $5), but the bank will probably lose money (relay fails to operate when it is supposed to) over the long run. The dynamic tester doesn’t need to do any research to make sure the right bills are dispersed, and they can complete the test in less time with a fancy test report.

Let’s look at the last link in the automated testing chain. The automated test needed to get input signals to receive inputs from the ATM, so the hopper connections were disconnected from the actual, in-service equipment (mapping a spare output) to get feedback to the test program. If you get a successful test with this method, do you even know if the hopper (actual trip output) is fully functional? No. There could be something mechanically wrong with the hopper that prevents it from operating correctly. Or wires could be crossed to disburse the wrong bills. The only sure way to test the actual hopper is to perform a full functional test.  Why not just skip the in-between part and perform a fully functional test from the beginning?

How Should You Test Protective Relays Summary

Testers who rely on automation without understanding what is happening in the background are essentially pulling the handle on a slot machine. The bells and whistles (test report passes) make it look like you’re winning, but your luck will run out eventually.

I encourage you to use manual test methods for all your testing until the procedures actually sink in (you’ve done it so often that you could do it in your sleep), and then start playing with automation.  All of your test procedures should follow this path.

  • Ask yourself, “What are the goals for this test?” and plan your test to meet those goals.
  • Find out what the relay is supposed to do without looking at the settings in the relay. Try to get as far away from the settings as you can (coordination studies for overcurrent elements, line parameters for distance relays, setting engineer descriptions of operation or working sheets, documented, version-controlled settings from the design engineer)
  • Review the site drawings and replace the CT/PT inputs with your test set as far from the relay as possible to test more of the circuit.
  • Use in-service digital inputs whenever possible.
  • Use the actual, in-use output contacts instead of spare-output simulations.
  • Create realistic fault simulations on either side of the element pickup to prove the relay will ignore faults on the wrong side of pickup, and trip for faults on the operate side of pickup.
  • If you can’t test an element without another element turning on, contact the design engineer to determine the scenarios that should operate without overlap. Usually the engineer doesn’t realize that an overlap has been set and will change the settings accordingly.
  • Make sure your test will actually meet your goals. For example, does your CT secondary loop test prove that the current flows into the correct terminal? Many test plans that I’ve seen used will NOT prove polarity. Have someone switch the wires and see if your test plan will discover the change.

Thanks for reading this soapbox rant to the end.  I’ll be working on more technical emails in the future.


Image Attributions:

  • Numeric keypad = Numpad.svg By Mysid (Self-made in Inkscape) [Public domain], via Wikimedia Commons
  • Telephone keypad =


About the Author Chris Werstiuk

Chris is an Electrical Engineering Technologist, a Journeyman Power System Electrician, and a Professional Engineer. He is also the Author of The Relay Testing Handbook series and founder of Valence Electrical Training Services. You can find out more about Chris here.

follow me on:

Leave a Comment:

Send this to a friend