In a previous article, I discussed the challenges and the importance of proper cable installation. Here, we turn our attention to the challenges of obtaining accurate test results. These best practices and processes come with a field technician’s ability to be efficient and confident in their test results.
I have been around the industry long enough to be present for the evolution and the debate surrounding the best method for testing medium-voltage cables. DC high potential (hipot) testing and VLF with tan delta are the two most common. Each of these tests has advantages and disadvantages. IEEE 400.2–2013, IEEE Guide for Field Testing of Shielded Power Cable Systems Using Very Low Frequency (VLF)(less than 1 Hz), is referenced throughout this article. I encourage you to read it to educate yourself and become more authoritative on these topics.
Definitions:
- Test end: The end of the cable where the connection of the test set to the cable will be performed
- Non-test end: The opposite end of the cable you are testing
- Hipot: High potential test method
- VLF: Very-low frequency
- Cable: Medium-voltage cable assumed
IT’S NOT THE CABLE — IT’S YOUR TEST SETUP
Whether we are talking about typical cold shrink or heat shrink terminations or the increasing use of load break and dead break terminations, proper isolation and test apparatus setup are crucial. I remember a panicked call from an electrical contractor regarding how six conductors he had terminated with load break elbows all failed a DC hipot test.
Now, I know this electrician, and he is a highly qualified cable splicer and terminator. He is one of the few cable terminators I know who has successfully taken and passed the demanding National Cable Splicing Certification Board (NCSCB) exam. You can learn enough about this program at www.ncscbinc.com to know you must be highly proficient and demonstrate hands-on skills to become certified. So, when this electrician called with SIX! failed cables, I knew the issue was unlikely his terminations and more likely the testing practices being used by the supposedly qualified previous test technician who, in all fairness, was probably doing the best they could with the experience, procedure, and materials available.
These cables were terminated with load-break elbows on each end. The test method was DC hipot. The recommended hold voltage for these types of terminations is 53 KV. The test technique had the cables hanging in free air with the high-voltage test lead stuffed into the load break to apply the voltage to the load-break probe (Figure 1).
The problem with this test setup was that as the voltage was increased above 30 KVDC, the elbow would flash over and trip off the test set. This tripping was what prompted the technician to fail the cables.
It may seem obvious that if the applied voltage is flashing over to the semiconductive layer of the load break, the issue is with the test setup, not the cable termination. After all, the load break isn’t designed with creepage distances between its inner and outer construction to be energized in free air. It’s designed to be plugged into an insulated stand-off or a bushing on a transformer or other electrical apparatus.
The solution to these issues when testing load breaks and dead-break rubber goods is to insulate them with the appropriate dead-end insulated stand-off plugs and create a test block that properly insulates the cable termination system. Simply put, you must incur the expense of purchasing the necessary dead-end plugs and junction bars and even build an elbow for a test point. Depending on the equipment rating, you will want your cable testing tool kit to include 15 KV, 25 KV, and possibly 35 KV insulated stand-offs to plug into one end of the cable that is terminated with load breaks. Figure 2 shows load-break plugs.
I have found a two-point junction bar (four-point junction bar for VLF) at the end you are testing from works best. This junction bar allows you to plug in the cable you want to test and also plug in a load-break elbow with a short piece of cable. Figure 3 shows the junction bar to test the cable. Figure 4 demonstrates it being used in testing.
By using the appropriate load break accessories, you can properly insulate the cable system to eliminate the corona and ultimate flashover that will occur at higher voltages. Safety note: It is always important to ground cable shields, and this includes the wire that is connected to the eye of the load break elbows.
Dead break T-bodies have similar requirements to those of load-break elbows. In the case of dead breaks, you need a minimum of two insulated plugs to insulate the non-test end. I recommend putting six plugs in your test kit to minimize having to move them from cable to cable. At the test end, you will test from an insulated plug, a dead-break connector (often referred to as a football around our office), and a dead-break-to-load-break adaptor. Figure 5 shows the cleanest installation and allows you to plug in your load-break elbow test point.
Another arrangement that can work is to use a second football and test through that point. Figure 6 shows the entire arrangement (without the dead breaks) that could be used to perform a VLF withstand of a set of cables terminated with dead breaks.
An increasing number of new construction projects have electrical apparatus (think GIS gear) that use IEC separable connectors (dead breaks) to make the required cable connections. These connectors present testing challenges similar to those already discussed. You need the correct parts to properly insulate the bodies during testing. In a recent project that utilized Nexan products, the contractor needed not only the connectors but also specially designed test rods that could be installed for testing (with an eight-week lead time). If you are not having these conversations and getting these details before the start of testing, you will inevitably have hurdles to obtaining satisfactory test results.
Figure 7 demonstrates the test rods used during a recent VLF withstand test. Note that our process was improved following poor test results when utilizing all that copper strap to tie the phases together.
Figure 7: Test Rods from Separable Connector Manufacturer Required for Appropriate Test Connections
Let’s now discuss the challenges of testing standard terminations (live front connections), including heat and cold shrink terminations. These terminations require less specialty equipment to properly insulate them than do the rubber goods. For DC hipot testing, a short piece of PVC pipe works very well to shield standard cable terminations.
Despite the ease with which standard terminations can be shielded (insulated or isolated), it is common to see cable test results that lack proper shielding (insulation or isolation) during testing. A proper DC hipot test pattern for a medium-voltage cable would have flat or linear leakage currents at each step up to the hold voltage. Large jumps in leakage current (generally above 30 KVDC) are a strong indicator of improper shielding of the cable end to control the corona that begins to occur. Once the hold voltage is achieved, the classic pattern would be to see the leakage current trend down over the course of the 15-minute hold time; this is the absorption current of the cable dielectric. Cables that exhibit flat leakage-current measurements (they start high and stay high) are another possible indicator of improper test setup.
Figure 8 is a cable test report that exhibits a significant increase in leakage current going from the 22.4 KV step to the 33.6 KV step and even more so to the hold voltage of 56 KV in addition to sustained high leakage current (no decrease in leakage current over time). Side note: Other procedural issues with this test example could lead a qualified evaluator to question the test methods used and the validity of the entire report.
VLF Withstand
Much of what I have mentioned for DC hipot holds true for VLF cable testing. If you are only doing a VLF withstand, you might be able to get away with poor shielding or insulating of your cable and associated tan delta numbers to be used as a comparison for the next example. This test by its nature is a pass/fail test with no criteria other than it doesn’t trip the test set off and the micro-amps hold steady throughout the test. Unlike DC hipot, you will not see the reduction of absorption current over time as the AC wave from of the VLF continuously charges and discharges the cable system. Therefore, a poorly shielded system will still produce a passing pass/fail result.
VLF with Tan Delta
The biggest challenge we have experienced is obtaining acceptable TD test results (see IEEE 400.2–2013 for those criteria) on short runs of cables (less than 100 feet and often as little as 40 feet). In these situations, the small amount of cable capacitance (measured in nano-farads) is directly related to small amounts of leakage current. The leakage current is so small that the test set has limitations to accurately measure it and provide accurate TD values. This small amount of leakage current also allows the inevitable leakage across the cable termination to predominate the TD reading. This can lead to higher overall TD readings as well as larger tip-up values.
The best solution anytime you are performing a VLF TD or withstand test, particularly on standard (live front) terminations, is to use corona balls (Figure 9a and Figure 9b). These metal spheres significantly reduce air ionization (corona) and thereby improve the overall readings.
Figure 10 and Figure 11 demonstrate the significant difference made by using corona balls.
Regardless, short runs of cable can exhibit values that exceed the standards laid out in IEEE 400.2–2013. Changes in the weather, such as increased humidity because of an incoming storm over the course of testing a short set of cables can wreak havoc on your test results causing significant increases in tan delta values. When weather conditions are less than ideal, our process is to postpone testing, particularly testing of short runs of cable.
One process challenge with VLF withstand testing is an occasional debate between technicians on the duration of time to perform a withstand test. IEEE is clear when it states:
- The recommended minimum testing time for an installation and/or acceptance withstand test on new cable circuits is 60 minutes at 0.1 Hz.
The standard immediately contradicts itself stating:
- A test time within the range of 15–30 min may be considered if the monitored characteristic (leakage current and or TD value) remains stable for at least 15 min and no failure occurs.
At issue here is one of providing consistency from test to test and a technician’s discretion about how long is long enough. Our solution, but certainly not the only option, is to have a written procedure that clearly spells out the time that a withstand is to be held. I opt for the full 60 minutes regardless of the monitored characteristic value. Consistency is key to avoiding debates on why there are variations in the reported time duration between cable sets.
VLF Is a Non-Destructive Test
A final note on VLF testing versus DC hipot. There is a myth in our industry that DC hipot is a destructive test and that VLF is non-destructive. DC hipot has largely been abandoned for maintenance testing because it can polarize voids leading to failure when the cable is returned to service, thereby making way for the non-destructive VLF test that doesn’t polarize voids. Regardless, if you have a poor splice or other system component, VLF testing will fail the cable system. IEEE 400.2 is clear again in Section 5.1.2 where it states:
- If the accessory or cable insulation is in an advanced condition of degradation (which you can’t necessarily tell by looking at it), the test can cause breakdown before it can be terminated when using test voltages above the operating voltage. (Note that “above operating voltages” is the key wording here).
- Crews should be prepared to install a new splice, cable, or termination if failure unexpectedly occurs during the test.
I can tell you from painful experience that VLF is destructive to cable systems that have “advanced conditions of degradation” such as flaws, voids, water trees, and poor installation practices (see Fall 2022 issue for more details). This was particularly true in a situation where our team was not made aware of splices in a manhole. Our VLF with TD test results (performed up to 1.5 Uo–12 KV in this case) indicated the cable was in satisfactory condition to continue increasing the withstand voltage only to fail during the withstand performed at 2.0 Uo (16 KV) in accordance with IEEE 400.2 Table 3 Standard Test Values. Following this experience, we have made two crucial changes in our processes.
- We communicate with our customers in writing that maintenance testing of cables comes with the risk of unexpected failure, particularly if splices are submerged in water in manholes.
- The cable test voltage values will not exceed the operating system line-line voltage. Therefore, if your power system is — let’s say 12,470 V phase to phase — then the maximum test voltage will be 12.5 KV. Please note that this highest withstand test voltage is below the 16 KV recommended by IEEE 400.2.
Ultimately identifying a cable fault during an outage may be preferred over an unplanned outage. While neither is a great experience, technicians failing a cable while performing maintenance inevitably will face this uncomfortable question from the client: “My cable was just in service an hour ago and now it can’t be returned to service?!?”
Cable Insulation Resistance
One of the first tests performed prior to a DC hipot or a VLF test is the insulation resistance test. NETA provides a reference for expected values of insulation resistance which are typically met by short runs of cables. However, as the cable length begins to extend into thousands of feet, the ultimate insulation resistance will be significantly lower than those NETA values. The reason this happens might be obvious to some. The longer the cable, the more leakage current is available in the system (higher leakage current equals lower insulation resistance). Stated simply, a cable that is 2,000 feet long could be expected to have half the insulation resistance value of a cable 1,000 feet in length. There is even a note in the standard that I have used to explain to customers why their cable doesn’t meet the industry standard.
When cables of significant length fall below the NETA values, it is important to perform the necessary calculations, based on information provided by the manufacturer, to determine the actual minimum resistance value. Figure 12 and Figure 13 provide examples of the calculations required to determine the minimum resistance value.
CONCLUSION
NETA Certified Technicians commonly perform medium-voltage cable testing. It is often one of the first tests a technician will begin to perform without supervision because it is perceived as a simple and repeatable test. That is not to say that it is not riddled with problems and challenges that can make even experienced technicians pull their hair out when failing or poor results are obtained. Our team has spent hours testing and retesting cables to eliminate all the possible issues with our testing process that could contribute to these questionable results. I have shared many of our lessons learned so you can implement procedures and obtain the necessary materials to always be confident in your cable testing projects.
Mose Ramieh is Vice President of Business Development at CBS Field Services. A former Navy man, Texas Longhorn, Vlogger, CrossFit enthusiast, and slow-cigar-smoking champion, Ramieh has been in the electrical testing industry for more than 25 years. He is a Level IV NETA Certified Technician with an eye for simplicity and utilizing the KISS principle in the execution of acceptance and maintenance testing. Over the years, Ramieh has held positions ranging from field service technician, operations, sales, and business development to company owner. To this day, he claims he is on call 24/7/365 to assist anyone with an electrical challenge. That includes you, so be sure to connect with him on the socials.