Hipot test

I have been involved in the design for production of many power electronic products, including switching power supplies, electronic ballasts and automotive electronics for Hybrid vehicles. What I have found is the consistency of the breakdown voltages is different for AC and DC. I've sampled 100 ballast from a high volume production line. 50 of the units were subjected to a slow ramp to a sufficiently high DC Hipot breakdown voltage. The other 50 units were subjected to the same ramp but with AC. We found that traces in the PCB that has either spacing issues or contamination broke down, tripping the Hipot tester to record the voltage. With the same pcb defect, AC failures were within 5 to 8% of each other. However, with DC the breakdowns were higher and non consistent varying as much as 50%. Usually, most companies limit the Hipot test to once. When we have a surface breakdown, carbon tracks lower the subsequent breakdowns. DC Hipot breakdown results are effected by humidity, wind, temperature, altitude, and rate of rise, especially when companies add 20% to the peak value of the Hipot test to do the test in 10 sec instead of 1 min. There are additional stresses applied during the ac Hipot test. The two power lines are tied together and Hipot voltages are applied to those two power leads and the case safety ground wire and output leads on an isolated design . The high voltages distribute across capacitive elements to ground, but that happens if a high common mode voltage occurs during normal operation. I always test any elements that are capacitive coupled to safety ground for corona inception and extinguishing voltages to understand the effects of the Hipot on the product.

Leakage thru transformer capacitance and "Y" caps are not determined by the Hi Pot test it is a separate test. Leakage tests are performed on the magnetics before they are placed. The "Y" caps have to be certified by the various agencies to be used for a Hipot test. The only products that I’ve had experience with that we had to disconnect the “Y” caps for Hipot testing are military equipment with high value “Y” caps. The steady state leakage current is accounted for in the settings of the Hipot tester, which can trip on arcing and excessive capacitive current.