Understanding hi-pot testing and its standards
Certifying a new product to relevant safety standards is an essential part of any design and development process. But faced with myriad standards, sometimes engineers need clarification about the nature of the tests appropriate for each standard, writes Kent Smith is applications engineer with XP Power.
Typically there are two voltages used to conduct isolation testing; 3kVAC and 4kVAC. For IT and industrial equipment the international safety standard IEC60950-1 applies to AC-DC power supplies.
This stipulates that your product must pass an input-to-output isolation test of 3kVAC. The IEC61010 standard applies to test and measurement equipment and also requires the 3kVAC test.
Medical equipment is covered by standard IEC60601-1. Input-to-output isolation tests need to be conducted at the higher voltage of 4kVAC in order for the design to comply with this standard.
Confusion can occur when the actual system hi-pot testing takes place. Typically there will be two different types of test.
Design verification (DVT) and safety testing may take place first, followed by production and customer qualification testing. Misinterpretation of the relevant standard may result in the product being subjected to a hi-pot design verification test to the levels described above.
However, the standards state that production hi-pot testing should not exceed 1,500VAC, or its DC-rectified equivalent of 2,121VDC, from input to output or input to ground.
The 3kVAC and 4kVAC tests are a stress-test for the main transformer only. To conduct these tests the main transformer is removed from the power supply and tested as a stand-alone part.
If these test voltages are applied to the supply assembly, a failure may be seen, but it’s not a failure of the transformer as a protection barrier, but rather a failure of the basic insulation parts that are tied to ground.
A 1,500VAC test checks the two isolation barriers of the equipment, namely input-to-output and input-to-ground.
For the test to be completed, certain conditions must be met so the product will pass. If the conditions are not met then they can cause false failures and possibly damage the power supply.
Typically input-to-ground and output-to-ground capacitors can be damaged. A failure can also occur in some primary side FETS and any surge protection devices.
A recent call to XP Power’s applications team illustrates the issue. One of its fleXPower series units that comply with both the IEC60950-1 and IEC60601-1 standards was being isolation tested from input to output at 4,242VDC (the rectified equivalent of 3,000VAC). A breakdown occurred from the input connection to the chassis.
This could have been anticipated, since the isolation from the barrier strip to the chassis is designed to withstand a maximum of 1,500VAC (2,121VDC).
Our applications engineer explained the context of this standard and that, in order to ensure that the transformer would meet the 3000VAC test, it should be removed from the supply and tested separately. Alternatively, the unit could be disconnected from its chassis, and its Y capacitors removed, whereupon it would pass the test.
These test issues are a particular problem in Class I equipment where a protective ground is employed to ensure safe operation. Class II equipment utilises double or reinforced insulation to ensure safe operation with no provision for protective ground.
Where Class II power supplies are employed the user is able to test from the input to the output of the power supply at 3,000VAC (or 4,242VDC) for ITE devices or 4,000VAC (5,656VDC) for medical devices to verify the insulation in the supply.
In another good example, the customer was testing a new product for monitoring cell phone networks. The testing had progressed well with the exception of the hi-pot test. This was surprising as the power supply it was using carried medical safety approvals in addition to the ITE approvals that it required, meaning that the isolation specification exceeded the standard in question.
After having verified that the test procedure was correct and that the supply was correctly installed, an investigation of the test voltage found that the voltage applied was in excess of 1,500VAC, but not at the 3kV or 4kV level. It was not clear why the higher voltage test was being applied but the test lab was adamant that it was correct.
After further investigation a joint call with the customer and the test lab to review the test process quickly concluded that the customer had labelled the product for an input voltage range of 80VAC-264VAC as that is the input range specified for the power supply it was using.
This label led the test lab to base the hi-pot testing on 264VAC rather than 240VAC. Once the label was corrected and the test was re-run at 1,500VAC the product passed as expected and the customer’s system is now approved.