Years ago Cirris provided 1000 volt hipot testers that were integrated into AMP's R-CAM Automatic Insulation Displacement (IDC) Ribbon Cable Assembly Machine. The machines built ribbon cables with wires on 0.050" centers that were, at best, rated for 300VDC. A 1000 Volt hipot test was found to be most effective in finding problems in damaged insulation on wires that passed between adjacent tines in the IDC connector; anything less than 1000VDC could miss errors. Was this a crazy idea? Did it damage or degrade the ribbon cables? What about cables that will not pass a high voltage test due to the use of small-pitch connectors in the assembly? Is the specification in use unreasonable? Let's explore some answers.
What is the goal of testing with high voltage?
To find "near shorts" that low voltage testing would miss, and to do this without damaging the cable being tested.
Myth: Use the wire or connector "maximum working voltage" specification as the limit for high voltage cable testing
A logical concern is that exceeding the rated working voltage of the wire or connector during hipot testing will weaken or damage the insulation. Most people have seen what happens when you get a short in an AC extension cord plug. Sparks fly, resulting in charred remains of insulation, melted metal embedded into the insulation where the short occurred, and insulation is burned away creating little barrier to repeated shorting. If this is what happens when you have a short at 120VAC limited by a circuit breaker of 15 amps or more, what might 1000VDC do to your cable assembly?
If you have a thorough understanding of electricity, you will recognize two things wrong with this comparison between a 120VAC short and High Voltage cable testing.
- The energy that most testers apply to a cable under test is usually very limited (most Cirris testers limit this to 0.03 Joules). Both the current and the time that the current is applied are carefully controlled. If you were to touch wires together on a Cirris hipot tester during a high voltage test, no damage would occur. This is because of the limited amount of current and quick shut down time. In contrast, an AC line short that triggers a circuit breaker takes maybe 200 milliseconds at perhaps up to 100 amps. This would be 0.2 seconds X 120 Volts X 100 amps = 2400 Joules, or 80,000 times more energy than a failure during high voltage testing on a Cirris tester!
- In our AC extension cord example, the wires must touch for the failure to occur! The 120VAC did not cause the failure. With Cirris testers, we can detect a breakdown even through a small air gap. Remember back to Paschen's law: an arc will not start in air until around 327 VDC or 231 VAC (the peak voltage at 120 VAC is 1.41 X 120VAC RMS = 170 VDC peak, which is not enough to cause an arc).
- If the direct short (breakdown) had not occured, no damage would have resulted.
Cable/Harness test specifications have not used "wire or connector working voltage" as a limit
IPC/WHMA A-620, and all military specifications we are aware of, do not use working voltage as a limit to test voltage. A-620 uses 1500 VDC or 1000 VAC to test Class 3 assemblies (High Performance Electronic Products . . . where performance on demand is critical).
High Voltage is also required in A-620 Class 2 assemblies (Dedicated Service Electronic Products . . . where high performance and extended life is required), with pin spacings less than 2mm (0.079") due to the higher risk of shorting due to close pin spacing.
UL labs require much higher voltages than 'working voltage' for their testing
The components, which are connected to the AC line and are typically found in appliance and technology products, have working voltage specifications of 120 to 300 VAC. Yet UL in their specifications 982, 1010 and 1082 for appliances, and 60950 for Technology Equipment (ITE) often require 100% testing at thousands of volts. The voltages used? Typically 1000VAC + 2 times the voltage of operation. This means that 1240 VAC is regularly applied to products with 120 VAC rated working components and 300 VAC wiring.
Your wire supplier likely applies much more than 1000 volts in test voltage to your wire
If you have ever had a chance to see insulated wire being manufactured, you may have seen a tester checking for defects in the insulation. This testing usually consists of a small bead chain (like those old key chains) or perhaps a brush raised to a very high voltage, through which the wire is passed. For 300V rated wire, several thousand volts is applied by the "spark tester." To date, there is no evidence that suggests that this testing has degraded the insulation. As a result, this practice is standard and required by most UL and MIL wire specifications.
It is interesting that people worry about potential damage from re-testing a cable or harness assembly at 1500VDC when these "spark tests" have previously been performed on every wire of their assembly. As an example, MIL-W-16878E TYPE EE (TFE hookup wire) rated at 1000 volts must be spark gap tested at 5,000 volts!.
Research supports the assumption that higher voltages do not permanently damage insulation
In 2003, Sandia National Labs tested aircraft wiring for degradation, not just at 1500 volts but all the way to breakdown requiring thousands of volts! You can read the report here: Glover2003.pdf
Damage was not identified until breakdown occured. The first detectable damage observed did not occur until 10 times the energy of a Cirris tester was used.
Dielectric breakdowns that damage insulation + the role of IR testing
There is no evidence that damage occurs if a cable is subjected to higher voltages--as long as a breakdown does not occur. However, what if a DWV failure (breakdown) does occur? The energy is limited so that the risk of damage is low. Moreover, if damage did occur, the IR test is designed to detect the carbon trail damage that may be created. This is the reasoning in A-620 why IR testing must be performed if a DWV test is performed, and it is not allowed to precede the DWV test.
While specifications such as A-620 are good for general guidance, are there considerations (like very small connectors) where test voltage should be reduced?
Specifications like A-620 give good guidance on test voltages. To be more rigorous, you can test assemblies as though they were in higher classes. But connectors do keep getting smaller. At some point, the connector itself cannot hold up at a specified test voltage.
- High voltage testing is effective in catching "near shorts" that low voltage testing may miss. These "near shorts" can turn into real shorts in the field (latent defects).
- It is NOT appropriate to use Rated Working Voltage of wires/connectors as a limit for test voltage.
- We lack evidence that high voltage testing damages or degrades the insulation of the wires or connectors in a harness assembly unless breakdown occurs. Even breakdown with high voltage at low energy levels is not expected to create damage.
- IR testing should be done concurrent with or after DWV testing to ensure that any possible damage from a dielectric failure is detected.