We are performing the DC inductive endurance (life) test on a 4-pole electromagnetic relay in accordance with MIL-PRF-83536 and are looking for guidance from engineers who have performed similar testing.
Test Setup

• Specification: MIL-PRF-83536
• Relay: 4-pole electromagnetic relay
• Endurance requirement: 20,000 DC inductive load operations
• Supply voltage: 28 VDC
• Contact current: 8 A per contact
• All four poles are tested in parallel.
• The load bank consists of one 25 mH inductor and one 3.5 Ω resistor for each contact circuit.
• Operating cycle:
  • 0.5 s ON
  • 3.5 s OFF

Issue
After completing the inductive life test, the relay fails the post-life dielectric withstand test.
While investigating, we monitored the voltage across the relay contacts during contact opening and observed transient voltage spikes of approximately 650-700 V.
Our original test setup had relatively long wiring:

• Approximately 20 m total wiring loop
• Approximately 5 m between the relay and the load bank

To investigate whether the test setup was contributing to the transient voltage, we moved one inductor and one resistor from the load bank much closer to the relay and repeated the test on one contact.
With the shorter wiring, the measured transient voltage reduced to approximately 300-350 V.
This suggests that the wiring layout and associated parasitic inductance/capacitance may be influencing the measured switching transient.
Questions

1. Has anyone performing MIL-PRF-83536 DC inductive life testing observed contact opening voltages as high as 650-700 V?
2. After reducing the wiring length, our transient reduced to approximately 300-350 V. Is this closer to what would normally be expected, or is it still considered high?
3. Some other test facilities have reported peak contact voltages of only 100-150 V under similar test conditions without using any snubber, flyback diode, MOV, or other suppression device, since these are not permitted by the standard. Has anyone else observed similar values?
4. Can the transient voltage alone be sufficient to cause a relay to fail the post-life dielectric withstand test, or should we also be investigating other possible degradation mechanisms?
5. Are there any recommended practices for the physical layout of the relay, load bank, and wiring that minimize parasitic effects while remaining compliant with the intent of MIL-PRF-83536?
6. If anyone has reference oscilloscope captures of:
   • Contact voltage during opening
   • Contact current during opening
7. for a 28 VDC, 8 A DC inductive endurance test, I would greatly appreciate seeing them for comparison.

Any practical experience or reference data would be greatly appreciated. Our goal is to determine whether the remaining 300-350 V transient is typical for this test or whether additional changes to the test setup are needed to better represent the intended MIL-PRF-83536 test conditions.

 

The high transient contact voltages could be contributing to the problem.
I would suggest decoupling the load with a large electrolytic capacitor in parallel with a 100nF ceramic capacitor directly from on side of the contact to the input to the RL load (below) with the load bank as close as possible to the relay contacts, to have a short a wire length as possible that's not decoupled.
That way the inductive kickback voltage will be due only to the 25mH inductor with no significant contribution from the wire length.
Also using twisted-pair wire where you can, significantly reduces wire inductance.

None of that should not be in conflict with the MIL spec requirements.