The connector is within its current rating. Why is it too hot?
Fictional engineering case. All case measurements and outcomes are constructed for this series; no company or reader is being quoted.
Last week’s answer: the probe was part of the measurement
The change from 18 V to approximately 14 V was a reason to investigate the measurement setup before modifying the converter.
In this fictional investigation, the long ground connection introduced additional inductance and created a larger pickup loop. Its interaction with the probe and circuit affected the displayed waveform. Shortening the connection reduced those effects, but a cleaner trace alone did not establish the remaining peak’s accuracy.
The team kept the converter’s operating conditions and oscilloscope settings fixed. It repeated each connection several times and varied the long lead’s position. The large overshoot depended strongly on that arrangement. The short connection gave a much more repeatable result.
As a pickup check, the probe was disconnected from the board and its ground clip connected directly to its tip. The resulting loop was held near the switching circuit without contacting it. A switching-related signal was still visible. That observation supported pickup; it did not quantify every error in the earlier trace.
A second, suitable low-capacitance probe then measured the switching node with a short connection. Its bandwidth and voltage limits covered the measurement. Its loading and connection arrangement were assessed, and it agreed with the roughly 14 V result within the stated measurement uncertainty. An unchanged DC output voltage alone would not have ruled out probe loading of the fast switching waveform. Some ringing remained with both arrangements.
The team therefore rejected the original 18 V capture as a reliable peak measurement. It retained the smaller observed ringing as something to assess across input voltage, load and temperature, with measurement uncertainty included. The result did not justify deleting the overshoot allowance from the design review.
This week’s challenge: the crowded connector
A controller uses a twelve-position wire-to-board connector. Its purchasing description says “8 A per contact”. The product carries 5 A through each of six adjacent loaded contacts, including the return paths. The other six positions are unused.
A single loaded contact passes an initial bench test. Inside the enclosure, the middle of the populated group becomes the hottest region. With a local air temperature of 55°C, the hottest measured contact reaches 102°C at thermal equilibrium.
The design’s documented acceptance limit is 95°C for this assembly. That is a case-specific project limit, not a general connector standard.
The team obtains the supplier’s original rating report. The headline figure came from a single-contact configuration with a specified wire gauge and test arrangement. The production wire gauge matches, but the number of energised contacts and enclosure conditions do not.
A four-wire check across each mated-and-terminated path, using defined sense points and excluding the bulk lead resistance, gives between 7.5 and 8.5 mΩ at low current after cooling. There is no obvious outlier. The crimps look acceptable, although neither appearance nor this cold measurement proves that every connection behaves correctly when hot.
Where would you put the next thermocouple?
Would you investigate the middle contacts, individual crimps, wire entry or local air temperature first? Propose a controlled loading sequence that distinguishes mutual heating from a poor termination.
Which supplier evidence would you need before concluding that the connector can carry 5 A in this arrangement? A single-contact rating is not enough to answer that question.
The answer follows on 22 October. The next challenge examines a replacement MOSFET that improves normal running losses but fails during startup. Share your investigation plan through the Troubleshooter page.
