As many of you recall from our post on December 24th, 2025, we hit a significant roadblock with our hardware testing. During the last shipment of one of the DevKits at our disposal, the power supply (which we had already replaced) and the T2080 motherboard sustained damage.
Since then, we have spent quite a few hours on the workbench trying to revive the NXP T2080RDB DevKit, but the situation remains critical.
A photo of the NXP T2080RDB reference DevKit board.
The symptoms
The behavior is consistent but frustrating. We hooked up the serial console, hoping for some life, but unfortunately, no information came out of the serial port. It is complete silence.
As soon as we power up the board, a specific LED close to the CPU turns on. After checking the technical documentation, this indicator seems to signal that the CPU is being kept in a sleep state.
The investigation
So far, none of our attempts have led to a breakthrough or even a solid theory, so we are currently clueless.
To dig deeper, we had to get creative with our quite limited diagnostic tools. We decided to investigate the electronic components and connections on the DevKit board using a piece of classic equipment we had on hand: a very old Tektronix T935A oscilloscope.
Photograph showing a flat signal in the oscilloscope connected to components of the T2080RDB devkit.
For those interested in vintage test gear, the Tektronix T935 is a 35 MHz dual-channel analog scope with a delayed timebase from the T900 series. The specific model we are using, the T935A, is an upgraded version of the standard T935 (and the single-timebase T932).
The “A-series” adds some very useful features that came in handy during our probing:
- Differential input (A-B)
- X/Y mode in full sensitivity for both channels
- DC trigger coupling and composite trigger
- User-selectable CHOP/ALT mode (Non-A series select this automatically based on sweep rate)
You can read more about this workhorse on the Tektronix Wiki.
What We Tested
We painstakingly probed the board using the T935A. Although it cannot detect the high-frequency bus operation of the CPU or DDR lines, many components can be seen. High-impedance readings allow us to check the pins, and we can verify the power lines.
Beyond signal probing, we also attempted to force the board to boot from different media. We altered the DIP switches to try booting from the SD card, the NAND flash, and the SPI flash. The oscilloscope should show us chip selects of them to show read activity.
We tested the electric signals associated with:
- The main power lines (fine)
- The SD card interface (no activity)
- The NOR and NAND flash interfaces (no activity)
- The oscillators that provide clock sources for CPU and RAM (fine: 66 MHz and 133 MHz)
- The reset lines of the T2080 CPU (one good, one shows no activity)
- The CPU clock input line (dead?)
Photograph taken during the electrical track testing session of some components of the T2080RDB devkit.
Current status and next steps
Unfortunately, even with a detailed signal analysis and configuration changes, the CPU remains stuck in reset. The CPLD attempts to pull the CPU out of reset; however, the clock input is dead, and the HRESET line stays low.
The boot phase fails very early. Perhaps the CPLD did not complete the sequence, or maybe it is corrupted. There may also be a hardware issue. A visual inspection shows nothing apparent; we removed the CPU heatsink. The CPU remains cool, which proves that it is not running.
If anyone has experience with T2080RDB boards getting stuck in a reset state or has suggestions on what to probe next, please reach out!
