Soldered 333020 MCP2518 board — SPI/internal loopback works, external CAN fails

TECHNICAL SUPPORT Communication Modules
1

Hello Soldered Support,
I need help diagnosing a possible fault with my Soldered 333020 MCP2518 CAN board, connected via SPI to a LilyGo T-2CAN (ESP32-S3).
It previously communicated successfully with a Solis inverter at 500 kbit/s using Battery Emulator firmware. Communication stopped after a power-supply change and restart. Returning to the original supply did not restore it.
Connections and measurements:

  • External regulated supply: 5.00 V measured between VCC and GND, with shared logic ground.

  • SPI: SCK GPIO38, MISO GPIO37, MOSI GPIO42, CS GPIO41; INT GPIO39.

  • Idle voltages: CANH–GND 2.55 V, CANL–GND 2.54 V.

  • Isolated bench bus: two 120 Ω terminators, approximately 60 Ω measured across H/L with power off.

  • CAN wires were also soldered directly to the terminal solder pads to bypass possible terminal-contact problems.
    Tests performed:

  • Full power cycles of both boards; replacement of two previously pinched wires.

  • SPI initialization and register readback succeed.

  • Five-minute internal loopback: firmware reported 3000/3000 correct frames, 3000/3000 correct INT transitions, and 3022 successful register checks.

  • External communication failed in both directions with two different peers: CANable Pro and the LilyGo’s native CAN channel.

  • During the original fault, CANable could see Solis frames at the same terminals, while the Soldered channel received none.
    External transmission diagnostics:

  • TREC: 0x00108800 — transmit error counter 136.

  • BDIAG0: 0x00001100.

  • BDIAG1: 0x00010000 — NBIT0ERR.

  • No transmitted test frame was received by the external peer.
    We corrected a bit-timing issue in the separate bench-test library and repeated the tests. Readback was NBTCFG=0x003E0F0F for 500 kbit/s with a configured 40 MHz oscillator. OSC=0x00000460. The physical oscillator frequency has not been measured. The original Battery Emulator uses a different CAN library.
    Could you suggest any further checks to distinguish a transceiver fault from a local supply or standby/control issue? Are there accessible multimeter test points? I do not have an oscilloscope or logic analyzer.
    Attached is a diagnostic package containing test logs, source code and an English summary. Please start with README_FIRST.txt, which explains the results and their limitations

    Update: Measured directly at U3: VCC ≈5.0 V, VIO ≈3.3 V, STBY ≈0 V
    Update2: Additional multimeter checks, performed external power disconnected, confirmed near-zero resistance on these paths:

    • U3 pin 7 (CANH) → CANH terminal solder pad.

    • U3 pin 6 (CANL) → CANL terminal solder pad.

    • U3 pin 2 (GND) → board GND.

    • U2 pin 1 (TXCAN) → U3 pin 1 (TXD).

    • U2 pin 2 (RXCAN) → U3 pin 4 (RXD).

    All connecting wires between the LilyGo and Soldered board also passed continuity checks. Powered measurements showed approximately 5 V at C3 and C6, and 3.3 V at C5.

    These checks found no obvious open connections. They confirm DC voltages and continuity, but not signal integrity or transient behavior. Given the successful internal loopback and failed external CAN tests, would you suspect the ATA6561 transceiver, or recommend another check before replacing the board?

2

Soldered_333020_diagnostics_2026-09-19.zip (103.2 KB)
I don’t see attachment in my opened topic. Maybe I forgot to add it, so I add it now via reply message.

3

Hello @paldeiv ,
Thank you for the very detailed summary. First thing that is worth trying is soldering a capacitor (preferably something like 470uF) between the VCC and GND pins of the breakout. This has been known to fix an unstable voltage coming to the chip. A clear sign of this would be SPI data being mirrored over the CAN lines.
If possible, please try this out and get back to us with the results.

4

I tried adding a 470 µF / 25 V low-impedance electrolytic capacitor directly between VCC and GND, with the correct polarity and all power disconnected during installation.

After powering up:

  • VCC–GND and C3 remain approximately 5 V.

  • C5 now measures approximately 1.7 V instead of the previously measured 3.3 V.

  • The area around regulator U1 becomes hot.

  • The PWR LED lights for approximately five seconds, then turns off, although VCC remains at 5 V.

  • MCP2518 initialization now fails; previously, initialization and internal loopback tests passed.

I removed the capacitor, but these symptoms persist. The board is now disconnected from power.

Could this indicate a regulator fault or an overload on the 3.3 V rail? Please advise on the next steps or replacement/ refund.

5

If the capacitor fix caused this, then there was definitely an underlying problem beforehand.
I am sorry that you have had this experience and you are entitled for a replacement. For a replacement/return, contact us via hello@soldered.com email with your order number and either a link to this forum thread or explaining the issue.