Hi,
I’m bringing up a custom i.MX6ULZ board and I’m stuck in an odd state where the Boot ROM is clearly alive, but I can’t get reliable USB SDP transfers or any proof that a raw OCRAM payload executes after jump-address.
Board / setup
- MCU: i.MX6ULZ, custom PCB
- USB: USB OTG1 over USB-C connector
- UART: UART1 broken out and connected through CH340 USB-UART
- No DDR populated yet
- Two separate assembled boards from the same PCB design show very similar behavior
What is definitely working
- Boot ROM over UART SDP works
- sdphost can:
- read registers/memory
- write a raw binary to OCRAM at 0x00900000
- Reading back OCRAM after write confirms the contents changed, so OCRAM writes are real
USB behavior
- Over USB, the board is recognized by UUU as:
- MX6ULL SDP: 0x15A2 0x0080
- USB attach is working well enough to enumerate
- uuu starts the SDP write, progress goes to 100%, then fails with timeout
USB hardware details
- USB-C used in USB2 device mode
- CC1 and CC2 each have 5.1k pulldown to GND
- D+ and D- initially had 4.7 ohm series resistors, later replaced with 0 ohm for testing
- ESD diode was also removed for testing
- None of these changes fixed the USB SDP write timeout issue
- D+ and D- both measure ~0V with DMM at idle
- USB_OTG1_VBUS was checked before BGA reflow and had raw 5V from VBUS at the MCU-side net
Power / internal rail observations
- POR_B is solid high when expected
- VDD_HIGH_CAP around ~2.5V
- VDD_SOC_CAP and VDD_ARM_CAP are present
- VDD_USB_CAP around ~3.0V
- NVCC_PLL around ~1.1V
- I did not observe obvious dips on VDD_USB_CAP or NVCC_PLL during the USB write attempt with a DMM
Long story short
- ROM over UART clearly works
- OCRAM writes clearly work
- USB SDP clearly enumerates
- Yet USB SDP write always times out
- And I still cannot prove my raw OCRAM code executes
Questions
- Is there any common mistake in custom i.MX6UL/ULZ USB OTG bring-up that could cause this kind of partial behavior?
- Does anything in the description stand out as suspicious or inconsistent?
Thanks,
Sam
here are screenshots of a writing attempt using uuu. as mentioned, it goes to 100% but then times out. I forgot to mention that once it fails, querying using uuu -lsusb takes much much longer to excute:
Hi @samprince
Could you please share your uuu file, and i will test your scenario on my board.
B.R