Post Go back to editing

LTC3300-1 (Daisy Chain) – Bottom Chip Not Responding to Read Balance Status (0xAC)

Category: Hardware
Product Number: LTC3300-1
Description:
We are using an STM32G4 MCU to communicate with three LTC3300-1 chips via SPI in daisy chain configuration. The SPI settings are:
  • Mode: 3 (CPOL = 1, CPHA = 1)
  • Clock Speed: 750 kbps
  • Data Frame Format: 8 bits
We have verified that:
  • TOS and VMODE signals are correct
  • VREG output for all LTC3300-1 chips is 4.8V
  • CS, SCK, and MOSI signals are toggling correctly and meet timing specifications

Issue:
The bottom LTC3300-1 in the daisy chain does not respond to the Read Balance Status command (0xAC). Specifically, MISO remains high instead of returning data.

Hardware Setup:
  • PCB is long, with the MCU-to-last-LTC3300 distance being ~17 inches
  • CS pullup: 10K to MCU's 3.3V
  • MISO pullup: 2K to MCU's 3.3V
  • Observed ripple on MCU's 3.3V rail
Troubleshooting Steps Tried:
  • Lowered SPI clock rate
  • Silenced all other SPI activity on the board
  • Adjusted MISO pullup resistor values
Given the PCB length and power rail ripple, could signal integrity or power fluctuations be contributing to this issue? Are there recommended mitigation strategies for long daisy chains or power stability concerns?

Thank you for your support.

Thread Notes

  • Hello,

    Are you able to see the signals on an oscilloscope? It would be good to confirm that FF is actually being sent or if the MCU is not reading it for some reason. 

    Is there a top IC also in this application? If so is that IC also being communicated with?

  • We were able to resolve the issue. Sharing the resolution here in case it helps others who encounter the same problem.

    The root cause was that the voltage levels of the SPI signals going into the LTC3300 were not within specification. According to the datasheet, the voltage levels on the CSBI, SCKI, and SDI lines should be VREG - 0.5V. In our setup, the MCU pins were connected directly to these SPI lines, but the output voltage level from the MCU was only 3.3V.

    To fix this, we moved the SPI signals to 5V-tolerant pins on the MCU and added 1kΩ pull-up resistors from CSBI, SCKI, and SDI to VREG. After making this change, we were able to successfully communicate with all three LTC3300 chips.