Skip to end of metadata
Go to start of metadata

You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 8 Current »

The major differences between the LTC6813 and LTC6811 can be summarized in the table below:

System Level Changes

There will not be any major changes to accommodate for the switch to an 18 cell battery monitor. The AFE connects directly to the battery module connector boards through connectors and cabling. No modifications are necessary to the board but the harness will need to be changed. Figure 1 and 2 shows a block diagram outlining the communication and harnessing between the AFE boards, BMS carrier, and the battery pack.

Figure 1: LTC6813 Block Diagram

Figure 2: LTC6811 Block Diagram

The LTC6811 features 12 cell taps while the LTC6813 features 18. This means that our current design, which requires 36 taps, can be reduced from three boards to just two but the boards may need to be expanded to accommodate the balancing network. Currently it is sized to 100 x 70 mm. The SPI bus communicating with the MUX still makes use of GPIO 3, 4, 5, leaving 6 extra GPIOs.

Switching to two boards means that 30 thermistors will not be accounted for. We could accommodate this by utilizing two of the current ADG731 MUXs for a total of 64 inputs per board that would monitor 45 thermistor cells each. This means that MUX1 would monitor 23 cells while MUX2 would monitor 22 cells. The remaining inputs will be dedicated to at least three on-board thermistors to measure the balancing resistors. These would be placed at the ends and middle of the battery stack as a priority. The output of the MUXs can be connected to GPIO1 and GPIO6 of the LTC6813.

Alternatively, we could use an addressable format to control 4 16:1 MUXs simultaneously. This would also remove the requirement to invert the signal clock. Careful planning of the thermistor connections will be required to prevent mixing up readings. Four of this component can be used to achieve this layout which meets our specifications:

https://www.digikey.ca/en/products/detail/texas-instruments/cd74hc4067m96/1507236

The conversion times of commands, such as the ADCV and ADSTAT, from both datasheets were found to take similar amounts of time, with differences in the range of microseconds, so there are minimal latencies in comparison. No further firmware changes are required as the LTC6811 has cross compatibility with the LTC6813. The major changes to the firmware would be the addition of the Group E and F battery cell commands for reading and balancing and thermistors.

The daisy chain can be used for reversible IsoSPI but the benefit that this provides may not be helpful for our requirements This is because we would need to add an additional redundant path for the MUX which would require further modifications to the BMS carrier board, prolonging development time and debugging.

Cost

The cost breakdown will assume that the quantity of passive components of the balancing and voltage taps will remain the same. Both systems will require the same number of thermistors, balancing resistors, FETs, and associated capacitors. Therefore, the price difference for the entire system will only need to look at the price difference of the different parts of the block diagram, namely the voltage regulator, MUXs, connectors, and ICs.

Based off the following spreadsheet, the 12-cell layout will cost the overall AFE system approximately $234.51 while the 18-cell layout will cost $204.70, allowing us to save $29.81 on the entire system. If the 64:4 MUX layout is used, then the total costs excluding complementary components is $166.34. so savings can be at most $81.61. The cost breakdown does not include the connectors, though they are included in the spreadsheet, as final harnessing and the necessary connectors are undecided and can differ substantially in comparison to the current revision.

 Pros and Cons

Pros

Cons

LTC6811

  • Firmware development is mostly complete

  • More compact board sizes

  • Lower total measurement error (1.2mV)

  • More debugging and assembly required which can increase error likelihood

  • Increased isoSPI distances which can increase error, though impact may be negligible

LTC6813

  • Reduces the number of boards to two

  • Connections can be robust

  • Less assembly required

  • More GPIO pins available

  • Need to develop firmware for extra battery groups and MUX

  • Greater overall board size

  • Greater planning required for wiring

  • Higher total measurement error (2.2mV)

Overall, the LTC6813 can be a good alternative for the BMS AFE board for a future revision. It would introduce less error potential while also improving the costs by a considerable amount.

  • No labels