• State
  • Locked Locked
  • Replies 2 replies
  • Subscribers 62 subscribers
  • Views 1056 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

BQ 78350 + AFE 76940 Inductive load SMbus communication problem

Matija
Prodigy 80 points
Other Parts Discussed in Thread: BQ78350

Hi all,

I encountered communication problems on SMBus during discharge with inductive load in the loop (motor driving). I'm using the EVM board and EV2300 with Battery Management Studio (BQ Studio). I have 2 EVM boards, both show the same problem.

When no or a continuous load (resistor bank) is applied, no communication problems occur, BQ Studio reads all registers and parameters. But if a motor is added as load, communication starts to break, the “No acknowledge” errors start to show up.

I've read on this forum of a similar problem, the answer was: "don't try to read too much data from it, because SMBus communication is not of highest priority for BQ 78350". Without reading any data at all, with just an idle connection to the gauge still produces the same result.

I measured the voltages on oscilloscope, to see how the noise is affecting 2.5V and battery voltage. There are higher spikes when the motor is running. Normal noise from 2.5V regout is 60mV p-p and with the motor it’s 350mV p-p. On battery the noise level is around 1V, but this is not so important. To check SMBus, on osciloscope are nice signals when no load applied, when motor starts, there is no extra noise on signals, but there is noticeable skew between SMC and SMD signals. This shows that noise doesn't directly affect (corrupt) signals, but chip itself with stream of data!

When the EV2300 is connected directly to the AFE (BQ 76940), with SCL and SDA jumpers to BQ78350 removed, the data in BQ769x0 Evaluation Software is read with no communication breaks, all measurements are stable. The problem doesn’t seem to be in the AFE part, only the gauge. 

On a custom board with this two chips and a host MCU, the MCU is working without a problem with this noise levels, but the communication with the BQ 78350 has the same problem. Does TI recommend connection from host MCU to AFE? 

Can anyone suggest a solution to this problem? Designing a low pass RLC filter on pack output is maybe one way to try silent the noise, but that shouldn't be so important part of BMS! Can't believe that BQ 78350 is so sensitive to the voltage noise in contrast to the other electronics (e.g. host MCU). 

Any suggestions are welcome! 

Best Regards,

Matija

 

  • 0
    TI__Guru 54715 points
    Certainly be cautious of the ground reference connections when operating a high current or inductive load.
    The AFE is expected to work with a MCU, but not if connected to the bq78350. Only one master should be on the I2C bus, connecting both a MCU and the bq78350 to the AFE will disrupt the bq78350.
  • 0 in reply to WM5295
    Prodigy 80 points

    Thank you for your response.

    AFE is always connected to gauge via I2C and MCU is connected to gauge via SMBus. AFE was connected to MCU directly only for test, because gauge wasn't working properly. So there is no case of both chips being connected to the same communication.

    About ground references...aren't the EVMs designed with proper ground references for this chips? Were EVMs ever tested with any kind of inductive load? This is not some high current condition, the motor current is up to 1 amper.