Actions
Forum Thread Details
  • Replies 6 replies
  • Subscribers 338 subscribers
  • Views 3835 views
  • Users 0 members are here
Related

NEED HELP with the ICM-20948 breakout board.

irs

Hi, I have created a ICM 20948 breakout. The schematic is almost similar to the one available online, the only difference is that I have only inserted the I2C pins as I was only using them for the communication purpose. Now, after getting the PCB printed and embedding the components onto it the board is not working. I used a multimeter to observer voltages on different points SCL and SDA which are used in I2C so I observed a voltage of 3.2V. I also  used the multimeter to check the continuity on the board and the sensor. So , continuity was well established. 

I also used a Logic Analyzer to observe the i2C signals but, there was no data available.

Now anyone who has created and assembled the board. Can you please help me out. Just please check my schematic if it's right or not and what possible 

  • I am not 100% sure but my gut tells me that the level shifters made from the BSS138LT1G Moffetts can only operate in one direction with the 1.8V logic signal at the source being the input and the 3.3V signal at the drain being the output.  If my hunch is true a signal like SDA that needs to be bidirectional won't work.  You also won't be successful in registering an interrupt on the 1.8V side of the Moffett.

  • in reply to dang74

    It looks like the BSS138LT's max Vgs-thres is 1.5 volts.So, it would probably be marginal for a 1.8 volt system, but without scope measurements we won't know for sure.

  • irs said:
    I also  used the multimeter to check the continuity on the board and the sensor. So , continuity was well established.

    For data signals like I2C, you really need to use an oscilloscope to see what is going on. A logic analyzer is useful to see if the signals are toggling but a scope will show you if there is a signal quality problem.

    You really need to see if both sides have activity and verify that both sides have a peak-to-peak voltage you expect.

    Schematic-wise, it looks like you did the level shifter correctly--assuming you schematic's pinout matches the pcb's footprint for the BSS138s.

  • Are you sure it's a working circuit? There's no decoupling capacitors near the ICM chip. Also, the ones near the voltage regulator are labeled 100pF. Are you sure that's correct?

    Photo of the assembled board would help too.

    Also, how does your chip-select work with that diode? Are you sure it will pull *CS lower than the input threshold (this is 1.8V logic..). Have you measured it?

  • in reply to baldengineer

    True.  Additionally, the only way the transistor turns on is when the source is made low by a signal from U2.  This satisfies the VTH with minimal margin as you point out.  When the logic level present at the source is +1.8V, the transistor is off.  To see why I think the circuit only works in one direction consider the case where the drain is made low.  The gate will still be +1.8V.  The source too is +1.8V by virtue of it's pull-up.  This keeps the transistor off, so the 0V at the drain does not make it through.  That I believe is the general flaw with the circuit.  My theory can be tested by applying 0V to the INT signal at the drain of U6 and see whether the signal at the source follows and becomes 0V as well or instead remains at +1.8V.

  • I can't load the image you have posted to have a look, unfortunately, it seems not working but I have designed an ICM20948-based sensor circuit and my PCB is working.

    Here is the diagram of the circuit.

    image