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.

P82B96: Clarification for the use of P82B96 for Single Master (MCU) & multiple slaves (Sensor) configuration

Part Number: P82B96
Other Parts Discussed in Thread: MAX232, AM26C31, AM26C32, SN65LVDS049

Hello Community,

The P82B96 datasheet suggests some configuration for I2C communications, in this, we wish to seek clarifications for the following queries:

Query 1: Configuration

We seek clarification on whether we can use a single pair of P82B96 I2C extenders to interface multiple Slave devices (Sensors) on the remote slave bus with a single master (MCU) on the local master bus with P82B96 transmitting signals over the buffer expansion bus.

The maximum required speed is 400kHz for our use case.

Query 2: Differential Drivers

We need to transmit the data over 3m distance using CAT5e cables. Do we need differential drivers in this case?

If at all required, for increased noise immunity of the I2C bus, can we use following two differential drivers & receivers:

  1. MAX232 Differential driver & receiver pairs alongside P82B96 at both ends with 120Kbps
  2. AM26C31 & AM26C32 with 10Mbps

Are there any known obvious factors that we may have overlooked when selecting the differential drivers for P82B96?

Kindly provide your insights on these assumptions.

Thank you.

  • 1. The number of masters or slaves makes no difference.

    2. CAT5e cables are designed for differential signals. When you are using them with single-ended signals, avoid crosstalk by pairing the signals with power lines (see figure 13). For noise immunity, you would indeed require differential drivers/receivers. The MAX232 is not differential (RS-232 uses single-ended signals at higher voltages). RS-485/422 is designed for industrial environments and uses more space and power. Consider using LVDS; there are 2+2 transceivers like the SN65LVDS049.

  • Hello ,

    We have AM26C3x for our SPI bus, we can use those with P82B96 as well, right?

  • Yes. But you would not need four channels per direction for one I²C bus.

  • Yes, 2 channels on each IC will remain unused. We have these ICs in stock & they are affordable, which drives our decision to use them.

    Also, on a side note, for better noise immunity (margin) can we use the P82B96 simply by increasing the buffered expansion bus voltage to 9V range? This will provide a significant noise margin as is, this is the right way to think about this concept, right? Assume power is not a concern in our use case.

  • To change the I²C signal, the noise would have to increase/decrease the current through the pull-up resistors. You do not need to increase the bus voltage to increase the current; you can just change the resistors.

  • Hi Neet,

    Query 1: Configuration

    We seek clarification on whether we can use a single pair of P82B96 I2C extenders to interface multiple Slave devices (Sensors) on the remote slave bus with a single master (MCU) on the local master bus with P82B96 transmitting signals over the buffer expansion bus.

    The maximum required speed is 400kHz for our use case.

    I agree with Clemens here. Only other thing to consider is the 400pF limit on the Sx/Sy input/output pins. Rise-time, IOL, and total parasitic bus capacitance would determine the maximum number of master/slave devices that can be present on the bus. 

    Query 2: Differential Drivers

    We need to transmit the data over 3m distance using CAT5e cables. Do we need differential drivers in this case?

    If at all required, for increased noise immunity of the I2C bus, can we use following two differential drivers & receivers:

    1. MAX232 Differential driver & receiver pairs alongside P82B96 at both ends with 120Kbps
    2. AM26C31 & AM26C32 with 10Mbps

    Are there any known obvious factors that we may have overlooked when selecting the differential drivers for P82B96?

    Kindly provide your insights on these assumptions.

    I agree with Clemens here as well. Another thing to note is CAT5e cable can be approximately 50pF / m. This would mean that over 3 meters the TX/RX pins of the P82B96 devices would need to drive this load. This is doable for this device.

    Differential drivers will have higher noise immunity than I2C. As Clemens already stated, MAX232 is not differential but point to point. 

    Also, on a side note, for better noise immunity (margin) can we use the P82B96 simply by increasing the buffered expansion bus voltage to 9V range? This will provide a significant noise margin as is, this is the right way to think about this concept, right? Assume power is not a concern in our use case.

    Making VCC = 9V and having the Tx/Rx/Ty/Ry lines pulled up to VCC would increase the input threshold margins since the thresholds are based off of VCC. 

    Increase in VCC would created larger margins for the high and low logic levels. This will effect the over rise/fall time on the buffered side, but with I2C speeds this is really not a problem (longer time to charge up to 9V VCC takes longer time than say 5V, falling from 9V to VOL takes longer than 5V to VOL). 

    Regards,

    Tyler