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.

Working of P82B715 at Vcc=3.3V

Other Parts Discussed in Thread: P82B715, TCA9509, P82B96

Hi,

I want to use P82B715 at Vcc=3.3V . But, it is told that "full performance over temperature is ensured only from 4.5 V. Specification deratings apply when its supply voltage is reduced below 4.5 V." Please specify the derating applicable in this case of Vcc=3.3V.

Also, please explain what this statement means, "Bus logic-signal voltage levels are clamped at (VCC+ 0.7 V) but, otherwise, are independent of the supply voltage, VCC", means?

Regards,

Rasif

  • Hello Rasif,

    I am looking into your first question at the moment, and will have to get back to you.

    I can answer your second question. The internal ESD protection circuitry will trip at a diode drop above VCC, this causes Bus logic-signals to get clamped at VCC +0.7V (1 diode drop)
  • Hello Rasif,

    I have looked through the characterization data and you can see in the datasheet that the DC electrical characteristics are specced down to 3V or have another column for 3V.

    Another difference I saw in between the 3V and the 5V-12V is that the rise and fall time delays for the 3V case are close to the 5V-12V, and in most cases, the delays are shorter. All of these numbers are still within the spec on the datasheet.
  • Hi Jonathan,

    Thanks. We have made note of the DC electrical characteristics.

    So, is it that there's no de-rating as such, as far as the rise and fall time delays are concerned? If possible, could you please specify the difference in the values? These are not specified explicitly in the datasheet.

    Regards,

    Rasif

  • Rasif,

    Unfortunately, I cannot share characterization data. I can tell you that for the most part, the 3.0 V case was quite close to the 5V results, and from what I see, all of the 3V cases were within the timing specs listed in the datasheet. The difference in values has more to do with process variation, so any differences you see between different voltages will not be uniform across a larger sample.

  • Hi,

    I am using this device for taking the I2C bus from one board to the other through wire of 23/.2 gauge and 20cm length. How can the capacitance offered by the wire be calculated?

    When calculations are done, using available formula it is found to be around 10pF. Do we require the device to be used for such a small capacitance offered? The capacitance offered is required to calculate the pull-up resistance required to be connected on the buffered line as mentioned in the datasheet.

    Regards,
    Rasif
  • Hello Rasif,

    To calculate the capacitance of a cable, you will need to determine the capacitance per unit length (F/m typically). This of course matters more with cables that have 2 conductors (such as a coaxial cable, which has the signal in the center, and the ground around the outside), since you need a ground reference to have capacitance. Some cable datasheets will contain this spec, otherwise for single copper wires, I would suggest searching the internet for a round wire capacitance calculator, which can make some assumptions based on the input you provide.

    The capacitance of the cable is only 1 part of the system capacitance. You must also take into consideration the capacitance loading of the master and any slaves/devices connected to the node. A typical rule of thumb is 10-20 pF per device.

    To decide if a repeater is truly necessary in your application, you need to know your total line capacitance (slaves and cables) and your maximum frequency you wish to communicate at. Typically, for 400kHz operation, the limit is about 200pF if you are using pull up resistors (instead of current sources).

    Another thing to consider is, with larger capacitance, to maintain the same communication speed, pull up resistor values will have to drop to provide more current to pull the lines up faster, which negatively affects power consumption. 

    Some people wish to use repeaters for small loads because it lets them use high resistance resistors, and keep overall power consumption low, and provide for clean signals.

  • Hi Jonathan,

    This is the system where we want to use the I2C bus extender. Where, we intend to place the extender before the connector on Card 1 and after connector n card 2. Can you please tell if the extender is required in this case? The connector cable between both the cards is through normal copper wires.

    Regards,

    Rasif

  • Rasif,
    You will almost certainly need an extender for 20m of cable. This is because the capacitance will likely be much larger than the 400 pf limit.

    I would suggest double checking your capacitance measurements, as I find it unlikely that 20m of any cable will keep you under 400 pf.
  • Hi Jonathan,

    The length is 20 cm. It is not 20 m.

    Regards,
    Rasif
  • Rasif,

    For 20 cm of length of standard cable, you can probably get away without the use of a repeater, or with a repeater designed for shorter runs (such as the TCA9509)
  • Hi Jonathan,

    The cable used here is not of standard type, but the normal multi-stranded insulated copper wire of 23/.2 gauge. For, which the capacitance for a length of 20 cm is unknown.

    Regards,
    Rasif
  • Rasif,

    In order to accurately determine if you require a repeater, the capacitance will need to be tested.

    However, If I were to guess, you will likely need some form of repeater for a 20 cm cable like you describe. The P82B96 should be able to handle this load (assuming total capacitance is below 4uF, which it should be), but I cannot accurate access if the repeater can be avoided in this application without capacitance of all nodes.

    The repeater would be the 'safest bet' in a situation like this. 

  • Hi Jonathan,

    Thank you for the valuable suggestions. We are going forward with P82B715 as of now, on a trial mode. 

    Regards,

    Rasif

  • Rasif,

    Great to hear! Please let me know if you have any further questions.