• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Answers 1 answer
  • Subscribers 33 subscribers
  • Views 432 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

SN65HVD62: SN65HVD62 on a 24VDC bus

R. E. Jozefzoon
Prodigy 120 points
Part Number: SN65HVD62

Hi,

I want to use several SN65HVD62's on a 24V bus, not the normal use case but.. This should  be possible since the usage of DC blocking capacitors.

This bus contains of one master and several slaves.

1. Where can I find information on the maximum number of SN65HVD62's I can place on the bus?

2. Where can I find the maximum cable length between SN65HVD62's?

Thank you

Roy

  • 0
    TI__Guru 62920 points
    Hi Roy,

    We don't have a lot of info on this since as you mentioned it is not all that standard of a usage. I agree it should be possible, though.

    In terms of the maximum cable length, you would need to know the insertion loss of the cabling at the AISG carrier frequency (2.176 MHz). The HVD62 transmitter will output a signal level of 5 dBm (minimum, assuming VRES = 1.5 V) and its receiver can detect signals as small as -12 dBm (maximum), so the difference between these two would give the loss budget for the cable.

    In terms of number of nodes, one of the limitations would be the loading of the driver caused by additional receivers placed in parallel. The minimum receiver input impedance is 11 kOhms, and each additional receiver will reduce this value. Reduced load resistance results in increased output current from the driver, which in turn means a reduced output voltage due to the finite output impedance of the driver (which is actually fairly small in the HVD62 itself, but the recommended implementation includes a series 50-Ohm resistance for impedance matching). So, on top of losses due to cabling, you should also consider the attenuation effect of this 50-Ohm effective output impedance driving the equivalent resistance of the parallel combination of receivers. For example, 10 receivers would yield an effective load resistance of 1.1 kOhms, which would result in 4-5% attenuation from a 50-Ohm source.

    Another limitation on the maximum number of connections would be the signal integrity impact of all the additional cabling to each node. This is something that may be harder to quantify without some prototype testing or simulations. In general, though, I would recommend keeping connections between each node and the main bus as short as possible so that the overall cable topology is linear (rather than something like a star network).

    Let me know if any of this is unclear or if you have other questions.

    Regards,
    Max
  • 0 in reply to Miguel Robertson
    Prodigy 120 points

    Hi Max,

    Thank you for your indepth answer. This is a good starting point for me to start my research and figure out if this is the right choice of component for me (I believe it is).

    Another question related to this is about addressing the node, giving them an unique address, in software rather than in hardware. I am not sure if I should start a new thread for this question.

    Do you know of a way to give the nodes addresses once installed in the system without giving them an address at front?

    For example you could first install  the master and then one by one install the nodes.  The  master "pings" for nodes. Once a node responds  to this ping it assigns it with an address. The next installed node(s) gets an updated  address.

    This is easy if you are able to install nodes one by one.

    If you power up the system with all nodes in place this way of assigning address to nodes will not work.

    Can you think of a way to assign address to  "blank" nodes? I don't want to program nodes during assembly because each node will get a dedicated location in the system and this is not known up front.

    Thank you,

    Roy

  • 0 in reply to R. E. Jozefzoon
    TI__Guru 62920 points
    Hi Roy,

    If the nodes are addressed automatically somehow, how would you know which node corresponds to which location in the system? (Or, is the location of a node not a concern here?) That's typically the issue that arises when trying to implement dynamic addressing. Typically you would need to either have a hardware-based identifier (e.g., DIP switches or high/low voltages on unused pins of a connector) or program the node with an identifier prior to installing it in the network, and in these cases you could use that unique identifier to address the nodes during serial communication.

    Max
  • 0 in reply to R. E. Jozefzoon
    Prodigy 120 points
    Hi Max,

    I think I found an easy way to assign an address to the nodes, by using a switch in a node.

    Upon activating the system all nodes are awaiting till their switch is pushed. Once it's switch is pressed it starts responding to the masters "ping" and gets an address. When the next switch (different node) in the system is pushed the process starts again, and gets an updated address. This way you know/the system knows which node is located where.

    Thank you for your support.

    Regards,

    Roy