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.

DS90LV804: Behavior with Contention?

Part Number: DS90LV804

Hello, I am trying to assess the behavior of contention between sets of LVDS outputs for the subject-line part.  To clarify, I mean the following:

  • Two separate boards, each with a DS90LV804 installed.
  • Each board has a common ground reference.
  • One set of OUT pins on one board gets connected to a set of OUT pins on the other board, while powered.

I see the short-circuit rating for this part, however, that is specified to ground, which is not the case here.  I'm thinking that contention between two sets of LVDS-driver outputs will not result in more than a 3.5mA draw between one or the other, and obviously the link will not function correctly.  However I'd like to confirm for this specific part. 

I'll add that this is not a normal operating mode for our design, but it is a possibility if a set of connectors in our system is mis-mated.  In our current situation, changing the connector pinout or connector type is not particularly simple, so I'm asking this question to assess the consequences of a mis-mated connector.  If it's simply loss of functionality, and no parts stress, we'll probably leave the design as-is as it's easily remedied.  

Thank you.  

  • Hi Josh,

    Check into using the tri-state feature. Tri-state buffers allow multiple logic devices to be connected to the same bus without damage or data loss. Contention arises when a device on a bus wants to drive a signal high, while another is driving low. If these devices send and receive data at the same time, a short circuit may occur if one is driving high to logic '1' (say to the supply voltage) and the other is driving low to logic '0' (say to ground). In the case of a short circuit, the device is rated to an absolute maximum of 90 mA, but the tri-state can be used for isolation from these cases.

    Regards, Amy

  • Hi - thank you for the response, and while I agree that tri-state helps prevent contention, it doesn't apply to this specific case.  

    However, your response implies that in normal operation of LVDS (i.e. a transmitter driving a receiver), it's the receiver (and not the driver) that limits the current to 3.5mA.  Is that correct?  

  • Hi Josh,

    Can you let me know how was the estimate of 3.5 mA was made? Is there a reason the Tri-state function cannot be used?

    Regards, Amy

  • How did I arrive at 3.5mA: I came up with 3.5mA because that is the value of the current source used for LVDS signaling.  My thinking here is that if the current source is located in the driver, that opens the possibility that two drivers shorted to each other could not produce more than what the current source can provide.  

    On tri-state:  it's not that we can't use tri-state, it's that it wouldn't solve this particular problem.  The concern that prompted my inital question is someone mis-mates the cables in our system, and we don't discover the mis-mate until the system is powered up and someone attempts to operate it.  At that point, the parts are no longer in tri-state. 

  • Hi Josh,

    The output pins were designed and optimized as outputs, so it is not recommended to use the device in situations with direct input. This device does have some degree of protection against events like ESD, but the only circumstance characterized for this device was the output short circuit current.

    Regards, Amy