SN65HVD10: SN65HVD10DR Fail safe resistor

Hi Dattatray,

I'm not able to see the schematic you shared. Could you please try including this again? You may have to use the "Insert/Edit Media" function, copy and paste generally doesn't work.

External fail safe networks are used to bias an idle bus to a known state when no transceiver is actively driving a signal. Such a network can be used to bias the state of the entire bus and therefore effectively keeps every transceiver in known state when idle. If this implementation is used, a single circuit may be sufficient for the entire network, meaning fail-safe resistors need only be included at one node. An example of this layout is described in the following app report.
RS-485: Passive failsafe for an idle bus

Modern RS-485 transceivers such as SN65HVD10 often have fail-safe features of their own which allow the receiver output to remain in a known state even when the bus is open, shorted, or idle. Because of this, external fail-safes are not necessary for typical functionality, but instead offer increased system resilience against noise in an open, shorted, or idle bus state.

Is this information what you are looking for? If you'd like to share your schematic again, I will be happy to comment on the use or need for external fail-safe biasing resistors.

Regards,
Eric Schott

Hi Eric,

Thanks for prompt reply. I have below queries.

1. In Slave use, if i used external fail safe resistor on all 32 devices, does it work ?

2. In Master use , if i used external fail safe resistor on Master & 32 slave devices, does it work ? 

Please find attached screen shot of RS 485 section 

Thanks,

Dattatray

2376.RS485.docx

HI Dattatray,

Including this fail-safe circuitry on every node in this system will likely overload the bus and make it impossible for the transceivers to drive the opposite polarity logic. With 32 1k-ohm resistors, this would equate to a 32-ohm path to ground (or 3.3V) which would essentially act as a weak short on the bus. 

If it's necessary to include fail-safe circuits at every node, I would recommend using the equations in the document I shared dto calculate the theoretical resistor value needed for the system. Once found, the resistors put in place at each of the 32 nodes should be valued so that the equivalent resistance equals this theoretical value. For example, the document calculates a Rfs value of 556-ohms for its sample setup. If we were to spread this value over 32 nodes in the system, each node would use ~18k-ohm resistors (556 * 32 = 17792). 

A simpler solution would be to only include the failsafe circuitry on a single node and use the calculated value for these resistors. This is the more common option as it has a lower system BOM count and allows any number of nodes to be used without needing to recalculate the resistors needed for each. 

Regards,
Eric Schott

Dear ERIC,

Thanks.

As per your suggestion its very difficult to add Fail safe circuitry on a single node. Because during production either i have option of solder fail safe circuitry or not. During deployment its difficult to trace.

Please give me solution such that with Fail safe circuity it should work in master & as well as in slave mode up to 32 slave. If someone connected 5 slave in that case also it should work & also with 10 slave...20 slave...25 slave...32 slave it should work. Also if same slave is used as Master it should work with any number of slave up to 32 devices.

Regards,

Dattatray

Hi Dattatray,

As I said above, the simplest solution is likely to only include the fail-safe circuit on a single node. If it is guaranteed that this system will always have one and only one master node, the circuit can be included here and on none of the slaves. Is this solution possible?

Unfortunately there's no elegant solution if all slave devices need to include the circuit as well if the number of total nodes is unknown. If each circuit needs to be strong enough to bias the system in the case where only 1 node is present, it will present too much load to the system in the case where 32 nodes are present. 

A secondary solution may be to include very weak biasing circuits at each node that will only work to bias the local transceiver to a known state during idle conditions. This would allow the transceiver to handle the idle bus case, but would not be as effective for fault cases. This is also commonly integrated into newer transceivers (including SN65HVD10) so a local biasing circuit may be redundant. 

Regards,
Eric Schott