TCAN4420: TCAN4420

Hi Adam,

Thanks for the question. I will get back to you tomorrow with some suggestions.

Regards,

Eric

Hi Adam,

For CAN networks with stubs like the ones you described, I would recommend TCAN1042 (standby mode with remote wake) or TCAN1051 (silent mode). Both are high-speed CAN transceivers with CAN FD and up to 5Mbps supported data rates. By internally suppressing ringing after a dominant period, these devices can operate well on buses where  heavy loading or reflections from stubs could interfere with communication. 

To dampen reflections in your network, I would recommend possibly loading all three nodes with 180-ohm termination. This will still result in 60-ohm load on the bus and allow dampening at all symmetrical branches on the network. 

We demo-ed these devices at the Electronica fair in 2016 with 20 nodes on a 40 meter network with stubs with varying lengths. The longest stubs in this demonstration were 2 meters. 

https://can-newsletter.org/hardware/semiconductors/170420_can-fd-demonstrator-with-20-nodes-at-a-40-m-network_ti/

With your desired network with only three nodes, the topology could take many forms and still achieve a reliable high-speed data rate. Star typologies are often problematic, but this system seems small enough that it should not introduce issues with newer devices. 

Let me know if you have any other questions.

Regards,

Eric

Thank you Eric for explanation.

TCAN 1042 would be suitable choice for me.

I've got some more questions:

Shouldn't it be 3x 360-ohm to have the 120-ohm differential resistance?

Is the below topology also correct (maybe better?):

Best regards,

Adam

Hi Adam,

Our CAN drivers expect two 120-ohm loading on both end-points of the bus. These 120-ohm termination resistors match the impedance of the bus and give a total loading of 60-ohms (two 120s in parallel). In the star case, adding some termination (though not perfectly matching cable impedance) helps reduce the reflections, but we also want to keep the total bus loading to 60-ohms. 180-ohm resistors would do a better job reducing reflections than 360-ohms because it is closer to the cable impedance. 

The daisy chain topology looks preferable. Without stubs, nearly all reflections can be effectively dampened by the 120-ohm termination the the end-points. This implementation would involve more cabling, but would be better for signal integrity. 

Regards,

Eric