TCAN1043A-Q1: Short-to-battery Termination Resistor Power Dissipation

Jon Knieper

Part Number: TCAN1043A-Q1

Tool/software:

I have a high-level question that I'm hoping you can shed some light on. We've been discussing CAN short-to-battery scenarios recently for a 48V program, and I've hit a bottleneck not with the TCAN104x CAN transceivers themselves (58/70V bus fault rating), but with the termination resistors. With multiple nodes on a CAN bus, I calculate the power dissipation through these resistors easily ends up being 16+ watts, which there is no way that can be meet in application. So I feel like I must be missing something.

Power Dissipation:

There are a couple of ways I'm looking at this. First, considering just the differential voltage across the CAN lines:

Given a short to 48V on CAN H or L, the voltage across the termination network would be approximately 48V−2.5V=45.5V.

With two 60Ω termination resistors in series, considering each resistor individually:

Per 60Ω resistor: W=V2/R=(45.5V/2)^2/60Ω=(22.75V)^2/60Ω≈8.6W @ 25°C

Factoring in temperature derating for 125°C (45% derating): W@125°C=8.6W×2.2=18.9W per resistor.

Adding a 10% margin: W=18.9W×1.10≈20.8W per resistor.

Alternatively, if I consider the CAN transceiver short-circuit output current, using the dominant current of 115mA, and factoring in just two CAN transceivers driving the bus dominant:

P=I^2×R=(2x115mA)^2×120Ω=(0.230A)^2×120Ω≈6.35W (16+W when factoring in temp and margin like above)

Do you have any insights into this problem? My hunch is that if the CAN transceivers on the bus all enter a recessive state during a short-to-battery event, this would drastically limit the current through the termination resistors. If this is the case, understanding the typical time duration a transceiver might remain in the dominant state during such a fault would be very helpful to know.

Any information you can provide would be extremely helpful and greatly appreciated.

5 months ago

+1 Eric Schott1 5 months ago

TI__Guru 53555 points

Hi Jon,

This is an excellent question and one with a more complex answer than it seems on the surface which you've already seemed to dive pretty well into.

You're correct that in the worst case conditions the termination resistors will be dissipating quite a lot of power - in most cases much more than you'd typically rate these components for. Luckily the CAN system has several safety mechanisms in place to help protect this hardware from any damaging conditions. The fist of which is the current limit on the transceivers which you've identified. The vast majority of time there will only be a single transceiver attempting to drive the bus dominant so the 115mA is going to be our typical fault current in a bus short-to-battery condition. In the cases where multiple devices are driving will be limited to arbitration or ACK bits which is a small portion of the total frame and would be a short enough duration that we can treat it as marginal additional heating. The rest of the heating cana be calculated by averaging the dominant-time of the CAN signal as a percentage of the total transmission time (see linked posts for detailed calculations).

The second safety mechanism we can count on is the CAN protocol controller recognizing strong bus faults based on data recognition. When a driving CAN controller drives its TXD signal dominant, it also checks the resulting RXD state to ensure that the dominant is transmitted correctly. In the case of a strong short-to-battery condition it is more than likely that the controller will see a discrepancy here and log it as an error and abort the current frame transmission attempt. This will quickly lead to a bus-off condition where the controller stops driving the transceiver and leaving the bus idle so no more current will be drawn through the termination resistors.

The last fail-safe for this condition would be the transceiver's thermal shutdown limit. As a large portion of the heating in a short-to-battery case will occur in the transceiver itself the thermal limit of the transceiver can be reached before the external components which will shut down the transmitter and stop the system heating.

This is all high-level as you mentioned - here is another thread with some similar descriptions for variety of explanation. For more detailed analysis see the second linked thread.

TCAN1042H-Q1: The power rating needed for the termination resistor?
Wattage Requirements for CAN-Bus Terminating Resistor with SN65HVD233

Let me know if you have any other questions or would like to discuss any specific portion of this topic and how it relates to our devices.

Regards,
Eric Schott