How to supply an isolated CAN transceiver from STSPIN32G4?

[mpf_gregory]

Dear all,

I'm planning to use an STSPIN32G4. Our product needs an isolated CAN transceiver. The EVSPIN32G4 offers a slot for a TCAN330, but this transceiver is not isolated. I could use an LTM2889, but it needs 140mA (225mA max) to drive the bus in dominant state. The STSPIN32G4's LDO provides max. 150mA. Of course the bus is not in dominant state all the time, so 150mA might be enough. Does anybody have experience with the power consumption of CAN transceivers? Do you think the STSPIN32G4's LDO can drive the LTM2889?

I appreciate your help.

Michael

[LCE]

Heat might be another issue.

If you're going to do some soldering / wiring with the LTM2889 anyway, another LDO / regulator wouldn't be that bad.

Get some 3-pin TO-220 LDO which can provide 500mA and attach to / a heatsink (5V * 200mA = 1W).

Or some 3-pin switching regulator (kinda 78L05 replacement).

[mpf_gregory]

You're right. Thanks or pointing out the heat issue. The LDO input voltage is 8V, so 4.7V must be dissipated as heat. This might be a problem if the full 200mA from the buck converter are consumed. In average the uC will draw 50mA and the CAN transceiver maybe also 50mA, so 0.5W of heat are generated by the LDO. That might be OK. But as you suggest it would safer to use another switching regulator connected to the battery instead of the STSPIN32G4. In this case it might be easier to use an isolated buck converter like the MAX17681 and an isolated CAN transceiver without integrated DC/DC, like the ISO1050.

But the question of the DC/DC output current remains. Can I get away with 100mA output current and a capacitor to handle current bursts in dominant state? Or should I provide 300mA of maximum output current to be on the safe side when transmitting in dominant state?

Maybe this questions would fit better in another subforum about peripherals, not about the MCU.

[LCE]

> Can I get away with 100mA output current and a capacitor to handle current bursts in dominant state?

Unless this is just for testing / playing OR you know exactly all the use cases for the product, I would always design to meet the maximum requirements, especially for industrial / automotive stuff.

[mpf_gregory]

True. But in this case it doesn't feel right to use a bigger DC/DC just to handle the high current demand for single dominant bits. The product sends out some CAN messages every 10msec, so about 99% of the time it's idling in recessive state anyway. But of course the product must not fail because of insufficient current supply, so I think I'll plan with another buck conveter connected to battery.

Thanks for your input, LCE!