TCAN1042-Q1: VIO control by microcontroller GPIO

Hi Hedy,

While the Vio pin on TCAN1042 only draws a maximum of a few hundred uA to power the device itself, this supply is also used as the pull-up source for external digital logic (RXD pin). This means that all current through the RXD pin to drive the receiver's output signal will also be sourced to Vio and could potentially exceed the limits of the microcontroller's GPIO driving capabilities. While it is possible that with low signal rates and minimal trace capacitance, this would not cause an issue, I don't believe that this design would provide a resilient solution. 

If the power-off high impedance state is required for this design, my recommendation would be to use the MCU's GPIO pin to drive a FET to switch power to TCAN1042's Vio. While this does add an extra component, the FET would relieve the MCU from sourcing this current and ensure that Vio can be adequately supplied under a larger range of RXD conditions. 

For those that have not read our previous conversation, I will note that TCAN1042's CAN pins will enter a high-impedance state when in standby mode with an input resistance around 25k-ohms (see "input resistance (CANH or CANL)" on datasheet). This state is designed to present minimal load to the bus as to not disrupt other communications while still allowing the device to monitor the bus for a wake condition. In your design, are you trying to completely power off the device because the standby impedance too low? This method would indeed increase the input impedance of the device, but it would no longer be able to monitor the bus for a wake signal and it would add complexity by requiring Vio switching capabilities. 

Regards,
Eric Schott

Hi Eric,

Thanks your great reply.

We need to power transceiver off and force CAN BUS be high impedance (~Mohm) because in our application, some old vehicle cable pin define is not high speed CAN bus. In that case, there will be voltage level on TCAN1042 CAN_H & CAN_L, as well as draining current. That will cause some error message and abnormal performance. This is we want to avoid.

Could you advice us, if TCAN1042 connect to non high speed CAN bus, how to avoid data transfer or communication? My idea is made transceiver enter "protected mode" (VCC=0V, VIO=3.3V, STB=0)

From your experience, if non high speed CAN (such as J1708) connect to high speed CAN, would damage transceiver?

Thanks.

Hedy

Hi Hedy,

I understand the concern of connecting an active CAN transceiver to a non-CAN line. I believe this may cause an issue if the transceiver is fully enabled and biasing the bus to a recessive level (~2.5V). However, when this recessive biasing is disabled (such is the case in protected modes), and the transceiver only presents the k-ohm impedance load, I don't believe this would cause an issue in such a system. If we estimate this input impedance simply as a resistive path to ground, it would only allow a small leakage current through the device in this state (Vcc = 0, Vio = 3.3, STB = 0) and would have minimal impact on the signal voltage. A far greater load will be presented by the termination network that exists between CANH and CANL.

In my experience, I have not seen a high-impedance (no recessive biasing) CAN transceiver present an issue in this way, though I must admit my exposure to such implementations is limited. I do not see how such a connection could risk damage to either the CAN or other driving devices if the transceiver is left in a protected mode. Depending on the drive capabilities of the other signal drivers, I also expect that the CAN transceiver will also not have a great impact on other communications on these lines. 

Regards,
Eric Schott

Hi Eric,

Can you provide the VIO current consumption? including drain current to TXD/ RXD/ STB at active mode.

Since I cannot get more information from datasheet. Thank you.

Best Regards,

Hedy

Hi Hedy,

The datasheet specifies the current draw from Vio in these modes when RXD is left floating. This specification includes current consumption from the TXD threshold detection circuitry and the standby/active modes of the device. Both of these limits are in the uA range. The absolute maximum specification for current draw through the RXD pin is defined as +-8mA. This current will be sourced from the Vio pin on top of the current needed for other functions. 

Because this supply pin has the ability to draw more current than it appears the microcontroller GPIO pin can supply, I don't recommend this solution without extra design considerations. If the current through RXD can be limited (ensuring leakage current through the MCU input pin and AC current from trace capacitance do not exceed the ~3.0mA) by limiting line capacitance and possibly including series resistance on this line, I believe the solution could be viable. 

Regards,
Eric Schott