THVD1406: RS-485 Communication Failure during Large Data Transfers in a Robotics Application

Charles (KAT) Zhang

Intellectual 1960 points

Part Number: THVD1406

Dear Team,

My customer Lenovo are developing an application for Robot and are using the THVD1406DR to implement a USB-to-RS485 interface (see attached schematic). This interface connects the robot's onboard MCU to a Battery BMS.

Issue: Short character transactions are stable. However, during large data transfers (e.g., 148KB 9600 bps. firmware download to the BMS), the communication fails with errors.

Questions:

Schematic Review: Could you please review the attached schematic? We want to ensure the biasing resistors, termination resistors are correctly implemented for stable long-term communication.

2. Signal Integrity (Direction Control): In the attached oscilloscope captures, They observed an ~8us spike in the A-B differential signal during transitions. Is this a normal characteristic of the automatic direction control feature in the THVD1406, or could it indicate a contention or improper switching timing that might cause errors during high-duty-cycle data transfers?

TX - UART_TX.

485 - A-B differential signals

Thank you

Charles

13 days ago

0 Charles (KAT) Zhang 13 days ago

TI__Intellectual 1960 points

Some figures are broken for some reason... re attach here:

0 Clemens Ladisch 13 days ago in reply to Charles (KAT) Zhang

Guru 327310 points

1. This side of the bus looks OK. Please show the schematic with the transceiver at the other end of the bus.

2. These spikes are normal. But the voltage after the spike (20 mV) might be too small for the other receiver.

0 Ethan White 13 days ago in reply to Clemens Ladisch

TI__Intellectual 2420 points

Hi Charles,

As Clemens mentioned this side of the bus looks ok, but sending the schematic of the other side of the bus would be helpful. Your failsafe biasing resistors, however, seem fairly large. Here's a resource we have on failsafe biasing: https://www.ti.com/lit/an/slyt324/slyt324.pdf.

The 8 us spike is exactly within the datasheet limits. During that time the THVD1406 is driving the signal as normal. After 8 us the auto-direction feature takes over and the device goes into high impedance mode, meaning the failsafe resistors would be taking over. See section 2.2 here: https://www.ti.com/lit/an/slla574/slla574.pdf

I agree with Clemens in that everything looks to be operating fine, you just have to play with the failsafe biasing a bit to get more reliable communication.

-Ethan White

0 Charles (KAT) Zhang 12 days ago in reply to Ethan White

TI__Intellectual 1960 points

Thank you,

Let me check with the customer

0 Charles (KAT) Zhang 10 days ago in reply to Ethan White

TI__Intellectual 1960 points

Hi Ethan, Clemens,

Please find the other side sch below

0 Clemens Ladisch 10 days ago in reply to Charles (KAT) Zhang

Guru 327310 points

The receiver also uses 4.7 kΩ biasing resistors.

The CA-IS3082 has fail-safe inputs that should work with a differential bus voltage near 0 V. If the transmission sometimes fails, then it's likely that there is noise on the bus. Try using stronger biasing resistors near the THVD1406 (470 Ω would result in a current of about 3 mA and an idle bus voltage of about 0.2 V).

0 Ethan White 10 days ago in reply to Clemens Ladisch

TI__Intellectual 2420 points

Hi Charles,

Thanks for sending the other side of the schematic. As Clemens mentioned you also have failsafe biasing on the receiver end. You should not have multiple failsafe biasing circuits within your network. I'd remove the failsafe biasing on the receiver end (R195,196) which should help with the drive strength.

-Ethan White

0 Charles (KAT) Zhang 6 days ago in reply to Ethan White

TI__Intellectual 1960 points

Thanks you Ethan and Clemens, I will check with customer on this

Interface

Interface forum

THVD1406: RS-485 Communication Failure during Large Data Transfers in a Robotics Application