SN65HVD3082E: RS485 SN65HVD3082EDR undershoot while TX enable

Artur,

Thanks for your information. Unfortunately the device behaves this way during mode transition in some conditions. The reason is that the driver circuitry takes some time to operate. You can see that when the control signal goes high, there is a very short period both A and B going low, which means the pins are not active yet. What's the edge rate of the control? Sometimes the undershoot disappears with a different rising/falling time. 

Regards,

Hao  

Hello Hao,

thanks for the fast response. The edge ist relatively slow because the signal is isolated with the optocoupler with a phototransistor output. The edge is about ~10 µs, i didn't measure it exactly.

In the 2nd figure the edge is slowed down by changing some parameters in the optocoupler circuit. The undershoot was still there.

If this behavior doesn't influence the communaction negatively I wouldn't change the circuit anymore. Can you agree?

Best regards

Artur

Hello Hao,

thank you for your help. 

I found the THVD1500 interesting. Its seems to be cheaper and better in all parameters. Its better for noise rejection, doesnt need bias resistors etc. etc.

I would prefere to use THVD1500 after I found this. Are there issues why not use THVD1500 instead of the SN65HVD3082E?

best regards

Artur

Hi Max,

now i have a board with THVD1500 and i measured another issue. I can see an overshoot while tx ist beeing disabled after sending the data. In this configuration the 120R termination is also active.

I tried different falling edges but it doesnt matter, the overshoot is still there. Is it something THVD1500 specific? Do you have any Idea how to remove that?

In the following again i measured A and B Signal of the SN65HVD3082E and THVD1500 while disabling tx. The SN65HVD3082E do an undershoot and THVD1500 an overshoot.

THVD1500

SN65HVD3082E

Hi Arthur,

I'd like to understand this a little better.  Could you share your schematic of the RS-485 implementation?  I'm wondering what sort of loading may be present on the A/B bus pins and how the logic controls (D/DE/RE) are handled.

Is the behavior observed when testing a node individually or only when multiple nodes share common bus connections?

Have there been any schematic changes when moving to the THVD1500 device?  (I was confused since the previous SN65HVD3082E waveforms didn't show this undershoot and they seemed to have a differential bias applied during the idle state.)

Regards,
Max

Hi Max,

in all my measurements i had my RS485 device (TX) connected to a commercial RS-485 device (Rx) with 120R termination at the bus end. The cable length is~1m. I also tried to load the bus just with 54R and 120R, without a 2nd device. 

The logic controls (D/DE/RE) are handled by a host processor. The main change in the schematic when I moved to the THVD1500 was removing the 1k bias resistors on A/B bus pins because the THVD1500 has fail safe function.

In the last two scope screenshots i just soldered jumper wire to the SO8-packages of the THVD1500 and SN65HVD3082E. I connected the transceiver to a stable 5V power supply and an USB-to-Serial Adapter so i was able to send some data. The screenshots were taken while i was manually switching the DE/RE from 5V to GND.

best regards 

Artur

Hi Arthur,

Where did you probe the enable and bus signal when you captured the shots? Can you try bypassing the common mode choke and adding some resistive load between A and B? It looks you might have the load, but I didn't see it on the schematic. I was wondering the inductive loading makes the overshoot.

Regards,

Hao 

Hi Hao

i probed the A/B signals after the chocke at the bus connector.

The last two scope screenshots have nothing to do with the schematic. That was like on a breadboard, where i only had the transceivers +120 Ohm load, power supply, and usb-uart-adapter.  So it cannot be the inductive loading making the overshoot.

Without the termination load the signal would look like that:

best regards 

Artur

Artur,

Thanks for your clarification. When you disabled the device, was the D pin still sending data? Have you ever seen undershooting from THVD1500 and overshooting from SN65HVD3082E if you toggled the DE/REB pin randomly? Have you ever tried with keeping the D low/high?

Regards,

Hao

Hi Hao,

I also see the overshoot/undeshoot while toggeling randomly DE/RE and i also tried to fix RE to VCC and only toggle DE, there was still the same issue. I also tried keeping the D low and high, nothing changed! It doesnt matter sending data or keeping the D pin low/high. Maybe it can be a measurement issue i dont know.

If i keep the DE high and just sending data everything is clean, no issue at all.

Artur,

I will do some bench testing to see what the possible issue was. I will update you sometime next week.

Regards,

Hao

Artur,

I did some bench testing on THVD1500 and SN65HVD3082E by toggling the DE/REB pins (with data on the D). However I didn't see the behavior you described. Here are the results for your reference.

THVD1500

THVD1500 (zoom in)

SN65HVD3082E

SN65HVD3082E (zoom in)

Can you check your setup one more time, like the connection, supply and ground? Do you install a 0.1uF cap between Vcc and ground? Does the supply have any transient current limiting?

Regards,

Hao