RS-485 Port surge immunity (IEC61000-4-5) protection circuit

I'm designing a RS-485 communication device for CE & UL standard certifications. My test standard is ±2KV (IEC61000-4-5). We have In-house lab test facility with ESD, EFT, Signal line surge generator, etc., For my various protection circuit, I can't able to pass the test more then ±1.5KV. Please review my circuit & share your commands.

The attached waveforms & test run procedures are below..,
2kv_sig_pos (2kv signal surge positive)
04.gdt_bu_2k (bourns GDT parallel with surge generator)[measured across A & GND point]
05.gdt_bk_2k (brightkings GDT parallel with surge generator)[measured across A & GND point]
06.gdt_bu_tvs_2kv1 (bourns GDT & SMAJ6.5CA TVS diode parallel with surge generator)[measured across A & GND point]
07.gdt_bk_tvs_2kv1 (brightkings GDT & SMAJ6.5CA TVS diode parallel with surge generator)[measured across A & GND point]
08.gdt_bk_tvs_2kv_ab (brightkings GDT & SMAJ6.5CA TVS diode parallel with surge generator)[measured across A & B point (differential)]
09.gdt_bk_res_tvs_2kv_ab (brightkings GDT, 10ohms resistor(between GDT & TVS) SMAJ6.5CA TVS diode parallel with surge generator)[measured across A & B point (differential)]

As per bourns "RS-485 PORT PROTECTION EVALUATION BOARD1" design notes, I changed my design as GDT, TBU(series), TVS. But this configuration also doesn't help me. The waveform similar to GDT+Resistor+TVS configuration.

Measurement tools are,
Oscilloscope (Agilent DSO-X 3014A, 100MHz)
Active differential probe (Pintek DP-150 PRO)
Surge generator (EMC Partner TRANSIENT 3000)
RS-485 Transceiver (MAX485)
GDT ( Bourns 2030-23T & brightking 2RL230M-6)
TVS (Vishay SMAJ6.5CA)

Regards
Udhay

You don't report test results.

FvM said:

You don't report test results.

Click to expand...

SN75176 IC failed, RS-485 Communication stopped. By changing the SN75176 IC alone, I can reestablish the communication.

Hi,

SN75176 IC failed, RS-485 Communication stopped. By changing the SN75176 IC alone, I can reestablish the communication.

Click to expand...

The SN75176 is damaged?

Can we see the PCB layout?
What data rate (baud rate) do you use?

With a profibus interface for CE and UL we use common mode data line chokes. I don´t see the surge current path in your application, so I´m not sure if this givex an improvement for your application.

VCC of SN75176 is on.y LC filtered. Check if the surge pulse causes ringing / overvoltage at the VCC line. (I rather see a (damping) RC than a high Q LC filter that causes ringing)

Klaus

KlausST said:

Hi,


The SN75176 is damaged?

Can we see the PCB layout?
What data rate (baud rate) do you use?

With a profibus interface for CE and UL we use common mode data line chokes. I don´t see the surge current path in your application, so I´m not sure if this givex an improvement for your application.

VCC of SN75176 is on.y LC filtered. Check if the surge pulse causes ringing / overvoltage at the VCC line. (I rather see a (damping) RC than a high Q LC filter that causes ringing)

Klaus

Click to expand...

PCB Layout attached (PCB ground plane & power supply hidden). Baud rate 9600.
Common mode choke can be used for EFT(capacitove coupling) test. I tested the circuit also with a common mode choke (Wurth electronics 744212100), but no improvment. The isolated 5v supply of RS485 section is powered directly with a DC/DC converter. So optioanaly I added a ferrite bead. I tested with & without ferrite bead, no change in output.

The next level I thinking of to put a high voltage RS-485 transceiver like SN65HVD1780 or LTC2862 & there are very expensive. SN75176 is very cheap price here. I'm looking for a protection circuit which is working with SN75176.

Expecting for any other circuit ideas. Also please verify my test circuit setup. I assume this CDN (coupling & decoupling network) network circuit is the suitable for RS-485 signal surge test. Please correct me if I'm wrong

Hi,

in your picture I can´t find the GND connections of all protection devices.
But this is very important to verify the current path.

Klaus

Klaus,
while I'm testing the RS485 surge, I seperate the protection device with a seperate PCB. The layout shown in the previous post is the original design, but not completed. before I start the actual PCB, I try to test the surge functionality seperately. I know that the ground return path impedance is not low, but this is a differential signal. Data can be acceptable even without the ground (DC ground - no earth connection). I'm connecting the protection devices across the device. Here I attached the picture of PCB used for testing

I found the solution for this problem. I replaced the SN75176(Ti) IC with SP485 (EXAR Corporation):thumbsup:. Working great :clap:. I tested it with the same setup for upto 3.5kv. No problem yet. The SN75176 & MAX485 failed at 1.5KV negative surge with the same test setup. I have test facility for surge upto 4.4KV. Today I will test it with upto 4.4KV or until damage the device & update the report.

The one thing I noticed while I testing, the waveform looks very different with very low in amplitude & the time extended very long. Loo at the attached waveform.

Today I will test it with TVS diode alone (without GDT).

Udhay