RS485 to USB communications

Thread Starter

tonyg1097

Joined Jun 26, 2023
3
We have a very simple RS485 to USB cable connection on our automotive inspection equipment.
To sum up the issue, we are having drops in communications between our board and any laptop that is charging!
No failures when the laptop is on battery. No failures when using a Desktop PC. We use a half-duplex configuration
and a max cable length of 20 meters, but our 8 meter cables seem to fail more often for some unknown reason.

Here is the microprocessor side. Cable connects to the 3-pin connector.
1687800838678.png

Here is a shot of the cable. J1 connects to the 3-pin above. Black and Red are the diff pair. White is Ground.
Drain wire is the shield. Cable shield is grounded to the USB case. No schematic for the dongle circuitry, although I
do see a micro and an RS485 converter IC.
1687801196828.png
We are starting to think it is a Ground issue - noise or Ground loop or something similar.
Thanks in advance for the help.
 

MrChips

Joined Oct 2, 2009
29,855
RS-485 is a balanced complementary transmission line. It requires two things, (1) proper cable termination, and (2) proper DC bias.
Make sure that both of these items are installed.
 

nsaspook

Joined Aug 27, 2009
12,320
Sounds like the common-mode voltage of a transceiver is being exceeded during the charge cycle.
https://ez.analog.com/ez-blogs/b/en...common-mode-range-rs-485-and-can-transceivers
Common mode voltage—it can be a scourge for serial communication systems such as RS-485 or even CAN communications in noisy industrial systems. High common mode voltage can be due to an electrically noisy environment or nearby equipment, such as an AC motor, transformers, or power cables. Or, it could stem from the voltage differences in ground planes; for example, two nodes or systems that operate from different transformers or ground levels. These voltage differences occur because cabling can be hundreds of feet long, especially in RS-485. For example, a 1000ft long 24 AWG cable (25.67 Ohms per 1000ft) carrying 450mA of current can cause a difference in ground potential of 11.6V. It's easy to see that in a large factory or building, such voltage differences can be quite common. So whether the high common mode voltage is due to noise or ground level voltage differences, it can hamper reliable communications. In an application area like an automated factory, such errors can mean lost uptime and productivity.
...
In serial communications systems, the RS-485 protocol defines common mode range (CMR) for the receiver side (-7V to +12V). This is the voltage range that allows messages to be received. But the RS-485 standard doesn't have such a definition for the driver side. This means that messages can't be sent if the common voltage level in the system goes above the transceiver's threshold.

What if we had a larger ground potential difference than the 11.6V in the example above? Most standard RS-485 transceivers would fail to communicate. The larger the CMR the transceiver can accommodate, the more reliable the system becomes. Thus, it behooves the designer to select components with the widest CMR
The ISL3175EIBZ CMR is the standard -7V to +12V common-mode input/output voltage range. That voltage can be reduced by adding a common ground between devices or using upgraded transceiver chips like this.
https://www.analog.com/media/en/technical-documentation/data-sheets/adm3095e.pdf

Certified Level 4 EMC protection on RS-485 A and B bus pins
IEC 61000-4-5 surge protection (±4 kV)
EC 61000-4-4 EFT protection (±2 kV)
IEC 61000-4-2 ESD protection
±8 kV contact discharge
±15 kV air gap discharge
Certified IEC 61000-4-6 conducted RF immunity (10 V/m rms)
RS-485 A and B bus pins HBM ESD >±30 kV
±42 V ac/dc peak fault protection on RS-485 bus pins
TIA/EIA RS-485 compliant over full supply range
3.0 V to 5.5 V operating voltage range on VCC
1.62 V to 5.5 V VIO logic supply
Common-mode input range: −25 V to +25 V


 

dendad

Joined Feb 20, 2016
4,375
Is the signal GND a solid connecting between the laptop and the equipment? It does sound like leakage from the mains through the laptop power supply is the problem. This is fairly common. Just for a test, plug in the laptop and measure the AC voltage from the laptop GND to the test equipment GND (without the USB connection).

Using an isolated USB module in line may help.
1687815714855.png
Ebay item shown above.
 

Thread Starter

tonyg1097

Joined Jun 26, 2023
3
Sounds like the common-mode voltage of a transceiver is being exceeded during the charge cycle.
https://ez.analog.com/ez-blogs/b/en...common-mode-range-rs-485-and-can-transceivers


The ISL3175EIBZ CMR is the standard -7V to +12V common-mode input/output voltage range. That voltage can be reduced by adding a common ground between devices or using upgraded transceiver chips like this.
https://www.analog.com/media/en/technical-documentation/data-sheets/adm3095e.pdf

Certified Level 4 EMC protection on RS-485 A and B bus pins
IEC 61000-4-5 surge protection (±4 kV)
EC 61000-4-4 EFT protection (±2 kV)
IEC 61000-4-2 ESD protection
±8 kV contact discharge
±15 kV air gap discharge
Certified IEC 61000-4-6 conducted RF immunity (10 V/m rms)
RS-485 A and B bus pins HBM ESD >±30 kV
±42 V ac/dc peak fault protection on RS-485 bus pins
TIA/EIA RS-485 compliant over full supply range
3.0 V to 5.5 V operating voltage range on VCC
1.62 V to 5.5 V VIO logic supply
Common-mode input range: −25 V to +25 V


thank you! This is very helpful.
Is the signal GND a solid connecting between the laptop and the equipment? It does sound like leakage from the mains through the laptop power supply is the problem. This is fairly common. Just for a test, plug in the laptop and measure the AC voltage from the laptop GND to the test equipment GND (without the USB connection).

Using an isolated USB module in line may help.
View attachment 297143
Ebay item shown above.
there is no connection between the ground at our board and the ground at the laptop. there is a 1meg resistor between the "Shield" pin on our board and ground.
1687815921077.png
 

dendad

Joined Feb 20, 2016
4,375
You need the GND connected. That is the problem. The common mode voltage of the drivers are being exceeded by the AC leakage.
Years ago, when I first started in my industrial control profession, we installed an RS485 network on a wire drawing factory. And as it was our first RS485 project, we too thought it was "A two wire network" and did not include a common GND. When the big motors started up, the GNDs at each place bounced in voltage and crashed the network. A common GND wire, using the cable shield, fixed the problem.
We only had a 2 pin connector on our PCBs :(
Since then, all our boards use a 3 pin RS485 connector!
 

Thread Starter

tonyg1097

Joined Jun 26, 2023
3
You need the GND connected. That is the problem. The common mode voltage of the drivers are being exceeded by the AC leakage.
Years ago, when I first started in my industrial control profession, we installed an RS485 network on a wire drawing factory. And as it was our first RS485 project, we too thought it was "A two wire network" and did not include a common GND. When the big motors started up, the GNDs at each place bounced in voltage and crashed the network. A common GND wire, using the cable shield, fixed the problem.
We only had a 2 pin connector on our PCBs :(
Since then, all our boards use a 3 pin RS485 connector!
I do believe you found it! We dropped that 1 meg to a jumper since you mentioned the ground commonality thing and things started working with one bad actor. Several more cable tests to do but you are correct. I noticed that when the batteries in the laptop are topped off, we get intermittent functionality - well not any more! Bada boom! Thank you! Case closed.
 

dendad

Joined Feb 20, 2016
4,375
We are pleased to help :)
FYI, here is a picture of one of our boards. It has provision for 2 RS485 cable runs for added reliability as it is generally used in building fire control.
UMDC16-net.jpg
 
Top