ADM2582E doesn't receive differential signal with low common-mode voltage

Hello, 

This is unusual behavior for the ADM2582E, the receiver should function correctly across the full -7V to +12V common mode range. 

I would recommend confirming that the VISO power supply and VCC power supplies are stable with an oscilloscope. Excessive AC noise on the power supply inputs may effect operation. 

Are these plots taken at room temperature, or does this behavior occur at a specific ambient temperature? 

The common mode noise on the bus side will appear across the isolation barrier. If the slew rate of this noise exceeds the common mode transient immunity specification of 25kV/us, data across the isolation barrier may be corrupted. 

From your first plot, the common mode signal on the RS-485 bus does not look this fast, but it would be good to rule this out as a cause. Would you have a math waveform of the RS-485 bus common mode signal (A+B/2)? 

Please also confirm that the DE & #RE signal of the ADM2582E remains a stable logic low during the remote device's response window.

Regards,

Neil 

Thank you for answer. I checked VISO and VCC power supplies - they are normal. DE and #RE work properly.

After I connected sine from generator to both lines RS485. Here is oscillograms.

 - A and B with respect ground:

 - The same A and B and else A-B (purple):

 - The same A and B and else A+B (purple):

 - One of lines (A or B) and RX line:

After I discover that if I unmount common-mode choke (it is mounted in differential line) and restore A and B lines that RX line work correctly. Is this normal behaviour of ADM? Can common-mode chocke affect to ADM?

Thank you for answer. I checked VISO and VCC power supplies - they are normal. DE and #RE work properly.

After I connected sine from generator to both lines RS485. Here is oscillograms.

 - A and B with respect ground:

 - The same A and B and else A-B (purple):

 - The same A and B and else A+B (purple):

 - One of lines (A or B) and RX line:

After I discover that if I unmount common-mode choke (it is mounted in differential line) and restore A and B lines that RX line work correctly. Is this normal behaviour of ADM? Can common-mode chocke affect to ADM?

Thank you for sharing this additional detail. 

Depending on the characteristics of the common mode choke, this could definitely affect the receiver performance. All CMC's are not symmetrical; some of the AC common mode signal can be converted into differential mode which can impact the differential voltage seen by the receiver. 

With the waveforms captured above, are these taken with the common mode choke in place? On which side of the CMC are the A and B voltages captured (on the connector side of the choke, or on the ADM2582E side of the choke)? 

Could you share the model # of the common mode choke used? 

Regards,

Neil 

Yes, the CMC was placed when I captured the oscillograms. They was captured in the connector side. The CMC is DLW21SN501SK2.

Now I took waveforms on the ADM side and signal was correct. But I saw on another channel where problem was not appeared. This is because the CMC had been took off on malfunctioning channels.

But another one problem was discovered. When transmitter is disabling, on RX line appears parasitic peak:

yellow line - rx, blue line - DE and #RE

And RX and A lines:

If to decrease common-mode voltage then parasitic peak passes off.

if #RE is always up then the problem does not cease.

This common mode choke is optimized for higher speed interfaces like LVDS, USB, etc but should also work correctly with the ADM2582E. ACT45B-110-2P-TL003 is another suitable option. 

The low RxD pulse on the HL edge of the DE signal shown in your recent plots is known behavior on the ADM2582E receiver.  A hardware fix for this is to place a 1nF capacitor between the A and B pins of the ADM2582E input. In end application the inherent cable capacitance may be sufficient. There are some more details in related post here: 

https://ez.analog.com/interface-isolation/f/q-a/542934/adm2682e-unnecessary-l-output

A software fix would be to flush the receiver buffer of the UART or similar after the DE signal is transitioned low. 

Regards,

Neil 

Thank you for this information. Placement of capacitor solved the problem. And we decided not to install common mode choke.