ISOM8118: ISOM8118 and interference @ certification

Just note - problems starts around are around 13.4-13.8 MHz.

R.

Hello Lada,

I am unfamiliar with this standard. Is IN_A and IN_B a differential signal?

Lada Lostak said:

When higher amplitude and higher frequencies are aplies we are loosing 'value

Can you provide me with more details about what this means? Maybe include a scope capture of the problem? 

ISOM811x has a limited bandwidth ~680kHz. I suspect the decreased amplitude at the output is related to the bandwidth limitation. For my reference, Comparable optocouplers are much lower bandwidth. What are the other circuits you have used? 

Hello, no differential signal. It is standard SMEMA - 24V switched by a relay on the other side. I am also not so much into norm for testing, but this test is simple :)

Right now, we just failed this test so we don't have scope yet. Loosing value means: the second side sends to us 0/24V (switched by relay) with 1second duty cycle. On the DUT we are sending the value throw Ethercat back to 'master' and these values are compared. If comparison is OK -> test passed. In our case - EtherCAT works all the time (another part of test which was passed) but value readed by DUT doesn't match with value sends by master device -> comparison failured -> test NOT passed. We passed most of frequencies.

When we tried some generic OPTO -> it worked. ISOM811x are new devices for us (by PDF it is opto-emulator). This device should be manufactured in 10K+ quantities, so it is important for us to understand what happens, since we would like to keep this OPTO in the device.

Bandwidth: static signal Hiz / 24V with ~1 second duty cycle. Don't know from head, but 20mS is allowed as 'turn around' for input-to-output comparison. Far enough...

Second side: GND directly connected, 24V from same source switched by SSR. Nothing else.

Schematics: master device - output ------[cable]----> DUT input ---> etherCat ----> master device

Power supply is the same. Conditions: wires between master and DUT are exposed to EMI source (exacly defined in norm, as far I remember 12Vamp signal). At given frequencies we lose output to input reading comparison. Most frequencies are OK, but the one mentoined above are problematic.

We need to analyze the problem deeper, so at this point the most likely cause is disturbing OOK signal in OPTO coupler. In this case we would need to add some additional low-pas filter at the input. And since standard opto-coupler is slow/clean analog, comparison is OK.

Thank you !

R.

Hello Lada,

The OOK frequency is much higher than the 13.4MHz signal that you are having a problem with (in the GHz range). To be clear, is the 13.4MHz problem frequency an input seen at IN_A and IN_B? What is the amplitude? 

Or is the ISOM8118 being used as a FB for a power supply here? 
Best,
Andrew

Hello, thank you! This issue is new, so first we have to repeat same conditions here to measure something (there is no much time in the official lab, since the count tests).

IN_B is connected to common group.

IN_A is connected throw wire to SSR and second side of SSR to common +24V.

IN_A wire is the victim.

3.3V power supply features LMR36006BRNXR with filters and seems to work OK (everything other is passed, including this test with different OPTO).

When OOK is not an issue (far different frequencies), then the problem seems to be in time window, where we're victim. EMI source is strong, so when EMI is aplied, there must be flowing current throw OPTO. At this window sensor state is 'unknown' at all. The different behaviour seems to be because of different construction of OPTO.

Does this sounds to you realistic to you ?

I asked SW guys and actually there is no lo-pass filter in SW (previous products used 'real optos'), which seems to be necessary.

R.

Thank you for confirmation !

R.