AMC1303M0520: Emissions Reduction Approach

Patrick Frye

Part Number: AMC1303M0520
Other Parts Discussed in Thread: ISO224, AMC1200, AMC1100

Hello, I am trying to pass EMC for my inverter which uses 3x AMC1303M0520DWV ADCs. During radiated emissions I get noise from the clock and data signals. I have tried shielding, rounding the clock edges, choking the power lines, etc. to no avail. I was considering adding LVDS drivers and receivers to the signal path, but I tested the ADC with the clock and data lines disconnected, and there is still too much noise coming from the IC pins. Do you have any advice on an emissions reduction approach for these devices? Am I really just fighting the clock/data signals, or does the isolation encoding also make noise?

Thanks,

Patrick Frye

4 months ago

0 Eva Mason 4 months ago

TI__Expert 4215 points

Hi Patrick,

Clock and data lines will show some noise. You could try embedding your clock and data lines between two ground layers to see if this helps.

I would also recommend taking a look at this white paper for best practices.

Best in Class Radiated Emissions EMI Performance with Isolated Amplifiers (Rev. A)

Please let me know if you can try this and if it helps, thank you.

Best regards,

Eva

0 Patrick Frye 4 months ago in reply to Eva Mason

Prodigy 10 points

Thanks for the white paper! The noise I am fighting is between 100MHz and 400MHz and it looks EXACTLY like the signature shown on figure 3-2. So, I now think I am chasing emissions from the isolation barrier. The whitepaper states that all isolated amplifier ICs have the improved noise design changes, except for the AMC1100, AMC1200, and ISO224 devices. How about my chip, the AMC1303? judging form my emissions testing I think it must have the same design as the AMC1200? Is there a compatible sigma delta ADC that reads +-50mV that has all of the latest emissions improvements?

Thanks,

Patrick

0 Eva Mason 4 months ago in reply to Patrick Frye

TI__Expert 4215 points

Hi Patrick,

I am glad the white paper was helpful. I think the AMC1303 should also have the improved noise design changes from the paper.

As for finding comparable devices, you could look at our isolated ADCs comparison page:

Isolated ADCs product selection | TI.com

Please let me know if you have more questions!

Best regards,

Eva

0 Patrick Frye 4 months ago in reply to Eva Mason

Prodigy 10 points

Hello, Based on your last response, ordered AMC1306M05QDWVRQ1. It looks like a drop-in replacement for the AMC1303M0520DWV, except that it does not generate a clock signal (which we can create).

One more question, what I see on my spectrum analyzer, and on figure 3-2, is a radiated emissions peak every 10MHz from 100MHz to 400MHz. Was this a result of the carrier signal crossing the isolation barrier? I ask because I have not seen an EMI signature like this and I am trying to understand what causes it.

Thanks,

Patrick Frye

0 Eva Mason 4 months ago in reply to Patrick Frye

TI__Expert 4215 points

Hi Patrick,

I think the 10 MHz spacing corresponds to a quarter-wavelength resonance effect. If you are getting similar results to the white paper, it sounds like you are getting good performance.

Best regards,

Eva

Data converters

Data converters forum

AMC1303M0520: Emissions Reduction Approach