Author Topic: Strong Magnetic field (wireless charger) and bus capacitance on I2C  (Read 850 times)

0 Members and 1 Guest are viewing this topic.

Offline BBBbbbTopic starter

  • Supporter
  • ****
  • Posts: 302
  • Country: nl
So I need a reasoning check, since this is a bit beyond my understanding.

I'm detecting an intermittent issue with the I2C communication between 2 components (PCB level, <10mm apart) in the presence of a strong EM field (freaking phone wireless charger). Components are fine, but I2C rise time is slow. Internal resistor (~28k, no lower value available) is used.
Obviously the value is over the recommended one, but in normal conditions it works fine.

Now this is where I'd like some help - explaining the effect of the charger on the RC value on the bus lines.

I'm guessing since the magnetic field is strong, capacitive coupling becomes significant enough to affects the bus capacitance (increase) and thus tips the rise time to logic high over the edge causing the issue on the I2C lines.

I'm still to play around with the com. speed, playing with pull ups is not possible without HW change, but I'd like to check with you all if my thought process is good or am I talking out of my ass.
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 17429
  • Country: us
  • DavidH
Re: Strong Magnetic field (wireless charger) and bus capacitance on I2C
« Reply #1 on: July 12, 2020, 06:10:41 pm »
I'm guessing since the magnetic field is strong, capacitive coupling becomes significant enough to affects the bus capacitance (increase) and thus tips the rise time to logic high over the edge causing the issue on the I2C lines.

Magnetic fields yield inductive coupling.  Electrostatic fields yield capacitive coupling.

The magnetic field from the inductive charger effectively creates a voltage source in series with the I2C driver and reciever.  The magnitude depends on the field strength and the area between the I2C signal and ground so using twist pair or twin lead or a coaxial cable will help reduce inductive coupling.
 

Offline BBBbbbTopic starter

  • Supporter
  • ****
  • Posts: 302
  • Country: nl
Re: Strong Magnetic field (wireless charger) and bus capacitance on I2C
« Reply #2 on: July 12, 2020, 06:25:16 pm »
Magnetic fields yield inductive coupling.  Electrostatic fields yield capacitive coupling.
Yes, and that holds for "general case" analysis, but in high magnetic field capacitive coupling becomes a factor as well. My question is can these phone chargers be responsible for it? or am I looking at something else describing the increased rise time.

The magnetic field from the inductive charger effectively creates a voltage source in series with the I2C driver and reciever.
What I'm noticing on a capture by an LA is a slow rise time issue.

using twist pair or twin lead or a coaxial cable will help reduce inductive coupling.
This is really pointless for the described case (2 ICs, PCB level, less then 10mm apart)
 

Offline Marco

  • Super Contributor
  • ***
  • Posts: 7045
  • Country: nl
Re: Strong Magnetic field (wireless charger) and bus capacitance on I2C
« Reply #3 on: July 12, 2020, 08:04:00 pm »
Are you quite sure the power supply rails of the ICs aren't being affected? That's more easy to believe to be a source of problems than a 10mm trace.
 

Offline WattsThat

  • Frequent Contributor
  • **
  • Posts: 788
  • Country: us
Re: Strong Magnetic field (wireless charger) and bus capacitance on I2C
« Reply #4 on: July 12, 2020, 08:18:07 pm »
I would expect your I2C pull-ups should be in the neighborhood of 4K7, maybe even lower due to the external field. I have not read the spec of late but 10K is the absolute highest value I can ever remember using for short runs. I’ve gone as low as 2K2 when going off-board and pushing the limits of cable length.

What device are you using that has internal pull-ups?
 

Offline BBBbbbTopic starter

  • Supporter
  • ****
  • Posts: 302
  • Country: nl
Re: Strong Magnetic field (wireless charger) and bus capacitance on I2C
« Reply #5 on: July 12, 2020, 08:51:43 pm »
I would expect your I2C pull-ups should be in the neighborhood of 4K7, maybe even lower due to the external field. I have not read the spec of late but 10K is the absolute highest value I can ever remember using for short runs. I’ve gone as low as 2K2 when going off-board and pushing the limits of cable length.

What device are you using that has internal pull-ups?
uC with 28k Strong pullup (I know it's quite big and a solution would be most likely be to go lower with an external ~4k7 but that is a change in a design, BOM increase...)
Again the raise time issue exhibited only when the charger field is present. That's why I'm trying to explain this by the added parasitic capacitance simply puts it over the edge.

Are you quite sure the power supply rails of the ICs aren't being affected? That's more easy to believe to be a source of problems than a 10mm trace.
supplied by a CR2032, without any disturbance noticed, and both ICs are working fine, only the rise time issue on I2C (captured by an LA - Saleae, so some room to doubt the conclusion there...).
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 17429
  • Country: us
  • DavidH
Re: Strong Magnetic field (wireless charger) and bus capacitance on I2C
« Reply #6 on: July 12, 2020, 09:26:19 pm »
The magnetic field from the inductive charger effectively creates a voltage source in series with the I2C driver and reciever.

What I'm noticing on a capture by an LA is a slow rise time issue.

Are you sure you are not seeing the rise from the coupled flux?  Inductive chargers are resonant so the output is a sine wave and not a square wave.

Conceivably the current induced by the flux is saturating the driver or receiver or causing substrate currents producing all kinds of bizarre behavior.

Quote
using twist pair or twin lead or a coaxial cable will help reduce inductive coupling.

This is really pointless for the described case (2 ICs, PCB level, less then 10mm apart)

And how much area is between commons?  The same rules about minimizing area apply to any current loop and when doing critical layouts, I often pairs signals with their commons.
 

Offline Kashif

  • Contributor
  • Posts: 17
  • Country: us
    • Explore Silicon
Re: Strong Magnetic field (wireless charger) and bus capacitance on I2C
« Reply #7 on: July 12, 2020, 10:44:41 pm »
Have you probed the i2c lines to check SI before and after the charger on?  What is your i2c speed? I am suspecting due to higher value of pull-up, you might have marginal VIH.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf