Isolated I2C and 3.3v DC Power

Thread Starter

robsworld

Joined Nov 5, 2015
86
Hi, I'm trying to isolate I2C data lines and DC power for a sensor, does this look correct?

pH_Isolation.jpg

I'm using the RFM-0505S to isolate DC, it's $3 cheaper for a 5v vs 3.3v so I added a regulator after it to get my 3.3v.
https://www.mouser.ca/datasheet/2/468/RFM-1710978.pdf

Then I have the ISO1540DR isolating the SCL/SDA lines. Is it correct to have both GND1/GND2 and VCC1/VCC2 come from the isolated side? I can't see a problem with it but datasheet with it's example it shows otherwise. I don't have non isolated 3.3v and it doesn't feel right to put non isolated GND there.
https://datasheet.lcsc.com/szlcsc/1811082013_Texas-Instruments-ISO1540DR_C179739.pdf

One other thing, in the AZ1117 regulator datasheet it says if a .1uf capacitor runs in parallel of the output it needs to be 5mm away. My board is small so it will be closer, can I run about 6mm of trace in a loop?

The I2C lines do have resistors on each side just not in the image.

Thanks
 

Attachments

Marc Sugrue

Joined Jan 19, 2018
222
Both sides of your ISO1540DR seem to be connected to the same 3.3V rail in essense making the isolation of the IC redundant, currently it is just a non isolated signal reapeater. To operate as isolated Vcc1 and Gnd1 need to be galvanically isolated from Vcc2 and Gnd2 to operate as isolated. You need a isolated power supply to power the IC.
 

Thread Starter

robsworld

Joined Nov 5, 2015
86
Thanks, how about this?

pH_Isolation2.jpg

The ISO1540 input is now all non isolated at 5v as the SCL/SDA in is 5v.

The output is all isolated and VCC2 is 3.3v so SCL/SDA should be 3.3v as it does logic level conversion, correct?

Now it looks like the images in the datasheet.
 

schmitt trigger

Joined Jul 12, 2010
868
Wit respect to the capacitor issue on your regulator, I downloaded Diodes, Inc datasheet, and could not find any mention of the 0.1 uf capacitor to be far away. The only capacitor related information was:

The AZ1117C is compatible with low ESR ceramic capacitor.
The ESR of the output capacitors must be less than 20Ω.
A minimum of 10μF output capacitor is required


Could you attach the datasheet where you found the information?
 

Deleted member 115935

Joined Dec 31, 1969
0
Im going to nit pick..
Circuits normaly read left to right,
in this case, I'd say thats from Power in to signals out
so circuit is back to front,


Re the AZ1117, Thats End of life as a part, though tit is a "generic" 117 LDO

Just to highlight, the LDO will NOT isolate the power in , you are relying upon your PSU on the left of your circuit.
 

dendad

Joined Feb 20, 2016
4,451
Just a thought, on your PCB design, I'd add a slot between the pins 2 and 3 of the RFM0505s isolated power supply, just to help keep the isolation voltage rating.
 

Thread Starter

robsworld

Joined Nov 5, 2015
86
Im going to nit pick..
Circuits normaly read left to right,
in this case, I'd say thats from Power in to signals out
so circuit is back to front,


Re the AZ1117, Thats End of life as a part, though tit is a "generic" 117 LDO

Just to highlight, the LDO will NOT isolate the power in , you are relying upon your PSU on the left of your circuit.
Your right, the schematic is backwards, this one should be better. ;)

pH_isolation4.jpg

I realize the LDO won't isolate, I'm only using it because the isolated DC-DC is 5v, a 3.3v is $3 more which makes no sense so this is cheaper overall, that's why it's on the isolated side.

The datasheet you linked was for the AZ1117E,

The datasheet that I found was for the AZ1117I,

So it all depends which version you are purchasing.
Interesting, I wonder why a difference. Turns out they're about 8mm apart so it passes.

Just a thought, on your PCB design, I'd add a slot between the pins 2 and 3 of the RFM0505s isolated power supply, just to help keep the isolation voltage rating.
Thanks, will do. I have separate ground planes that split between those pins and a nice space where the slot can go.
 
Top