# Is shorting two power rails with a capacitor like this OK?

I am using this CAN transceiver which runs on 5 V but, if you provide it with the microcontroller voltage on its VIO pin, then it will automatically apply voltage level correction for the micro.

The following has been mentioned in the design notes:

As you can see, they suggest creating a high-frequency short circuit between the 3 V and the 5 V using a 100 nF capacitor.

Is this recommended practice? Must I be wary of anything because of it?

• How is it any different from the high-frequency short between, say, 5 V and ground? – Dave Tweed 2 days ago
• Well, the note at the bottom explains why they recommend it. There isn't much more to expand on this, I think. It certainly is specific to this component, though. I wouldn't do it for another random component with dual supplies, unless the datasheet of this other component also suggests this. – dim 2 days ago
• I've also got … questions about the reasoning of *** there, @dim. It's great that you've got less high frequency noise on V_IO relative to V_CC that way, but who is helped by that if the point of V_IO is to share an IO reference voltage with some microcontroller? That just feels like it misses the point. – Marcus Müller 2 days ago
• @MarcusMüller I agree, there seems to be little point in wanting to have V_IO and V_CC sharing the same noise, rather than having both supplies decoupled from ground only, as you would expect. Which is why I said I wouldn't do it for another component. Now, as for "who is helped by this": maybe the level translation circuitry within the chip itself prefers both supplies being coupled together, or there is more risks of glitches... That's the only reason I could imagine, but this is beyond my level of expertise at this point, anyway. – dim 2 days ago