Noise on 3.3v rail - is this normal?

[blapcb]

I have a 3.3v rail and several boards that are working basically fine with the same power section. The LDO is MIC5209 (schema/use as per the data sheet). I have been experiencing interference on some other communication lines (SPI), and put a scope to the 3.3v rail. To my horror, I discovered a lot (is it a lot?) of spikes/noise which get severely aggravated when a RS232 cable/device is connected (I used MAX3232 line drivers).

See attached images.

My question is, how much AC should there be on such a rail? Is it normal? If not, how do I reduce/eliminate it?

 

[mtwieg]

That's far more "noise" than you should see on your supply rail. Looks like you need some additional bypass caps on your supply. Make sure to distribute the capacitance so that each IC gets at least one.

Are those spikes happening at the baud rate of the RS232? What's different about the third image?

 

[dick_freebird]

The current into the too-capacitive RS-232 load needs
to be softened, and the input supply needs to be stiffer
(more, higher quality decoupling for spikes; better LDO
dynamics for the overshoot).

 

[blapcb]

The current into the too-capacitive RS-232 load needs
to be softened, and the input supply needs to be stiffer
(more, higher quality decoupling for spikes; better LDO
dynamics for the overshoot).

Thanks... but... can you translate that in to something actionable that someone like me can understand and do? (softened? stiffer?)

Just to mention that I have been using these two parts, the LDO MIC5209 and line driver MAX3232 on many boards and this is the first time I have come across this problem, both in terms of the amount of noise on the 3.3v rail as well as when the RS232 is connected (I just checked on 3 other different boards using the same components and this does not happen). So I must assume it is something in the layout/design of the board in question.

All of my ICs have the standard 0.1uF decoupling capacitor.

 

[mtwieg]

0.1uF may not be enough. Try adding a few more uF and see if that helps.

Again, what events are actually correlated with those spikes? They're obviously caused by some sort of switching or transition.

 

[BradtheRad]

There are things to watch out for when using several boards (in addition to suggestions which were already mentioned).

It brings in greater opportunity for faulty ground connections, ground loops, echo spikes, exposure to 60 cycle hum, etc.