Noise on the power rail

[John80]

Hello to everyone!

I was debugging a small hobby board that includes an FTDI FT230XS usb to serial IC. The circuit looks like that:



The FTDI_TXD and the FTDI_RXD are going to a PIC microcontroller (directly connected at the pins that support hardware UART). The +2.8V are generated by an LDO and this LDO is supplying the FTDIs +2.8V and the PIC with +2.8V. All the decoupling capacitors are in place! And somewhere here I am start having problems! ...

When the circuit is at the power off state (no power at the input of the LDO) just the USB cable attached (so the FTDI is powered on because I am having 5V from the USB) and without to open a terminal program at my PC... at the +2.8V power rail I am having something like that:



With exactly the same configuration (no power at the LDO, USB cable attatched) but now I've just opened a terminal program and press connect with the FTDI IC I am having something like that at the +2.8V:



The peak to peak noise at the first one is around 180 - 270mV (the picture is showing the highest one that I was able to catch). The peak to peak noise at the second capture is around 600mV to 910mV! Both of those are on top of a 600mV DC... so, they are not around zero.

I know that the circuit includes an error at the component LH1 (a ferrite bead) that should be before the capacitors C11, C25 but I don't think that this is the cause of all this huge noise! Especially when the terminal is connected with the FTDI IC (have in mind that it is just connected and I am not exchanging any data at all at least from my side)!

The biggest problem is that the noise is transferred to to PIC power rail, creating sometime strange UART transmits/receives and in general affecting the normal operation when I am power the LDO and running my program

What is the cause of that? Any info? Any update of how I can solve this?

PS: During measurements I've used the probe's ground ring and not the gnd cable with the alligator ring in order to have the lowest environment effects as possible. I've also tested a couple of USB cable with pretty much the same results

[tszaboo]

Can you give more information about the 2.8V PSU, type bypassing type etc...
Also, your reset pin is not connected (and it is in the datasheet), maybe it resets itselfs every now and then...

[Dave]

I know that the circuit includes an error at the component LH1 (a ferrite bead) that should be before the capacitors C11, C25 but I don't think that this is the cause of all this huge noise! Especially when the terminal is connected with the FTDI IC (have in mind that it is just connected and I am not exchanging any data at all at least from my side)!

Bodge a decoupling capacitor directly between VCC and GND.
You want capacitance directly across the supply pins, so the loop that handles current spikes is as small as possible, therefore it has the lowest inductance possible. Your loop has an inductor in it.

I wouldn't be surprised if the bodge cap solved all of the noise problems.

[John80]

Ohh... so, in other words this ferrite bead is the cause of all problems! Is that true? I will try to remove it and check the results.

As for the reset, I think that i've read that you can left it unconnected if unused. Probably already pulled up by and internal resistor

[Monkeh]

As for the reset, I think that i've read that you can left it unconnected if unused. Probably already pulled up by and internal resistor

You think you've read? Check!

[Dave]

Page 15 of the datasheet:

RESET# can be tied to VCC or left unconnected if not being used.

My first guess would be that the chip resets itself, because the supply voltage dips too low.