Go Down

Topic: extreme noisy board voltage 3.3V on Nano 33 IOT (Read 520 times) previous topic - next topic

KaiMatzen

Hi,

i realize a project with the nano 33IOT. The sketch runs only properly, when the USB - Cable is connected to a PC. If I use the Vin at 7V hogs the processor after a few seconds.

I looked for the failure and I did some measurements at Pin <Vin> and <3.3V out>. For these test a have used a new nano without any personal programming - right out of the box. I tested 4 devices.

Every device seems to have the same issue. The board voltage oscillating with 1Vpp around the 3.3 V main level. See the attached scopes.

So I used a stable voltage source at 3.3V and connected the board thru the <3.3V out>. All nanos works properly. No crashes.

I think, that the sept down converter MPM 3610 is the problem. These oscillations can't erased by an 100uF capacitor.

Have anyone any suggestion, what I can do? Have someone the same issue?

I tested the Nano 33 BLE. This board is ok. I tested also the Nano Every. Here I can see horrible oscillations until 1.6Vpp.

What the hell...

Greetings

Kai


Klaus_K

I did a few measurements with Nano 33 IoT.

Yes, I can see noise. The pattern of the noise depends on the radio usage. There are peaks in groups about 500ns apart with a clear repeating pattern. The pattern looks different between BLE and WiFi. I suspect the noise comes from the increased current needed by the radio when transmitting a packet.

Looking at the datasheet the transmit current is at least 2-3 times the average current (measured by my lab supply). When I limit the current between the average and the peak current, I can see the power supply voltage crashing. So, the radio module is switching on and off very rapidly as expected.

The behavior is not affecting my Nano at all, as long as the current limit on my supply is set high enough. The sketches run without any issues.

I can see the same pattern with slightly lower amplitude when running from USB. And it clearly starts with a delay in both cases, it takes a few seconds for the radio to initialize.

For analog measurements it is probably a good idea to switch the radio off or synchronize the measurements with the quiet times.

KaiMatzen

Hi Klaus,

It would be crazy, that the processor crashes while the onboard radio is used, perhaps the device is uselessly.

I don't use the radio in my sketches. I'm not sure about the state of the radio while first running after unpacking the device.

I can see the noise / oscillating on a new device on 3.3V pin.

The datasheet for the step down converter MPM3610 shows a reference pcb design, that I can't see at the Nano family. The manufacturer talks about critical design rules ... or the device is still oscillating ...

If I use an external voltage source on 3.3v pin while the MPM is inactive, then I can't determine any oscillations or transitions at the power lines.

May be, that you talk about an other issue ...

Thanks anyway

Klaus_K

Hi KaiMatzen,

I did some more measurements. You are right when I shut down the radio and put the microcontroller to sleep, I can still see the noise. It is weaker but definitively there. The 500ns I measured is likely coming from the 2MHz switching frequency stated in the datasheet of the MPM3610.

I also looked at the PCB and layout design. You are right they did not follow the layout guidelines at all.

The output capacitor is connected through a thin trace 1mm away and the input capacitor is even further away. The input capacitor placement is mentioned at least twice in the datasheet as being critical.
The feedback network seems to be exactly where it is not supposed to be, when I read the datasheet PCB Layout guidelines point 4 correctly.

Go Up