Issue with Raspberry Power Supply

Thread Starter

Xavier Pacheco Paulino

Joined Oct 21, 2015
728
The Pi has always been kind of selective when it comes to the power supply. They usually recommend purchasing a 2.5A PSU. I'm using two microcontrollers in my project, an L432KC nucleo which claims to be an ultra-low power MCU, and a Raspberry. I chose this AC-DC converter to power up my project. But it was a bad idea, since the board which has the converter is located 7 feet away from the Pi. So, I used a "good quality" 7 feet cable to power the Pi. It turns out that the Pi throws the "Undervoltage detected" warning. When I measure the voltage going to the Pi, I get like 4.95/4.98V. So, I'm losing about 0.05V in the wiring. I've changed the cable twice, but it remains the same. If I use the original Raspberry PSU, everything is fine. The idea is to power up the whole project with a unique power supply. So I chose the converter mentioned before to be 4A to leave a good margin for the Pi and for other minor loads which also need 5V. A better idea that I should've considered is to have 12V in the main board, and convert to 5V next to the Pi. That way, I'm transporting 12V instead of 5V, which lead to minor losses. So, what do you think the real problem is? The fact that I'm transporting 5V, or I'm still getting bad quality cable? Or worse, the converter itself? The original PSU that came with my Pi is also about 7 feet length.
 
Last edited:

Sensacell

Joined Jun 19, 2012
2,568
The AC-DC converter is a little tiny thing...

Not much room in there for "bulk capacitors"
You might need to add a large electrolytic capacitor near the Pi.

You don't mention the wire gauge, ? what is it?

Raspberry Pi's are very finicky about power input, they also draw large spikes of current, so measurements can be misleading.
 

danadak

Joined Mar 10, 2018
4,057
Is it startup transient you are not observing with a standard meter. A
scope would be better at this.

Better bypass cap at Pi due to distance.

Not all caps are equal -



Regards, Dana.
 

Thread Starter

Xavier Pacheco Paulino

Joined Oct 21, 2015
728
If I power up the whole project with RPi power supply (2.5A), everything is fine. I mean, this what I'm doing: I'm feeding the Raspberry through the micro usb connector. Then, from the 5V pin of the Pi, I feed the rest of the circuit, which is less picky than the Pi. Seems like the official RPi supply is way more efficient than my converter.
 

LesJones

Joined Jan 8, 2017
2,664
Have you measured the voltage AT THE RASPBERRY PI END OF THE USB cable. I have found that many micro USB cable have quite a high resistance. When I had similar problems I ordered the shortest cables I could find but when they arrived they did no thave the lowest resistance of the selection of cables I had.

Les.
 

Thread Starter

Xavier Pacheco Paulino

Joined Oct 21, 2015
728
Thanks for your suggestions. Indeed, I tried adding the capacitor with no success. But I think my problem is the cable I'm using. It's a CAT5E Ethernet cable. The reason why I chose that cable is because I need 6 lines to communicate the main board with the Pi. I need 4 wires for RX and TX (differential), and 2 wires for VCC and GND. As the ethernet cable brings 8 wires, 2 are leftovers, which I connected to VCC and GND respectively. The serial communication works perfectly. But the power is not being effectively delivered. I wanted a well structured cable, instead of having the six lines in the air. The PCBs in communication, both have ethernet female connectors, so I would have to change the design in order to use another type of cable. Also, note that although I'm using ethernet cable, the purpose of the cable is rather different. I've seen many devices using this type of cable for other purposes. If you have a better suggestion, let me know. Anyways, in the new design, I'm considering using 12V, and convert to 5V next to the Pi.
 

Sensacell

Joined Jun 19, 2012
2,568
CAT-5 is usually #24 AWG.

1 foot of this stuff has about .025 Ohms resistance, your round-trip is 14 feet = 0.35 Ohms.

That's probably enough to cause your problem.
 

LesJones

Joined Jan 8, 2017
2,664
I suggest starting with something like 9 volts at the supply end of the cable and regulate it down to 5 volts at the raspberry Pi end of the cable using a low dropout regulator.

Les.
 
I've found with a current project I've been fighting with that for some reason the standard ethernet connectors don't always work perfectly for connecting power. There's times part of the project stops working right until I disconnect and reconnect it or wiggle the connection. I have connections to six identical devices between 15 - 50 feet from the controller and when I switch the connections around the problem moves with the change in connections. I've changed the connectors themselves several times and it always helps for a little while, but eventually it goes back to the way it was before. I've also changed the other end to make sure there wasn't a faulty unit at the other end.
 
Last edited:
Top