2

I am feeding 5.41 V on the microUSB power supply pins to a raspberry Pi 3B. Yeah, I know, USB = 5V +/- 5%. Because reasons, I have to keep doing this for a while. What I find is that dmesg constantly spews out undervoltage messages (and then voltage normalized). Is this known behavior? Ie, does the Pi interpret over-voltage as under-voltage?

The voltage is being measured by a decent DMM. Not nat'l bureau of standards good, but it's decent. And it's corroborated by other instruments of lesser quality. So it really is 160mV (or thereabouts) high.

8
  • Hi @i wish it would work, Welcome and nice to meet you. Ah, let me see. First brainstorming ideas: (1) Constraints: I am only talking about Rpi4B, 1/2/4GB, official and unofficial supplies. (2) Measuring tools: I usually use cheapy digital multi-meters, yes, two to auto/cross calibrate. I do have two digital storage scopes, 30/50MHz, again to cross check, because I ... (101) I do have 101 points to make, but let me jump to (102) Almost all electronics things have a "voltage regulator" at the Vin, eg. 5V microUSB connector. – tlfong01 Apr 16 '20 at 1:33
  • This regulator is very forgiving, say you can input 6V and it will step it down to around 5V. If your Rpi is taking 1A say, then the wasted power (6V - 5V) = 1V * 1A = 1W, is dissipate as heat energy, warming up the Rpi's PCB. In other words, it is OK to stress supply input 5.5V, safely, at least for one hour. Yes, I once supplied 6.5V to a Rpi3B+, 16 hours a day, or 14 days in a row, and nothing happens, ... – tlfong01 Apr 16 '20 at 1:44
  • 1
    Could you expand on the reasons you're feeding it with 5.41V ? I suspect overclocking... It may be possible to explore something like POE to work around the issue. – Criggie Apr 16 '20 at 9:03
  • 1
    @tlfong01 what are you even mumbling about – Swedgin Apr 16 '20 at 13:37
  • 1
    @tlfong01 that's better. And idd 5.4V shouldn't be a problem, it depends on the component that has the lowest max voltage rating. – Swedgin Apr 16 '20 at 15:02
4

"Is this known behavior?"

This is common for poor supplies and/or inadequate cables.

The warning is for transient undervoltage - you can't measure with any kind of DMM.

Trying to compensate by increasing voltage is likely to damage the Pi.

Use an official supply

2

Two things can get your measurement all wrong:

  1. Where do you measure the voltage? If you measure it at the power supply side, it's quite possible that you have a voltage drop in the cable and the voltage is much lower on the RPi side.

  2. How fast do you measure the voltage? I have a cheap multimeter with a refresh rate of 0.5 seconds. Anything faster than that simply cannot be seen. Another expensive multimeter I own has a data logger mode with a sample rate of 0.001 second (it comes with a USB stick), which is enough to monitor the supply voltage. A cheap oscilloscope in trigger mode can catch voltage drops in nanosecond range. If your DMM is close to the first option, it's perfectly feasible for the Pi to complain about undervoltage which you simply cannot see.

As for the voltage you use, the power management IC (MXL7704) has the input voltage range of 4.0V to 5.5V, so you can't kill that one with 5.41 V. I don't own a Pi 3 so I don't know which other ICs are there which are powered with 5V. Find them out, google the datasheet and check how much voltage they can withstand. Do the same for all the USB devices you want to plug in. Then you'll know the limit.

Note that running at 5.41V (with only 0.09V of safety margin) will kill the Pi sooner or later. There will be a voltage surge at your local power plant, or the motor in an HDD will fail and dump some extra voltage into USB, and you will instantly find yourself over the limit.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.