I am trying to power a raspberry pi 3 over a long distance (5 meters). I can power the raspberry pi with the standard power supply from Raspberry Pi itself, but when I use these: https://www.digikey.ca/en/products/deta ... -5/2529127 https://www.digikey.ca/en/products/deta ... 6B/9837906

The raspberry pi power led flashes somewhat randomly and does not boot up. When I use the exact same power supply noted above from digi-key, and a really short usb cable, it works fine. So it must be the long cable causing the issue? If so, would a lower gage usb wire solve this? Or am I missing something else here

  • To power a Pi over 5m you would need to use impracticality large cables. Think welding leads. – Milliways Jul 29 '20 at 2:21
  • @Milliways: That is not necessarily true. – Seamus Jul 29 '20 at 18:47
  • 1
    Please fix the broken links to digikey's website. It spoils an otherwise good question. – Seamus Jul 29 '20 at 18:51

Assuming you have got standard copper cables, over 5m you will drop over 0.6V across that distance assuming you are pulling 3Amps and 0.4V at 2Amps. Both of these drops will give a Pi real issues.

It's worth using a voltage drop calculator such as this one (others are available) as soon as you start increasing cable length beyond that supplied.

As a bit of fun, I ran a few calculations and at 3Amp draw you could get away with cables (not cable plus insulation) of 2.5mm thick over 5m...

  • 1
    Of course you are then faced with the problem of terminating the cables on the Pi. The header pins are spaced 2.54mm apart - don't even think of a µUSB or USB-C plug! – Milliways Jul 29 '20 at 7:15
  • @Milliways :-) did not think of that! – user115418 Jul 29 '20 at 10:16
  • @Milliways You can solder your 2.5mm wire to a short piece of standard 0.5mm wire and crimp that (or even solder to a micro USB plug), no problem. Electricity is not like water (where the smallest pipe diameter limits the flow), there's no problem with a thin wire as long as it's short. – Dmitry Grigoryev Jul 29 '20 at 10:20
  • @DmitryGrigoryev Duh! I am perfectly aware of this - this was partly sarcasm, but a polite way of pointing out that the question was dumb, the answer, while technically correct ignored the physical realities. For instance it did not consider the cost of 10m of 2.5mm cable or the physical imprecalities of interfacing at both ends - we don't know because the links in the Question are broken. – Milliways Jul 29 '20 at 12:11

In my experience you need something like AWG25 to reliably power a Pi over 1 meter, which has an equivalent of 0.2 Ohm series resistance and produces a voltage drop of 0.2V per every Ampere your Pi consumes.

Extrapolating these figures to 5 meters, you will need AWG18 wires. I have never seen USB cables with such a rating, so you'll have to make a custom cable.


You could use a DC/DC inverter (or one of the existing modules which provide these). Then you can power the PI over a great distance with a wide voltage range usually. That will require something like a lab power supply, though, because the input voltage would need to be significantly larger than 5V then.

  • 12V Wall warts are quite common, no need for a lab PSU. – Dmitry Grigoryev Jul 29 '20 at 13:19

Yours is a good application for a remote sensing power supply. In a nutshell remote sense power supplies regulate the voltage at the load instead of at the source - this is done to compensate for voltage drops in the cabling. Here is an example:

enter image description here

Remote sense power supplies may be purchased from numerous sources, but they are considerably more expensive than the "wall warts" used to power most RPis. They also tend to be higher-powered units that may not be suitable for RPi's relatively modest power consumption. For example, this unit made by Mean Well. Many "bench supplies" used in electronics labs have separate remote sense jacks or inputs to bring the voltage at the load back into the bench supply.

Note that the current carried over the remote sense wires is a wee fraction of that drawn by the load. Consequently, sense wires may be small gauge wires (larger AWG numbers) - they have virtually no voltage drop because the current they are required to carry is small.

There are, of course, DIY options to buying. There is nothing magic about remote sense - again, it simply takes its sensed voltage from where the load is located instead of where the power supply output is located. This blog on the Instructables website gives a fairly detailed account of hacking an older bench supply to use in remote sense mode. Similarly, this Electronics SE Q&A covers some considerations for doing the same thing to a different vintage power supply.

You may find a remote sense supply in that old desktop PC: Some ATX power supplies have been built with a +5 V sense wire (typically colored pink) connected to one of the red +5 V wires. However, as this was never part of any official ATX standard REFERENCE, a detailed inspection will be necessary to "qualify" its use as a general-purpose remote-sense 5V supply.

If you wish to power the RPi using its USB-C connector, you'll need a detailed wiring diagram for the USB-C cable. The USB-C cable itself may be a good choice for connecting a remote-sense power supply to the RPi as the shielded data lines should be good choices for carrying the sense voltage back to the PSU.

  • Nice idea, though I'd still go for a 12V PSU and a 12V -> 5V converter on the other end: less wires and less failure modes. I bet a broken remote sense wire would produce a spectacular failure :) – Dmitry Grigoryev Jul 29 '20 at 13:21
  • @DmitryGrigoryev: Yes - if it's not properly designed, bad things can happen. But that's true in everything we do, no? The facts are that we have "run out of rope" wrt powering RPi with a "conventional" arrangement" such as we have now. You can't even find a USB cable with 20AWG power wiring; existing PS output voltages have been pushed to the top of what the spec allows, and still RPi4 sees 4.8V at the input connector. Something's gotta' give. – Seamus Jul 29 '20 at 18:45
  • With a conventional PSU or a step-down converter all you risk is not getting enough voltage. If the remote sensing like breaks, the PSU will keep increasing the voltage until the Pi blows up. – Dmitry Grigoryev Jul 29 '20 at 21:41
  • @DmitryGrigoryev: No - not if it's designed properly. No idea what allows you to come to that theory... – Seamus Jul 29 '20 at 22:25

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.