Glock24 Posted January 21, 2021 Posted January 21, 2021 I bought a set of Leviton T5632 in-wall USB chargers before I even thought of tinkering with SBCs. The spec sheet says it delivers 3.6A@5V, so I thought it should be enough to power a board and a couple of USB disks. Has anyone used a similar device to power an SBC? I still don't have my device and I don't have a USB power meter at hand, when I have them I'll update this. I bought a set of 3 for about $25 off eBay, so price is decent. Here are some pics of the insides in case anyone is interested. Board: Not on the list
Werner Posted January 21, 2021 Posted January 21, 2021 Will most likely but there is no guarantee. The general issue with such PSUs is that they are designed to deliver a constant current for charging purpose. However when plugging in devices like SBCs their input current changes depending on the load. If the output capacitance of the PSU is too small it might not be able to compensate and the voltage drops which then leads to unpredictable behaviour. Ripple voltage could also become an issue if the PSU is VERY cheap. Just try'n'n error. Plug 'em in, run a couple of stress tests and check if anything weird occurs. Of not you are good to go.
arox Posted January 21, 2021 Posted January 21, 2021 "I don't have a USB power meter at hand" Then, you should perhaps check voltage on gpio ... Typically, 5V PSU deliver 5.2 V. But if you draw 2 Amps and have 0.1 ohm in the circuit (cable, usb contact, board fuse, switch or meter shunt) you lose 0.2 V. I use a adjustable PSU at 5.28V. I have an INA219 to check voltage and amperage, but my RPI4 is not happy with the additional 0.1 ohm shunt when all cores are solicited and it reach 2 amps. And I should warn everyone that the first symptoms are generally strange network problems ... So : I have some doubts when the manufacturer say you need (their) 4 amp 5V PSU. If the board is badly designed or you use bad cables or you have bad contacts, you never will get a stable voltage even if the PSU provide good regulation. Powering SBC with 5V was all in all a bad idea from the RPI designers : boards need many step down circuit anyway for chips or CPU and step up for Ethernet.
Glock24 Posted February 14, 2021 Author Posted February 14, 2021 I just got my Nanopi M4v2 today and ran a few quick tests, about a week ago my USB power meter arrived also. The first test was powering the board using a Macbook Air 29W USB-C charger and running s-tui to put a full load on the CPUs. The voltage drops to 4.5-4.6 when the amperage goes above 2A. I let it run for a while (about 10 mins) and when a came back to check the board had rebooted. This power adapter is not suited to power this SBC Next test was using a PoE to 5V USB-C adapter rated for 2.4A. Again I ran s-tui and let it go it's thing. After about an hour or was still running. Add for voltage this power supply actually increases voltaje as amperage goes up. It starts around 5.08V with ~400mA power draw and goes up to around 5.26V work the board at full load. With an hdd connected to the USB doing a full disk read and s-tui running I get a max draw of 2.3A with some 2.42A peaks. Tomorrow I'll try the in-wall charger, as right now I can't connect a display near the outlet with the charger.
Glock24 Posted February 14, 2021 Author Posted February 14, 2021 I connected the board to the in-wall USB charger with no monitor attached, logged-in using ssh and ran the test again. The charger seemed solid delivering 5.12V@2.4A I overclocked the CPU to 2GHz and power consumption went up to 2.71A (with USB disk attached doing full read and s-tui stress test), voltage remained stable @5.12V. So this Leviton in-wall USB charger seems ideal to power 5V SBCs.
Recommended Posts