some findings on tinkerboard's power consumption with a newer power supply

xAda · January 13, 2023

probably a post bragging of my christmas gifts, but mostly just documentation because it really confused and stumbled me from the get go

so, my setup was powering tinkerboard (the first revision, without the new fangled eMMCs) via GPIO because it confused the hell out of me how to power this beast. this was powered through the XL4015 buck converter 12.1/12.2 V on input, with ammeter and voltmeter included on the board. (the board was current limited to 5A, because why not? the rk808 is 6A tops)

ok from the get go the board powers (not load tested, just power through and run kodi-gbm UI) from 3.4V to 6.2V. undervolting or overvolting it doesnt shuts itself down (no red light), then bringing back to operating voltages boots the board up

also take not that operating voltages for RK808 was 2.7V to 5.5V. the missing 0.7V was probably the diode drop people are talking about. maybe? without a multimeter (only using the power supply's voltmeter-ammeter) i can't check it that much. but i'm fine that it has wide range of operating voltages.
on all observed voltages, 3A seems to be the max current. it needs to be sustained though. also full load without peripherals (and HDMI, HDMI seems to be active than passive) is observed 2.3-2.5A. booting up the board needs about 0.7A to keep on going

this would probably contradict as well the "increase voltage to 5.25 or smthing" given that no matter how many volts you put into it, it would only use 3A.
weirdly (since this contradicts the block schematics from tinkerboard), #1 and #2 are observed in both USB and GPIO with no difference (significant or not) under load.

it's weird since
3.1. if USB lines were protected, there should be a diode drop *only* on USB lines. but this was seen on GPIO as well.
3.2. some claims microUSB only has max 2.5A but this is not seen in this case: the setup sustained 2.7A going 2.9A.

also, powering both GPIO and USB with the same power block doesnt seem to affect stability at all (still hard limits at 3A.)

so with that, there are few suggestions for powering the board as well (and minor observations. opinionated parts ahead)

people talking about overvolting or undervolting the power rails doesnt really help at all at slightest. it'll still top at 3A. what overvolting could help were powering via USB peripherals, but i wouldnt recommend it plugging straight given the fact that 2.5A on full load? you only have 0.5A to spare on heavy load.

get a usb y splitter (tape over the USB power lines to the board. idk why but i have usb resets whenever i dont have them. then again plugging a hard drive in there shouldnt be happening in the first place)
the raspberry pi 3B+ adaptor (5.25V, 2.5A)? good enough actually! as long as it can deliver 5V across the device, AND as long as you dont have power heavy devices on it. it doesnt matter that much so long as your USB cable can handle phone-charger level of amps. rk3288 was designed with mobile phones in mind. it'll always power at 3.7V. i always having issues when im using a cable that can only do 0.5A across it. no can do.

probably for me, but this is why you dont need to be hung with the minor details. starting from this, im wracking my brain around if tinkerboard was sensitive to power drops (like 5.1v or 5.2v would show stability improvements) but nope! just chuck it right in, it'll be fine. just check your connections and stuff.
powering through the GPIO pins is recommended of course, but it's an overkill. probably the reason for it was most USB chargers only go 2A max and that would underpower the tinkerboard, doing boot loops instead of starting correctly. then again YMMV: my place has 3A or 5A chargers everywhere. and with the introduction and interests of fast chargers, you'll find chargers (and cables) that can deliver more than you expect.
on sidenote on powering through the GPIO: PLEASE GET YOUR GAUGES RIGHT. i powered it via the ATX FDD (remember those?) connector, with 12V pin removed. it turns out for some reason its max current is 1A and its just thin flimsy as hell. get a thicker wire.

it didnt drop the current (thankfully) but the wire became so warm to touch, it's weird. i wouldnt wanna run this on full load. at least i didnt get burned wires. cooling fan helps a lot
also you cant bypass that 3A limit. that's 3A total, both GPIO and USB inputs. even powering both USB and GPIO (like that one person suggested) doesnt really do anything. you wouldnt get that much power in, but at least you can protect your wires?

which is weird since block schematics shows GPIO and USB has different lines? so it should have different protections and stuff. a multimeter can help see if those inputs are shorted (GPIO vs microUSB vs peripherals) but i dont have one at the moment

i would also like to measure the voltage across the peripherals if any load situations will lead to undervolting the outputs, but i dont have a USB tester at the moment.
RK808 seems to be programmable but interfacing with it is outside of my skill set. any help would do

i hope this post helps everyone well. no one uses these boards anymore except for niche use cases but i still enjoy it. i hope in the future i can get myself a huge NAS server with an RK3588S on it. kinda lonely to be honest

Edited January 13, 2023 by xAda
minor details might edit it again