Adequate power supply for the Orange Pi Zero?

[k4in]

I've got an OPi Zero v1.1 for a home project, It was supposed to be something very simple, connect the board to wifi, make a web app for external control and switch a power relay, I did finished it and everything was working fine but I had sudden disconnections and one day just didn't wanted to reconnect, I've read before about how this board has an horrible wifi chip and trying to avoid further headaches I went and bought an USB WiFi dongle TP-LINK_TL-WN725N, I got it today and connected it to the board using its 13 pin header, then I powered it on via micro usb and worked right out the box, no need for extra config, to finish it off I rewired it to this little box where I have it tucked away aaaand didn't work...

 

After testing why I found that the wifi would work only on the microusb power with my laptop, inside the box where it was working before I have this AC/DC 5v 1A module which power the OPi through 5v pin of the expansion header and relay (1 channel model), whenever I tried using the converter in the box the wifi didn't work, looked like it didn't  even power since there was no led activity of the dongle, but the board itself was powered and booted correctly, or at least that is what I believe because the board led would behave the same as when it would boot and connect to wifi.

 

In short, when I power it via microusb usb wifi works, when I use the 5v 1A converter mentioned above it doesn't, and for a last test I disconnected the relay from the circuit to feed just the OPi but wifi still wouldn't work, then I read this answer about usb power, if right it would mean that my laptop isn't supplying more than the ac/dc module, or is it wrong?, also I have video output and onboard wifi disabled, I really don't know what I'm missing here.

[chwe]

Congrats! You're one of the first thinking about non appropriate powering, so we don't have to move the thread in this subforum.   

 

1 hour ago, k4in said:

I read this answer about usb power, if right it would mean that my laptop isn't supplying more than the ac/dc module, or is it wrong?

The USB of my notebook delivers more than 1amp stable on USB, the reason why I never mention it in terms of SBCs is that it is above the specifications for a USB port. Not every notebook will deliver and I don't recommend it at all. 

 

1 hour ago, k4in said:

I have this AC/DC 5v 1A module which power the OPi through 5v pin of the expansion header

1A is a bit low, normally we recommend 2A at 5V. But let's see if we get it reliable working. Cause your use case seems to be not that calculation intensive, it might be OK when you rund your OPi0 with a minimal power consumption setup. I suggest you read through this thread here:

Reduce peak consumptions seems to be important for your case (which could be problematic with wifi ). 

 

For debugging, do you have a USB-TTL dongle so that you can connect to your board in case wifi isn't on but it is powered through your 'box setup'?

 

 

[StuxNet]

I've gotten higher than 1A ratings from a USB port (usb 3.0) before and absolutely gotten less than 1A from a similarly rated 'ac/dc module'

Can never go wrong with getting an appropriate, reliable setup. In my experience, FriendlyArm boards are no where near as fickle (power wise) than OPi boards. I can boot any of them from just about any power source I connect them to. For this reason I prefer the Neo.

OPiZ's are absolutely a different case. For me, they rarely work just from USB, unless I use a specific, good quality cable.  Forget compiling/installing more robust libraries.  I can't speak too much on the relays you mention. I can tell you though that I've purchased 4+ OPiZ boards, all serve a purpose. Require dependable power supply setup. I have a Powerstrip that genuinely pumps out 2 amps + to USB port. So I use that with a good cord and I've never had a power issue since.

Tl,DR; OPiZ over all design quality isn't the best but it's still proven to be a working product. I've purchased many of them, the only power issue I've really run into has been at fault of my own setup.

[k4in]

9 hours ago, chwe said:

The USB of my notebook delivers more than 1amp stable on USB, the reason why I never mention it in terms of SBCs is that it is above the specifications for a USB port. Not every notebook will deliver and I don't recommend it at all.

I guess my laptop is the same, when I connect my phone it doesn't complain about slow charging as it does with other usb ports, but I wasn't sure because of those USB specifications and didn't wanted to waste another power supply to find out it was something wrong with my wiring or somewhere else, but that cleared my doubts.

 

I'll try that minimal consumption config and if that doesn't work I'll throw a higher power supply into it, sadly I dont have a USB-TLL adapter, I saw it when ordering my parts but I thought I wouldn't need it...  oh well that's a lesson for me.

 

3 hours ago, StuxNet said:

I can't speak too much on the relays you mention. I can tell you though that I've purchased 4+ OPiZ boards, all serve a purpose. Require dependable power supply setup. I have a Powerstrip that genuinely pumps out 2 amps + to USB port. So I use that with a good cord and I've never had a power issue since.

 

If the low power config doesn't work I'll power the board with a 2Amp supply independently from the 1Amp powering the relay just for good measure.

 

Thanks a lot for the help!

[martinayotte]

5 hours ago, k4in said:

sadly I dont have a USB-TLL adapter, I saw it when ordering my parts but I thought I wouldn't need it...

Not having any USB-TTL nearby while playing with SBC is like doing any electronics without a multimeter, especially that those are costing only $0.99 ...

[tkaiser]

21 hours ago, k4in said:

2Amp

 

This is a number or a rating. This does not tell anything about 'quality'. On each of the sub-forums there are two topics linked at the top (except here for whatever reasons) and one of those two is this: https://forum.armbian.com/announcement/1-1-check-power-supply-check-sd-card-and-check-other-people-experiences/

 

While 1A is a bit I low anyway I would also do a web search for 'psu noise ripple filter'.

[DaveKimble]

My house is 350 metres from the meter at the front gate, with the cable in a 600 mm trench, so 240 VAC becomes 220 V all the time, 208 V with the kettle on, and even less with the toaster on too.  The area frequently has thunderstorms so surge protection is essential.  5V powerpacks usually have the Amps rating on them, but without some voltage data from the manufacturer and direct feedback from the board, the whole thing is pointless.  Apart from learning that "it doesn't have to be exactly 5.000 V", still no one has answered the question "What is the acceptable range?".  And if the board doesn't tell you continuously what it has actually got, ANYTHING can attributed to a "power failure".

[chwe]

1 hour ago, DaveKimble said:

Apart from learning that "it doesn't have to be exactly 5.000 V", still no one has answered the question "What is the acceptable range?".  And if the board doesn't tell you continuously what it has actually got, ANYTHING can attributed to a "power failure".

okay some starters:

 

 

1 hour ago, DaveKimble said:

And if the board doesn't tell you continuously what it has actually got, ANYTHING can attributed to a "power failure".

Most boards don't have a PMIC which is capable of reporting the DC-In voltage. That's a fact and it's likely that this won't change. Those SoCs are mostly made for TV-Boxes, so the corresponding PMICs are developed with the assumption that the (TV-Box) boardmaker understands how proper powering works (e.g. barrelplug with a known good PSU), unfortunately SBC boardmakers decided that 'the RPi way of powering' (e.g. telling your customers that they can use their crappy phone charger with a crappy USB cable first before 'providing' a 'RPi approved'  microUSB PSU which delivers 2.5A at 5.1V to overcome all those undervoltage situations). And that's why we're where we are.. having boards and dealing with undervoltage issues and trying again and again to convince people that undervoltage might be a reason why *random board* doesn't run as expected. If there would be a known 100% reliable way to detect undervoltage on each board we support, I'm sure @tkaiser would've implemented it the day he knows that it's working to save himself a bunch of time explaining again and again (and again) why *unexpected error* is likely related to underpowering. If the error would look exactly the same everytime, I'm sure such a script would also be shared, underpowering can have all sorts of funny errors and it's an error which can be avoided buy being not the scrooge trying to save 2-3 bucks by using random crap. 

If random OPi board has an barrel-plug, buy opis PSU my OPi PC+ works over a year 24/7 without issues (headless). If you plan to stick random crap via USB to it probably with turned on wifi as well and then mix it with some heavy tasks on the CPU, even the one provided by xunlong my f*ck up and quit. If you've an microUSB powered one (e.g. OPi Zero etc.) go for good microUSB cables (e.g. ~AWG20 for powerwires - you can even find sometimes such cables on aliexpress). I use 1m cables branded 'Fonken' with AWG21 powerlines and AWG28 for datalines together with a Ikea handycharger named 'KOPPLA' and never got fooled by them (doesn't mean that nobody ever will be fooled by this combination but at least headless OPis worked quite well). If you consider to stick random SSDs or HDDs on a microUSB powered SBC over USB powered by the board itself, please reconsider your plan or use Y cables and ask @TonyMac32 when his  "mezzanine" board:

will be available (he might need some testing not that he gets an article in the Register similar to team RPi https://www.theregister.co.uk/2018/09/11/raspberry_pi_poe_hat_issue/ ).  Or build your own GPIO powered solution.

Sometimes a simple Multimeter can help you to measure between GND and a random 5V pin on the pinheader to get a clue if the average voltage is okay.. Some of the shitty PSUs have low voltage spikes which aren't detected by an el Cheapo multimeter which makes debugging even harder. Some SoCs and their boards may be more sensitive to undervoltage than others. But the solution is simple don't save 2-3 bucks by going for the cheapest one you find... Btw. I assume that RPi's official PSU might also be considered as a 'good one' cause it was developed after team RPi realized that average phonechargers are to crappy to power their board reliable... 

[iamwithstupid]

I'd say one could simplfy that all and recommend good power supply manufacturers like Meanwell and Sunny from the start

 

which provide datasheets for their powersupplies (for example here) and properly messure their performances, efficiency and reliability.

 

Also a simple hint that you should optimally run a power-supply at only up to 80% of their max range to ensure that they work properly (super simplification, but you know ... safe margins work best with starters)

 

@tkaiser maybe a recommendation here (not only for Barrel-Plug connectors, but also for powering via Headers) like this could save you guys also a lot of headaches regarding power supply questions ?

 

That said - bottomline - don't use USB-Powersupplies like mentioned in the thread above.

 

You're just asking for troubles - especially those kind of detachable ones where you can use another USB-Cable.

 

If you really want an recommendation I'd look at original ones from phone manufacturers like Oneplus or Xiaomi (certainly NOT Samsung, with their ultra crappy power supplies) ...

[chwe]

16 hours ago, iamwithstupid said:

That said - bottomline - don't use USB-Powersupplies like mentioned in the thread above.

 

You're just asking for troubles - especially those kind of detachable ones where you can use another USB-Cable.

 

and that's wrong as well.. The OPi0 isn't a super powerhungry board which needs a special treatment to work reliable.  A reliable 2amp PSU combined with a good USB cable will work without issues. If you then start to combine this with all sort of other fancy add-ons (e.g. usb attached storage powered through the on board USB port) things will go worse. I don't promote microUSB powered boards, there are better boards in a similar price range. But those boards work as well, as long as you're aware of the issues.

Even my tinkerboard runs now since ~1 month USB powered (headless as a small git server). All repositories on this small server get a daily health check (e.g git fsck) - until yet, no signs that something isn't in a healthy state. It all depends on being aware that there can be an issue (I wouldn't recommend the tinker as a git server, but I have it spare and the gui I use can be memory hungry 2gb in a 32bit OS seems to be sufficient). I don't know how long @tkaiser let his OPi0 serve as a torrent server for armbian but I assume it was/is also micro USB powered.

 

bottomline - think first what your use-case is. Powerhungry board combined with powerhungry peripherals, microUSB isn't the way to go. Every crappy mobilephone charger as well, they're simply not made for it. Keep in mind they're made for charging a phone which has an more advanced PMIC - they can mostly deal with voltage drops (e.g. my android phone combined with a crappy charger under heavy usage reports that charging isn't sufficient). Those simpler SBC PMICs don't - but as long as you threat them as you should (e.g. check that the voltage which arrives on the SBC, a multimeter can be helpful) they will run without headache. 

[tkaiser]

9 minutes ago, chwe said:

I don't know how long @tkaiser let his OPi0 serve as a torrent server for armbian but I assume it was/is also micro USB powered

 

Powered via Micro USB from an USB port of the router next to it (somewhat recent Fritzbox). Storage was a 128 GB SD card, no peripherals, cpufreq settings limited the SoC to 1.1V (fixed 912 MHz) and so maximum consumption was predictable anyway (way below 3W). I could've allowed the SoC to clock up to 1.2GHz at 1.3V but performance with this specific use case (torrent server) was slightly better with fixed 912 MHz compared to letting the SoC switch between 240 MHz and 1200 MHz as usual.

 

I used one of my 20AWG rated Micro USB cables but honestly when it's known that the board won't exceed 3W anyway every Micro USB cable is sufficient.

[k4in]

On 9/24/2018 at 8:24 AM, iamwithstupid said:

That said - bottomline - don't use USB-Powersupplies like mentioned in the thread above.

 

You're just asking for troubles - especially those kind of detachable ones where you can use another USB-Cable.

 

On 9/25/2018 at 2:36 AM, chwe said:

bottomline - think first what your use-case is. Powerhungry board combined with powerhungry peripherals, microUSB isn't the way to go. Every crappy mobilephone charger as well, they're simply not made for it. Keep in mind they're made for charging a phone which has an more advanced PMIC - they can mostly deal with voltage drops (e.g. my android phone combined with a crappy charger under heavy usage reports that charging isn't sufficient). Those simpler SBC PMICs don't - but as long as you threat them as you should (e.g. check that the voltage which arrives on the SBC, a multimeter can be helpful) they will run without headache. 

 

ha! I just wanted to add, as a follow up of my original post, that is exactly what I did, I bought a cheap noname phone charger of 2amps and that was enough to power the board (connected directly to the expansion header), power relay and wifi dongle, granted as explained before my use case is really simple and the board is waiting idle most of the time.

Up time:       57 days

Last downtime was due to power outage, I wouldn't know if that power supply is insufficient on heavy load but I'll keep it in mind.

 

I did have some wifi disconnections but those were because of the driver power management options, I fixed it adding these options to a module config file

cat /etc/modprobe.d/8188eu.conf 
options 8188eu rtw_power_mgnt=0 rtw_enusbss=0 rtw_ips_mode=1

So far it has been running good since then. 

[bayindirh]

Just wanted to add some, albeit not very scientific, testing and observation results for further reference.

 

TL;DR: Do not use 1A power supplies with OrangePi Zero. The board is small, but with ethernet connected, it really needs and shines with 2.0A and more power to spare.

 

Long Version

I'm running an OrangePi Zero 1.4 with Armbian 16.04LTS. It runs DNSMasq for home network and also acts as a rsync/backup server for couple of android phones. I've powered the board with a high quality Artwizz 1.0A USB power supply around a year. Sometimes I experienced small problems like DNSMasq becoming non-responsive on a bi-weekly basis, or the SSH connections fail to the board rarely. Also, the board was working a bit sluggish. I attributed the problems to the CPU power and the overall quality of the board, but it was bothering me nevertheless. All in all, the uptime and general stability of the system was acceptable from that kind of board though.

 

Yesterday, after losing SSH connection again, I thought about increasing the power budget of the board, then I found this thread and replaced the adapter with an 2.5A one (I'm a kind of engineer who over-engineers for reliability and headroom). DNSMasq is visibly snappier, the temps are down 2-3 degrees Celsius.

 

I will observe the system more and will update here with long term test results.

Edited January 3, 2019 by bayindirh
better word choices.