Jump to content


Photo

NanoPI NEO / AIR


451 replies to this topic

#121 tkaiser

tkaiser

    Advanced Member

  • Moderators
  • 2660 posts

Posted 12 August 2016 - 07:41 AM

Full debug output (a bit scary regarding ehci_irq): http://sprunge.us/SZaR (on the other NEO with my patched kernel currently doing DRAM consumption tests there is no such message -- strange)

 

On a related note: just figured out that the physically exposed USB port on NanoPi NEO is usb3 and not usb1 as assumed. So maybe these USB messages will disappear when disabling usb1 and usb2 in fex file?

 

Will test that another time, now I'm trying out 624 MHz DRAM clockspeed with lima-memtester. Yesterday I tried it without fan and just FA's heatsink. Board pretty soon deadlocks even with low DRAM clockspeeds (maybe other components on the SoC might overheat as well like voltage regulators?). When I went to bed I let it run with active fan blowing above the heatsink and between heatsink and PCB at 672 MHz but this deadlocked too during the night.


Please don't send personal messages! Use the forum so others can participate and benefit!

 

Before you report any problem please be aware that crappy SD cards and insufficient power supply are reason N° 1 why things are failing. Try to rule this out first please, check 'getting started' recommendations and check/provide 'sudo armbianmonitor -u' output first!

 

Did you check out custom google powered forum search already (before opening new threads or asking questions)?


#122 cb88

cb88

    Member

  • Senior Members
  • PipPip
  • 11 posts

Posted 12 August 2016 - 02:30 PM

That is a bit of myth in some people's thinkings. If they were acting like that, it would means that LED is in conducting mode.

Same thing applied when LED is connected in Sink mode, it doesn't act as PullUp, otherwise it will glow ... :rolleyes:

While yes it is a stretch... depending on the LED's voltage drop it could be preventing the IO line from switching logic levels. With the LED's voltage drop being approximately somewhere between 3.2 and 0.7 volts. Also the "glowing" would only be with a sufficent current... which a floating IO isn't going to provide.

 

And yes I have seen glowing LEDs but not in this particular situation ;-) (PCB mounted on a conduit also connected to a 480AC drive.. it mostly went away after grounding the conduit)



#123 tkaiser

tkaiser

    Advanced Member

  • Moderators
  • 2660 posts

Posted 15 August 2016 - 05:49 PM

Since FriendlyARM sent me a few NEOs and also two of their heatsinks for NEO I did a test today: I used new preliminary Armbian settings for NEO, tested those out 3 times on the NEO and later the same image with identical settings on 2 Orange Pis (all Armbian images for any H3 device run on any other H3 board with only some minor glitches -- in this case there weren't any since hardware is too identical)

 

Bildschirmfoto%202016-08-15%20um%2018.31

 

From left to right:

  • NanoPi NEO/256 w/o heatsink lying flat on a table, SoC/DRAM on the lower PCB side so no airflow possible (~480 MHz average throttling)
  • NanoPi NEO/256 with FriendlyARM's own heatsink operated vertically to let convection help somehow (~840 MHz average throttling)
  • NanoPi NEO/256 with tkaiser's standard H3 heatsink operated vertically (~690 MHz average throttling)
  • OPi Lite with tkaiser's standard H3 heatsink operated vertically (~900 MHz average throttling)
  • OPi PC with tkaiser's standard H3 heatsink operated vertically (~980 MHz average throttling)

Using FA's own heatsink is an improvement compared to cheap heatsinks both regarding heat dissipation as well as stability (FA's heatsink is mounted perfectly and board + heatsink are ready for heavy vibrations). But as tests with OPi Lite and OPi PC show obviously PCB size and condition matter (copper layers inside PCB and the larger the size the better the heat dissipation -- Orange Pi Plus 2E for example performs better under identical conditions most probably due to its larger PCB size) 

 

In case you want to buy NEO (or NanoPi Air later -- I still believe they share the form factor and heatsink) you better order FA's heatsink too if you plan to operate the device under constant high load (which is IMO not being made for!). Regarding my 'standard H3 heatsink': That's the one that can be seen on these pictures and costs below $0.5 when ordering a few on Aliexpress.

 

The workload used for the test was cpuburn-a7. We don't know of any other program that is able to heat up H3 more (so please consider this as a 'worst case' workload you'll never be able exceed with normal 'real world' tasks since the NEO has no display output and GPU cores can't be used -- that's the only known exception to cpuburn-a7: using lima-memtester with Mali400 MP being clocked at 600 MHz like Armbian does. Then SoC temperature increases even more)


Please don't send personal messages! Use the forum so others can participate and benefit!

 

Before you report any problem please be aware that crappy SD cards and insufficient power supply are reason N° 1 why things are failing. Try to rule this out first please, check 'getting started' recommendations and check/provide 'sudo armbianmonitor -u' output first!

 

Did you check out custom google powered forum search already (before opening new threads or asking questions)?


#124 rm-rfstar

rm-rfstar

    Newbie

  • Senior Members
  • Pip
  • 9 posts

Posted 16 August 2016 - 01:37 AM

my nanopi neo arrived today and i got a chance to try the new heatsinks i purchased off aliexpress for about $10 (20 heatsinks)
 
 
i can't seem to post images because of my low post count, but here is google+ album with details
 
 
TLDR: 41c at idle, 74~75c at full load (using cpuburn-a7)
 
hopefully someone finds this interesting. also notice that the heatsink doesn't quite fit over the soc and memory, so i split the difference.
 
Regards!


#125 tkaiser

tkaiser

    Advanced Member

  • Moderators
  • 2660 posts

Posted 16 August 2016 - 12:45 PM

TLDR: 41c at idle, 74~75c at full load (using cpuburn-a7)

 

These tests should at least run 20 minutes (better 30) to get the idea about worst case conditions (since we experience with the NEO that the whole PCB and all components starts to heat up -- I've been able to burn my fingers yesterday when I did some cpuburn-a7 tests afterwards).

 

BTW: Another more realistic approach to test the performance efficiency of a specific heatsink is to use cpuminer (a bitcoin miner that can use NEON instructions and knows a benchmark mode). Simply grab https://sourceforge....tar.gz/download then untar it, chdir into cpuminer-2.4.5 directory and then:

sudo apt-get install libcurl4-gnutls-dev
./configure CFLAGS="-O3 -mfpu=neon"
make
./minerd --benchmark

But please don't be surprised that performance numbers reported will be lower compared to other H3 devices. NEO uses a single bank DRAM configuration and DRAM clockspeed is way lower than on all other H3 boards. Therefore performance will be lower anyway but using cpuminer's benchmark mode you might get the idea how different heatsink/cooling solutions 'perform'. But to be honest: NEO is not made for performance anyway so better use the heatsink as a simple matter of precaution and forget about benchmarking this tiny board at all :)

 

In case anyone wants to build a HPC cluster with NEOs (weird to say the least ;) ). I prepared an archive some time ago to do reliability testing with Pine64 that contains already a script to collect cpuminer benchmark numbers and feeds them into RPi-Monitor template therefore the efficiency of the cooling approach in question can be measured/compared directly: http://linux-sunxi.o...sting_on_Pine64 (see the screenshot there to get the idea)


  • rm-rfstar likes this

Please don't send personal messages! Use the forum so others can participate and benefit!

 

Before you report any problem please be aware that crappy SD cards and insufficient power supply are reason N° 1 why things are failing. Try to rule this out first please, check 'getting started' recommendations and check/provide 'sudo armbianmonitor -u' output first!

 

Did you check out custom google powered forum search already (before opening new threads or asking questions)?


#126 rm-rfstar

rm-rfstar

    Newbie

  • Senior Members
  • Pip
  • 9 posts

Posted 16 August 2016 - 02:13 PM

These tests should at least run 20 minutes (better 30) to get the idea about worst case conditions (since we experience with the NEO that the whole PCB and all components starts to heat up -- I've been able to burn my fingers yesterday when I did some cpuburn-a7 tests afterwards).

 

 

 

 

 

 Thanks for the insight. I have a question about the FA supplied heatsink, how did FA avoid hitting the Ethernet jack posts? I came pretty close to hitting them myself when placing a heatsink on the CPU. Is the heatsink contoured on the bottom?



#127 tkaiser

tkaiser

    Advanced Member

  • Moderators
  • 2660 posts

Posted 16 August 2016 - 02:44 PM

I have a question about the FA supplied heatsink, how did FA avoid hitting the Ethernet jack posts? I came pretty close to hitting them myself when placing a heatsink on the CPU. Is the heatsink contoured on the bottom?

 

Nope, their heatsink uses some sort of spacers and a silicone heatpad in between (they're said to be not that efficient but this pad + large heatsink show better heat dissipation than my usual cheap 14x14mm heatsink). Just take a look at the pictures: http://linux-sunxi.o...Pi_NEO#Pictures

 

BTW: I would fear that placing such a huge heatsink like yours will both cause troubles to get enough contact between SoC's and heatsink's surfaces and might short things out...


Please don't send personal messages! Use the forum so others can participate and benefit!

 

Before you report any problem please be aware that crappy SD cards and insufficient power supply are reason N° 1 why things are failing. Try to rule this out first please, check 'getting started' recommendations and check/provide 'sudo armbianmonitor -u' output first!

 

Did you check out custom google powered forum search already (before opening new threads or asking questions)?


#128 rm-rfstar

rm-rfstar

    Newbie

  • Senior Members
  • Pip
  • 9 posts

Posted 16 August 2016 - 03:24 PM

spacing is an issue, but doable. I thinking about cutting one of the heatsinks in half and placing it on the memory. Has anyone measured the memory temp when using a benchmarking/burning tool?



#129 tkaiser

tkaiser

    Advanced Member

  • Moderators
  • 2660 posts

Posted 16 August 2016 - 04:10 PM

Two PRELIMINARY TEST images available here with latest settings. Xenial/Jessie headless: http://kaiser-edv.de/tmp/6Mmziq/ UPDATE: You can grab the test images from official download page instead.

 

 

 _   _                   ____  _   _   _            

| \ | | __ _ _ __   ___ |  _ \(_) | \ | | ___  ___  

|  \| |/ _` | '_ \ / _ \| |_) | | |  \| |/ _ \/ _ \ 

| |\  | (_| | | | | (_) |  __/| | | |\  |  __/ (_) |

|_| \_|\__,_|_| |_|\___/|_|   |_| |_| \_|\___|\___/ 

                                                    

 

Welcome to ARMBIAN Debian GNU/Linux 8 (jessie) 3.4.112-sun8i 

System load:   0.06            Up time:       2 min

Memory usage:  14 % of 241Mb  IP:            192.168.83.25

CPU temp:      31°C           

Usage of /:    15% of 7.2G   

 

 

New to Armbian? Check the Armbian H3 Mini FAQ first:

 https://github.com/i.../H3_mini_faq.md

Changing password for root.

(current) UNIX password: 

Enter new UNIX password: 

Retype new UNIX password: 

 

Thank you for choosing Armbian! Support: www.armbian.com

 

 

 

In case you want to give any feedback, please try these out now and report back in this thread (not via PM!). Areas of testing

  • USB (all USB ports enabled)
  • Audio (enabled, not tested, see this config tweak for details and please report back)
  • USB gadget stuff
  • Booting behaviour with noisy DC sources (check dmesg and read through settings link above in case you encounter problems)
  • thermal behaviour

BTW: The jessie image should also be appropriate for users that use NEO/512 with Ethernet/USB unpopulated since on this image an USB-Ethernet dongle connected to Micro USB should work as eth0 out of the box (didn't work with Xenial for me -- no idea why).

 

Maybe we will enable g_serial for NEO later like we did for OPi Lite -- up to Zador :)


Please don't send personal messages! Use the forum so others can participate and benefit!

 

Before you report any problem please be aware that crappy SD cards and insufficient power supply are reason N° 1 why things are failing. Try to rule this out first please, check 'getting started' recommendations and check/provide 'sudo armbianmonitor -u' output first!

 

Did you check out custom google powered forum search already (before opening new threads or asking questions)?


#130 cb88

cb88

    Member

  • Senior Members
  • PipPip
  • 11 posts

Posted 17 August 2016 - 12:14 AM

Is anyone running the mainline kernel?

 

3.4 doesn't have the i2c driver for ssd1306 oled displays (fbtft has a driver for this chip but only SPI, mainline has a driver for SSD130X).... and I don't care about anything currently missing it mainline I think (except for ethernet which I don't *really* need I could always use a USB ethernet dongle if needed)

 

I mainly want to run low power, have the USB3 port working, and display output to the oled.



#131 Peter Scargill

Peter Scargill

    Newbie

  • Senior Members
  • Pip
  • 8 posts

Posted 17 August 2016 - 05:26 PM

I wonder if anyone can help..

 

I have Armbian for the NEO - indeed I've now tried yesterday's version and I get the same thing every time (I can use FriendlyArm's Ubuntu but I don't want Ubuntu - so the hardware is fine)...

 

When I try to use WinSCP in Windows - which works for everything normally, using root and 1234,  I see it try to login - for an instant I see it has succeeded and then instead of getting the normal window full of files in the root directory - WinSCP just goes back to the login dialog with no pane of files coming up... as if it had failed to login but but clearly it is not failing as the view is very different if I deliberately put the wrong password in.

 

Never had this happen before so I'm at a bit of a loss. Anyone seen this happen?

 

Pete.



#132 zador.blood.stained

zador.blood.stained

    Advanced Member

  • Moderators
  • 1609 posts

Posted 17 August 2016 - 05:41 PM

When I try to use WinSCP in Windows - which works for everything normally, using root and 1234,  I see it try to login - for an instant I see it has succeeded and then instead of getting the normal window full of files in the root directory - WinSCP just goes back to the login dialog with no pane of files coming up... as if it had failed to login but but clearly it is not failing as the view is very different if I deliberately put the wrong password in.

You need to log in first (via SSH, serial console or keyboard+display) and change password for root (it will be triggered automatically), it is marked as expired by default.


Spoiler

#133 pezi

pezi

    Newbie

  • Senior Members
  • Pip
  • 6 posts

Posted 17 August 2016 - 05:50 PM

Hi!

 

 


When I try to use WinSCP in Windows - which works for everything normally, using root and 1234,  I see it try to login - for an instant I see it has succeeded and then instead of getting the normal window full of files in the root directory - WinSCP just goes back to the login dialog with no pane of files coming up... as if it had failed to login but but clearly it is not failing as the view is very different if I deliberately put the wrong password in.

 

I think the problem is caused by the fact, that the first login on the system causes a request changing the root passwort. Please login on the NEON via putty. Change the root passwort. This should solve your problem.

 

Peter



#134 Peter Scargill

Peter Scargill

    Newbie

  • Senior Members
  • Pip
  • 8 posts

Posted 17 August 2016 - 05:55 PM

You need to log in first (via SSH, serial console or keyboard+display) and change password for root (it will be triggered automatically), it is marked as expired by default.

#
 

Hi 

 

I am trying to login - via WinSCP  (scp).  I can't log in via keyboard and screen as the Neo doesn't have a screen. I would have expected then - in a terminal - to change the password.



#135 Peter Scargill

Peter Scargill

    Newbie

  • Senior Members
  • Pip
  • 8 posts

Posted 17 August 2016 - 06:00 PM

#
 

Hi 

 

I am trying to login - via WinSCP  (scp).  I can't log in via keyboard and screen as the Neo doesn't have a screen. I would have expected then - in a terminal - to change the password.

Actually yes, PUTTY did work - hmmm  wonder why WinSCP wasn't having it - ok - thank you for solving the problem.



#136 zador.blood.stained

zador.blood.stained

    Advanced Member

  • Moderators
  • 1609 posts

Posted 17 August 2016 - 06:05 PM

I can't log in via keyboard and screen as the Neo doesn't have a screen.

Sorry, didn't look at the title of this topic long enough  :)

 

Actually yes, PUTTY did work - hmmm  wonder why WinSCP wasn't having it - ok - thank you for solving the problem.

SCP/SFTP protocols (and WinSCP as a result) don't support any interactive log in processes like forced password change or using "sudo" to use superuser privileges when logging in as non-root user.


Spoiler

#137 Peter Scargill

Peter Scargill

    Newbie

  • Senior Members
  • Pip
  • 8 posts

Posted 17 August 2016 - 06:32 PM

Sorry, didn't look at the title of this topic long enough  :)

 

SCP/SFTP protocols (and WinSCP as a result) don't support any interactive log in processes like forced password change or using "sudo" to use superuser privileges when logging in as non-root user.

Thank you for that - one learns something every day...   Well, it's now up and running and installing my home control script - so hopefully soon I'll have Node-Red, SQLITE and MOSQUITTO running on it - fingers crossed.



#138 rm-rfstar

rm-rfstar

    Newbie

  • Senior Members
  • Pip
  • 9 posts

Posted 17 August 2016 - 09:48 PM

Thank you for that - one learns something every day...   Well, it's now up and running and installing my home control script - so hopefully soon I'll have Node-Red, SQLITE and MOSQUITTO running on it - fingers crossed.

 

not sure if i missed a step, but i'm compiling node from scratch having failed to launch node-red because of node missing. it's taking some time since it's only using 1 core. not sure if i can specify (using make) more than one core.



#139 Peter Scargill

Peter Scargill

    Newbie

  • Senior Members
  • Pip
  • 8 posts

Posted 18 August 2016 - 05:18 PM

One more question - all running with NEO - seems fine -  erm - why in the setup for Armbian do you ask about screen resolution - given that the board does not have a video output?  Am I right in saying that the graphical environment is not installed?



#140 Igor

Igor

    Administrator

  • Administrators
  • 2888 posts

Posted 18 August 2016 - 05:22 PM

One more question - all running with NEO - seems fine -  erm - why in the setup for Armbian do you ask about screen resolution - given that the board does not have a video output?  Am I right in saying that the graphical environment is not installed?

This is about to be fixed with this commit ... and will be in action within next image / first proper one. Current one is officially called prerelease, since we just got the boards.


Want to send me a private message? Stop and think again. If there’s no reason for it to be private, it should be public. Thank you!


#141 tkaiser

tkaiser

    Advanced Member

  • Moderators
  • 2660 posts

Posted 18 August 2016 - 05:23 PM

One more question - all running with NEO - seems fine -  erm - why in the setup for Armbian do you ask about screen resolution

 

Thx for mentioning it, fixed by Igor already: https://github.com/i...463bb5aadb54a0c

 

Please remember, the most current NEO release is still called 'preview' -- we want to collect as much feedback as possible to provide then a mature image for NEO.


  • Igor likes this

Please don't send personal messages! Use the forum so others can participate and benefit!

 

Before you report any problem please be aware that crappy SD cards and insufficient power supply are reason N° 1 why things are failing. Try to rule this out first please, check 'getting started' recommendations and check/provide 'sudo armbianmonitor -u' output first!

 

Did you check out custom google powered forum search already (before opening new threads or asking questions)?


#142 Serge B

Serge B

    Newbie

  • Senior Members
  • Pip
  • 2 posts

Posted 18 August 2016 - 08:55 PM

Thank you for that - one learns something every day...   Well, it's now up and running and installing my home control script - so hopefully soon I'll have Node-Red, SQLITE and MOSQUITTO running on it - fingers crossed.

I tried last two days to install Node-Red, Mosquitto and SQLITE with you script on debian jessie image and failed. It failed on gyp rebuild stage twice. Will try it today with xenial image.

 

Edit. I have read that ubuntu is not supported by the script. I've tried 3. time with clean jessie image and it worked! Node Red is up. Update the webmin as suggested was a bad idea (. Look  the next evening on it.



#143 Peter Scargill

Peter Scargill

    Newbie

  • Senior Members
  • Pip
  • 8 posts

Posted 18 August 2016 - 10:16 PM

I tried last two days to install Node-Red, Mosquitto and SQLITE with you script on debian jessie image and failed. It failed on gyp rebuild stage twice. Will try it today with xenial image.

Hang fire - Xenial - isn't that a version of Ubuntu? It is highly unlikely to work on that, it was put together for the Raspberry Pi. I have it working on the NEO but there are comments inside for non-Raspberry Pi use - like making sure you have a Pi user, that it belongs to various groups and that you run the script from the pi home as pi.  If you don't do all of that there's no chance of it working.



#144 Zammy

Zammy

    Member

  • Senior Members
  • PipPip
  • 15 posts

Posted 19 August 2016 - 11:22 AM

Hi,

 

could you please explain your use cases for NanoPi Neo? 

I was thinking of buying one, but then found out that I can do same things with Raspberry Pi Zero. 

 

For the price of 11 EUR or 12.45 USD (shipping included) you can have same functionality as IoT node/modern linux core and processor that needs minimal passive or almost no cooling.

If you consider the price of Neo 512Mb version 10 USD + radiator 3 USD (overpriced, probably) = 13 USD and ability to run on 4 cores only small periods of time, bigger vertical size we end up with a draw or very specific usage. 

 

4 cores of Neo for me would combine perfectly with computer vision, but it lacks camera connector. 


Spoiler

#145 tkaiser

tkaiser

    Advanced Member

  • Moderators
  • 2660 posts

Posted 19 August 2016 - 11:56 AM

If you consider the price of Neo 512Mb version 10 USD + radiator 3 USD (overpriced, probably) = 13 USD and ability to run on 4 cores only small periods of time, bigger vertical size we end up with a draw or very specific usage.

 

Sorry, but that's simply not true :)

  • if you want the NEO as featureless as a RPi Zero simply order it like that (FA sells the 512MB variant with Ethernet and USB jack unsoldered -- you then still have 1 real Ethernet port ready for soldering experiments compared to zero on the Zero, still 3 real USB host ports ready for soldering experiments compared to zero on the Zero and 1 Micro USB port that can be used for both communitation and powering just like on the Zero -- clear win for the NEO ;) )
  • NEO works wonderfully with no or only an el cheapo $0.28 heatsink with new Armbian settings (the heatsink I used in post #123 above had this price -- more info on heatsinks here)
  • The workload above to demonstrate efficiency of different heatsinks in post #123 was cpuburn-a7 making use of specialized NEON instructions. You'll never find any other workload that is only half that heavy regarding temperatures and consumption. With normal workloads you'll be able to benefit from all 4 cores but unlike on RPi 2 and 3 you can disable cores to save energy on demand, at least with RPi 3 you loose performance and increase consumption -- see post #5 here, also for general consumption numbers and tips&tricks for various RPi and H3 boards!)
  • Unfortunately the NEO shows higher idle consumption than similar H3 boards with identical settings (Orange Pi One/Lite/PC/PC Plus or NanoPi M1 all consume less) and also temperature increases a lot more. So if you want a H3 consumption and thermal champ better choose the 2 little Oranges instead (or wait whether FA will produce a 2nd PCB revision that shows better consumption and thermal behaviour)

You'll find a lot of information regarding consumption / performance in the aforementioned thread -- definitely worth a read if you care about that (even if you only deal with Raspberries)

 

Regarding camera: all Orange Pi have a camera connector, since a few weeks we are able to use Allwinner's HW video encoder for video (freeing up the CPU completely just like with Raspberry's raspivid that uses VideoCore for h.264 encoding!) and unfortunately the camera cable necessary to connect an RPi Zero to any of the supported camera modules costs as much as the Zero itself: £4 for the Zero and another £4 to connect the camera :(

 

In my personal opinion the NEO is nice due to its size and the ability for tinkerers to solder the stuff they need while the device needs less space than a RPi Zero (simply order without Ethernet/USB jack soldered!). There are some drawbacks (consumption, temperature, deadlocks when running full load at higher clockspeeds without active cooling) but we tried to compensate from this with new Armbian settings (most importantly lowering default cpufreq and DRAM clock settings). That being said the NEO might make up for a nice IoT node as long as you keep in mind that it's not suited for heavy workloads (still magnitudes faster than any single core RPi). If you need better performance and lower consumption while still being able to use the whole H3 feature set then better choose NanoPi M1 or any of the smaller Orange Pis.

 

And for an overall comparison between Raspberries and H3 boards like the NEO better take an hour and read through the aforementioned thread which might explain a lot (or only just how to save energy on which device at which 'price' ;) )


Please don't send personal messages! Use the forum so others can participate and benefit!

 

Before you report any problem please be aware that crappy SD cards and insufficient power supply are reason N° 1 why things are failing. Try to rule this out first please, check 'getting started' recommendations and check/provide 'sudo armbianmonitor -u' output first!

 

Did you check out custom google powered forum search already (before opening new threads or asking questions)?


#146 Zammy

Zammy

    Member

  • Senior Members
  • PipPip
  • 15 posts

Posted 19 August 2016 - 02:52 PM

tkaiser, I see your point. I might give a Neo without Ethernet and USB a try.


Spoiler

#147 YonNomNom

YonNomNom

    Member

  • Senior Members
  • PipPip
  • 10 posts

Posted 20 August 2016 - 03:11 AM

Tkaiser, I have been the latest Xenial image you put out and it looks like it's thermals are better than the previous version. 

 

But the unit still will just thermal runaway and crash if there is no airflow.

 

I have disabled the 3 extra cores that I don't need but it makes not effect on the thermals also only a 30Ma reduction in current draw.

 

17:42:57:  240MHz  0.43  12%   9%   0%   2%   0%   0%   79°C
 
and crash. 


#148 HarfTarf

HarfTarf

    Member

  • Senior Members
  • PipPip
  • 11 posts

Posted 22 August 2016 - 09:41 PM

I am having issues using a usb wifi device:

root@neo:~# dmesg | grep rt2
[    0.811567] sw_uart_get_devinfo()1503 - uart2 has no uart_regulator.
[    1.592428] uart2: ttyS2 at MMIO 0x1c28800 (irq = 34) is a SUNXI
[   14.352607] Registered led device: rt2800usb-phy0::radio
[   14.352852] Registered led device: rt2800usb-phy0::assoc
[   14.353151] Registered led device: rt2800usb-phy0::quality
[   14.353452] usbcore: registered new interface driver rt2800usb
[  856.017333] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
[ 1100.839666] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
[ 1357.002345] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
[ 1362.002895] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
[ 1367.000936] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
[ 1372.002929] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
[ 1378.006304] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
[ 1384.001270] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
[ 1388.004151] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
[ 1393.005972] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
[ 1404.001861] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
[ 1408.001750] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
[ 1412.002048] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
[ 1418.001127] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
[ 1423.007031] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
root@neo:~# lshw -c network
  *-network:0             
       description: Ethernet interface
       physical id: 9
       logical name: eth0
       serial: 22:43:d7:1f:63:8e
       size: 100Mbit/s
       capacity: 100Mbit/s
       capabilities: ethernet physical tp mii 10bt 10bt-fd 100bt 100bt-fd autonegotiation
       configuration: autonegotiation=on broadcast=yes driver=sunxi_geth driverversion=SUNXI Gbgit driver V1.1 duplex=full ip=192.168.10.29 link=yes multicast=yes port=MII speed=100Mbit/s
  *-network:1 DISABLED
       description: Wireless interface
       physical id: a
       bus info: usb@4:1
       logical name: wlx000f6004d395
       serial: 00:0f:60:04:d3:95
       capabilities: ethernet physical wireless
       configuration: broadcast=yes driver=rt2800usb driverversion=3.4.112-sun8i firmware=N/A link=no multicast=yes wireless=IEEE 802.11bgn
root@neo:~# 

 

It seems as though the firware blob cant be found.

 

 

Any suggestions?

 



#149 HarfTarf

HarfTarf

    Member

  • Senior Members
  • PipPip
  • 11 posts

Posted 22 August 2016 - 09:58 PM

Nevermind, I found the solution:

root@neo:~# apt install linux-firmware
Reading package lists... Done
Building dependency tree       
Reading state information... Done
The following package was automatically installed and is no longer required:
  libnl-route-3-200
Use 'apt autoremove' to remove it.
The following NEW packages will be installed:
  linux-firmware
0 upgraded, 1 newly installed, 0 to remove and 0 not upgraded.
Need to get 0 B/32.8 MB of archives.
After this operation, 124 MB of additional disk space will be used.
(Reading database ... 58450 files and directories currently installed.)
Preparing to unpack .../linux-firmware_1.157.3_all.deb ...
Unpacking linux-firmware (1.157.3) ...
dpkg: error processing archive /var/cache/apt/archives/linux-firmware_1.157.3_all.deb (--unpack):
 trying to overwrite '/lib/firmware/ti-connectivity/wl18xx-fw-4.bin', which is also in package armbian-firmware 5.17
dpkg-deb: error: subprocess paste was killed by signal (Broken pipe)
Errors were encountered while processing:
 /var/cache/apt/archives/linux-firmware_1.157.3_all.deb
E: Sub-process /usr/bin/dpkg returned an error code (1)
root@neo:~# apt download linux-firmware
Get:1 http://ports.ubuntu.com xenial-updates/main armhf linux-firmware all 1.157.3 [32.8 MB]
Fetched 32.8 MB in 11s (2,872 kB/s)                                                                                                          
W: Can't drop privileges for downloading as file '/root/linux-firmware_1.157.3_all.deb' couldn't be accessed by user '_apt'. - pkgAcquire::Run (13: Permission denied)
root@neo:~# cd /tmp/
root@neo:/tmp# apt download linux-firmware
Get:1 http://ports.ubuntu.com xenial-updates/main armhf linux-firmware all 1.157.3 [32.8 MB]
Fetched 32.8 MB in 11s (2,869 kB/s)                                                                                                          
root@neo:/tmp# ls
linux-firmware_1.157.3_all.deb
 
Then extract rt2870.bin from the deb and put it here: /lib/firmware/rt2870.bin
 
Firmware can also be found here: http://git.kernel.or...lain/rt2870.bin
 
 
 
 

 

 

 

I am having issues using a usb wifi device:

root@neo:~# dmesg | grep rt2
[    0.811567] sw_uart_get_devinfo()1503 - uart2 has no uart_regulator.
[    1.592428] uart2: ttyS2 at MMIO 0x1c28800 (irq = 34) is a SUNXI
[   14.352607] Registered led device: rt2800usb-phy0::radio
[   14.352852] Registered led device: rt2800usb-phy0::assoc
[   14.353151] Registered led device: rt2800usb-phy0::quality
[   14.353452] usbcore: registered new interface driver rt2800usb
[  856.017333] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
[ 1100.839666] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
[ 1357.002345] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
[ 1362.002895] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
[ 1367.000936] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
[ 1372.002929] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
[ 1378.006304] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
[ 1384.001270] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
[ 1388.004151] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
[ 1393.005972] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
[ 1404.001861] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
[ 1408.001750] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
[ 1412.002048] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
[ 1418.001127] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
[ 1423.007031] phy0 -> rt2x00lib_request_firmware: Error - Failed to request Firmware.
root@neo:~# lshw -c network
  *-network:0             
       description: Ethernet interface
       physical id: 9
       logical name: eth0
       serial: 22:43:d7:1f:63:8e
       size: 100Mbit/s
       capacity: 100Mbit/s
       capabilities: ethernet physical tp mii 10bt 10bt-fd 100bt 100bt-fd autonegotiation
       configuration: autonegotiation=on broadcast=yes driver=sunxi_geth driverversion=SUNXI Gbgit driver V1.1 duplex=full ip=192.168.10.29 link=yes multicast=yes port=MII speed=100Mbit/s
  *-network:1 DISABLED
       description: Wireless interface
       physical id: a
       bus info: usb@4:1
       logical name: wlx000f6004d395
       serial: 00:0f:60:04:d3:95
       capabilities: ethernet physical wireless
       configuration: broadcast=yes driver=rt2800usb driverversion=3.4.112-sun8i firmware=N/A link=no multicast=yes wireless=IEEE 802.11bgn
root@neo:~# 

 

It seems as though the firware blob cant be found.

 

 

Any suggestions?

 

 



#150 ChrisK

ChrisK

    Member

  • Senior Members
  • PipPip
  • 21 posts

Posted 23 August 2016 - 02:51 AM

... With such a conductive silicone pad like this you can make thermal pads for 49 H3 (14x14mm). ...

 

It should be noted, however, that such thick SILPADS are quite bad when it comes to thermal resistance. While in some cases using such a thick pad and some metal may be better than nothing at all, it is better to avoid them and find a better solution. Using thermal pads or paste or glue is mainly to bridge air gaps due to imperfect surfaces. That stuff isn't really that thermally conductive at all, it just happens to be better than air. The general rule is that such layers should be as thins possible.

 

Using a small 14x14x10mm heatsink, attached with thermal glue, is very likely to be far better than a 4mm SILPAD and a big piece of metal. Nice, suitable heatsinks for the H3 would be these: http://www.ebay.de/itm/172290970476. Add a very small dab of heat conductive glue, for example http://www.ebay.de/itm/181026615992, then put on the heatsink, align it and finally press it down _really_ good (most of the glue should squeeze out on the sides). Let it sit for a while until it's cured, done.

 

Greetings,

 

Chris


  • tkaiser likes this