Stability problem Tinker Board


NicoD
 Share

3 3

Recommended Posts

Sorry, at the day job, now on lunch.  

 

I never used the Tinker Board kernel, so the DVFS settings were always patched in using the Tinker and Rockchip references.  I will verify the settings 

 

As far as power, there is no debate.  The 1.8 A is the minimum spec for current capability, I have information that says these are tested far beyond that, but again, multiple mate/unmate cycles would reduce that.  So, having dozens implemented with excellent supplies and top quality cables only plugged in a handful of times is not representative of a realistic situation for most users.  I applaud the quality of the deployment in this case, but consider it anecdotal evidence at best that Ohm and Kirchoff can be safely ignored.

 

39 minutes ago, Tido said:

by the way, either you have industrial  or  powering via Micro-USB

 

This is independent of Tinker.  While it's not uncommon to see in the field, "USB" and "industrial" are a bad combination.

Link to post
Share on other sites

Donate and support the project!

@NicoD Render completed with no problems. Took 2h 26m to finish.

 

Maybe the image you downloaded has some kind of problem. I'm going to build a current dev image for you, with all the multimedia stuff implemented, so you can use it in your tests.  I'll try to have it ready in one or two days.

 

Screenshot_2018-04-09_19-41-02.thumb.png.89ef8695c71deb8522574014b33b751b.png

Link to post
Share on other sites

@JMCC Wow, you did the 1080p one. That proves it's stable.
I'll try again with another sdcard to be sure it's not that. I also tried again and again a crash. This time my mouse did move but the desktop was frozen.
Thank you so much, I should have tried another installation first actually.

Let me know when it's ready.
Thanks, sorry for all the troubles :)

Link to post
Share on other sites

@JMCC
I've tried again on another sd, again crashed. Then downloaded Armbian again and tried. No luck. Now I've tried another keyboard, without upgrading, downloading the bmw file, with my power bank,.... All no luck.

So. It's probably my board. I've had these things before. My XU4 doesn't work with any 1080p display. Only with my 7" Display. And it's my second Tinker, the hdmi burned out on the first one.

Do you have any tips which good sbc I should buy next? You guys know a lot better than me what the state of them is.

I thought the potato, but that's almost exactly the same as my Odroid C2. RockPro64 looks interesting. Are banana pi's any good?
Thanks again for the efforts.

Link to post
Share on other sites

2 hours ago, TonyMac32 said:

This is independent of Tinker.

Not in this case, what I want to stress is that just because @freak is using tinker board in an industrial environment  -->  does not make it an "industrial" heavy duty, ruggedized solution.

If he had questioned his actions, he would never have written like he did.

The smallest calculation 4 x USB2.0 Ports = 4 x 500mA = 2A  |  if you just look at it from this perspective tinker board and every other SBC powered with Micro-USB and more than 2 x USB Ports = failed by design.

And not even considered, Ethernet, WiFi/Bluetooth, GPIO or the RK3288.

Link to post
Share on other sites

This is my blender log, one time there was an error when it crashed blender itself and not Armbian, all other times (+20) it was a hard crash, now I`m doing it in 1080p instead of 720p.

Cheers

 

me@tinkerboard:~$ sudo blender
connect failed: No such file or directory
Home directory not accessible: Permission denied
AL lib: (WW) alc_initconfig: Failed to initialize backend "pulse"
Home directory not accessible: Permission denied
AL lib: (EE) UpdateDeviceParams: Failed to set 44100hz, got 8000hz instead
read blend: /home/me/bmw27/bmw27_cpu.blend
skipping driver '100*power', automatic scripts are disabled
skipping driver '-90*brake', automatic scripts are disabled
skipping driver '90*brake', automatic scripts are disabled
skipping driver '-100*power', automatic scripts are disabled
skipping driver '100*power', automatic scripts are disabled
skipping driver '-90*brake', automatic scripts are disabled
skipping driver '90*brake', automatic scripts are disabled
skipping driver '-100*power', automatic scripts are disabled

Link to post
Share on other sites

4 minutes ago, NicoD said:

Is that the output of "sudo armbianmonitor -u"? It doesn't look like an Armbian log at all, looks like libreelec or something alike. Maybe some typo, can you please try again?

 

About board recommendations, I'd wait for RK3399 boards to be released. And don't worry about this Tinker, let's take a deep look before concluding that it's broken.

Link to post
Share on other sites

1 minute ago, JMCC said:

Is that the output of "sudo armbianmonitor -u"? It doesn't look like an Armbian log at all, looks like libreelec or something alike. Maybe some typo, can you please try again?

 

About board recommendations, I'd wait for RK3399 boards to be released. And don't worry about this Tinker, let's take a deep look before concluding that it's broken.

Oops, it's a big K, didn't think it was case sensitive.
http://ix.io/17pK

Link to post
Share on other sites

Just now, JMCC said:

@TonyMac32 He's using 4.4.119, while I'm using 4.4.126 dev. I'm building a 4.4.126 image, so he can test with the same version as mine.

(note: make sure you use this link: http://ix.io/17pK)

 

I didn't upgrade yet there. Now I have and still the same. Tried in 1080p too. Now I'm trying without my fan being powered by the Tinker. Didn't think of that. Looks good for now. It passed 20minutes. I"ll keep you up to date.

Link to post
Share on other sites

7 hours ago, NicoD said:

Again Blender crashed

 

Anyone of the RK guys here ever looked into DRAM timings? Or did you already run memtester?

 

On Allwinner boards we used in the past this to test for DRAM related instabilities: http://linux-sunxi.org/Hardware_Reliability_Tests#Reliability (the idea is to force the GPU consume as much memory bandwidth as possible while memtester is running in parallel -- this unlike 'memtester only' is way better suited to discover memory timing issues that of course can result in freezes/crashes)

Link to post
Share on other sites

17 hours ago, freak said:

 

 

The wrong side of the usb cable?   I'm inline between the tinker and power supply.  That's how you do it.  Here it is with a keyboard, mouse dongle and usb speakers plugged into it.

MVIMG_20180409_101152.jpg

 

Hello @freak,

 

How about try below method to get the actually current system Voltage?

ZQP7lsk.jpg

 

that would more close with the true voltage. (But cannot get the current)

Also if measuring voltage between PSU and Cable, it would miss the Cable’s effect.

(if cable also cause the voltage drop…)

Link to post
Share on other sites

6 hours ago, tkaiser said:

On Allwinner boards we used in the past this to test for DRAM related instabilities: http://linux-sunxi.org/Hardware_Reliability_Tests#Reliability

I didn't know about that test. However, we can't use it,  since it requires specifically Mali 400, and RK3288 has Mali 760.

 

But, if it is about stressing memory while running GPU-demanding tasks,  I've done that with the WebGL Aquarium and other CPU, GPU and memory intensive tasks in parallel. Off the top of my head, I remember memtester, Octane, CPU mining m7m, cryptonight, lyra2rev2, GPU mining lyra2rev2 and skein... Never got a crash, fishes got dizzy of turning around in the aquarium for hours.

Link to post
Share on other sites

16 hours ago, Tido said:

Not in this case, what I want to stress is that just because @freak is using tinker board in an industrial environment  -->  does not make it an "industrial" heavy duty, ruggedized solution.

If he had questioned his actions, he would never have written like he did.

The smallest calculation 4 x USB2.0 Ports = 4 x 500mA = 2A  |  if you just look at it from this perspective tinker board and every other SBC powered with Micro-USB and more than 2 x USB Ports = failed by design.

And not even considered, Ethernet, WiFi/Bluetooth, GPIO or the RK3288.

 

Correct the tinker is not an "industrial" solution.  Just because we use them in a factory to run a web browser does not mean I consider them industrial.  Yes if you are trying to draw 2A out of the tinker you are going to have problems.  If you have such large power requirements you probably shouldn't be looking at sbc's at all.  Use them where they make sense and stop complaining when they don't perform like a desktop computer.

 

Link to post
Share on other sites

I could put together a simplified equivalent cuircuit diagram for the web page to cover this sort of power confusion, since it is universal and based on very basic electrical concepts.

As for the instability concern: what are the cooling options being used? I have a large heatsink, and left off a Tinkerboard-only CPU throttling patch when the GPU was in use. I've never had instability, and given the Tinker does not have the worst cooling in the RK3288 world, assumed the patch was unnecessary (I also have the thermal throttle set to 80 instead 85 C).

I'm not ready to blame this strictly on the hardware just yet, unless it's possible your supply suffers from small "micro-break" events under heavy load. In this situation the device fails to provide stable power for some number of microseconds, potentially starving the board for power. I'll build a 4.4 kernel with the WZYY cpu/gpu patch for OP to evaluate.

Sent from my Pixel using Tapatalk

Link to post
Share on other sites

1 hour ago, TonyMac32 said:

Tinkerboard-only CPU throttling patch when the GPU was in use

In my experience, thermal throttling has been enough. When I was doing all the tests on both CPU and GPU at the time, cpu throttled down when it got too hot,  and there was no further problem. Even more, with a fan and a cool room temperature, throttling was minimal.

 

 

1 hour ago, TonyMac32 said:

I'll build a 4.4 kernel with the WZYY cpu/gpu patch for OP to evaluate.

The OP already got the development 4.4.126 image I built, so feel free to do the patch over the current development kernel instead of release.

Link to post
Share on other sites

Hi all.

I've gotten the 4.4.126 Armbian image from @JMCC

I've done a lot of tests and here are the results.
/*

I tried Blender and it crashed again, tried Kdenlive and it crached.

The video playback in Chromium is a lot better. 720p good picture, almost no lost frames.

But in 1080p there was some stuttering and frame loss. Also the framebuffer emptied itself and started buffering again. That made the video stop, go to a black screen and continiue again(Lubuntu does that too).

It then also crashed while playing a youtube video.

 

I've done a ramtest and that was ok. I then tried Blender again in Lubuntu. That worked fine.

*/

@TonyMac32 I use the stock Tinker heatsink with a 5V fan that's not powered by the Tinker Board. It doesn't go over 55°C. I tried different power supply's, and also with my power bank. Also by powering with the GPIO pins.

Always the same result. Different kinds of crashes, sometimes Blender closes itself, most of the times it's a hard crash and nothing works, sometimes the mouse still moves but the desktop has crashed, and sometimes blender stops working while still open and the desktop still works. Always between 15minues and 20m of rendering.
 

If anyone has any more ideas, please let me know. Thank you all for the help and for all the great work you're doing.

Greetings,

NicoD

 

Link to post
Share on other sites

6 hours ago, TonyMac32 said:

I could put together a simplified equivalent cuircuit diagram for the web page to cover this sort of power confusion, since it is universal and based on very basic electrical concepts.

If we cannot put it on the website, we can put it into the documenation. I like that idea - will show us a draft of what you have in mind?

Link to post
Share on other sites

Guest
This topic is now closed to further replies.
 Share

3 3