Wifi issues on A13-OLINUXINO-WIFI board

[peter_beadle]

Not wanting to be a pain but I bought a board from Olimex with WiFi pre installed, loaded their standard image. I expect it to work. I don't expect to have to waste client time trying to find workarounds.

[diracsbracket]

Exactly !

The WIFI on my 2nd A13-OlinuXino-WIFI board does not work properly either.
This is my 2nd board because the first broke due a mysterious USB hardware problem,
possibly started by the WiFi module ? (Even after unsoldering the module,
the USB subsystem remained utterly broken).

I build my own SD card image according to the instructions on the Olimex site:

http://olimex.wordpress.com/2014/01/15/building-debian-linux-bootable-sd-card-with-hardware-accelerated-video-decoding-and-kernel-3-4-for-a13-olinuxino/

BUT:

1. The WIFI module gets HOT!? I don't remember this happening on my 1st
   board, when it was still working.

2. The connection breaks in the middle of an apt-get install, and is
   the module is useless afterwards, unless I reboot!
   Even ping does not work anymore (hangs), even though ping works correctly
   before I start the apt-get install (of course).

In another post in this forum, this problem was magically "solved" by using
the Release 9 official image from Olimex (according to Michel De Meester),
but that does not say what was wrong in the first place.

So, to test this, I replaced the contents of my SD card with that official image
(again) and tested it. SAME PROBLEM.

Furthermore, as I mentioned, this is my 2nd board.
I had no such WIFI problems then, even though I used the same images.

During boot, with the "official" image Release 9, I still get the following bunch of
error messages:

Code Select Expand

ehci_irq: port change detect
[    2.769029] ehci_irq: port change detect
<3>usb 2-1.4: device descriptor read/all, error -71
[    2.800664] usb 2-1.4: device descriptor read/all, error -71
<3>hub 2-1:1.0: cannot disable port 4 (err = -71)
[    2.814737] hub 2-1:1.0: cannot disable port 4 (err = -71)
ehci_irq: port change detect
[    2.822837] ehci_irq: port change detect
<3>hub 2-1:1.0: cannot reset port 4 (err = -71)
[    2.835062] hub 2-1:1.0: cannot reset port 4 (err = -71)
<3>hub 2-1:1.0: cannot reset port 4 (err = -71)
[    2.848687] hub 2-1:1.0: cannot reset port 4 (err = -71)
<3>hub 2-1:1.0: cannot reset port 4 (err = -71)
[    2.862313] hub 2-1:1.0: cannot reset port 4 (err = -71)
<3>hub 2-1:1.0: cannot reset port 4 (err = -71)
[    2.875935] hub 2-1:1.0: cannot reset port 4 (err = -71)
<3>hub 2-1:1.0: cannot reset port 4 (err = -71)
[    2.889561] hub 2-1:1.0: cannot reset port 4 (err = -71)
<3>hub 2-1:1.0: Cannot enable port 4.  Maybe the USB cable is bad?
[    2.900789] hub 2-1:1.0: Cannot enable port 4.  Maybe the USB cable is bad?
<3>hub 2-1:1.0: cannot disable port 4 (err = -71)
[    2.916233] hub 2-1:1.0: cannot disable port 4 (err = -71)
<3>hub 2-1:1.0: cannot reset port 4 (err = -71)
[    2.930064] hub 2-1:1.0: cannot reset port 4 (err = -71)
<3>hub 2-1:1.0: cannot reset port 4 (err = -71)
[    2.943684] hub 2-1:1.0: cannot reset port 4 (err = -71)
<3>hub 2-1:1.0: cannot reset port 4 (err = -71)
[    2.957312] hub 2-1:1.0: cannot reset port 4 (err = -71)
<3>hub 2-1:1.0: cannot reset port 4 (err = -71)
[    2.970934] hub 2-1:1.0: cannot reset port 4 (err = -71)
<3>hub 2-1:1.0: cannot reset port 4 (err = -71)
[    2.984560] hub 2-1:1.0: cannot reset port 4 (err = -71)
<3>hub 2-1:1.0: Cannot enable port 4.  Maybe the USB cable is bad?
[    2.995783] hub 2-1:1.0: Cannot enable port 4.  Maybe the USB cable is bad?
<3>hub 2-1:1.0: cannot disable port 4 (err = -71)
[    3.011235] hub 2-1:1.0: cannot disable port 4 (err = -71)
<3>hub 2-1:1.0: cannot reset port 4 (err = -71)
[    3.025063] hub 2-1:1.0: cannot reset port 4 (err = -71)
<3>hub 2-1:1.0: cannot reset port 4 (err = -71)
[    3.038684] hub 2-1:1.0: cannot reset port 4 (err = -71)
<3>hub 2-1:1.0: cannot reset port 4 (err = -71)
[    3.052310] hub 2-1:1.0: cannot reset port 4 (err = -71)
<3>hub 2-1:1.0: cannot reset port 4 (err = -71)
[    3.065935] hub 2-1:1.0: cannot reset port 4 (err = -71)
<3>hub 2-1:1.0: cannot reset port 4 (err = -71)
[    3.079560] hub 2-1:1.0: cannot reset port 4 (err = -71)
<3>hub 2-1:1.0: Cannot enable port 4.  Maybe the USB cable is bad?
[    3.090789] hub 2-1:1.0: Cannot enable port 4.  Maybe the USB cable is bad?
<3>hub 2-1:1.0: cannot disable port 4 (err = -71)
[    3.106233] hub 2-1:1.0: cannot disable port 4 (err = -71)
<3>hub 2-1:1.0: unable to enumerate USB device on port 4
[    3.116767] hub 2-1:1.0: unable to enumerate USB device on port 4
<3>hub 2-1:1.0: cannot disable port 4 (err = -71)
[    3.131233] hub 2-1:1.0: cannot disable port 4 (err = -71)
<3>hub 2-1:1.0: hub_port_status failed (err = -71)
[    3.145322] hub 2-1:1.0: hub_port_status failed (err = -71)
<6>usb 2-1: USB disconnect, device number 2
[    3.154843] usb 2-1: USB disconnect, device number 2
<6>usb 2-1: new high-speed USB device number 7 using sw-ehci
[    3.439272] usb 2-1: new high-speed USB device number 7 using sw-ehci
<6>[mmc-msg] mmc 0 detect change, present 1
[    3.592765] [mmc-msg] mmc 0 detect change, present 1
<6>hub 2-1:1.0: USB hub found
[    3.628831] hub 2-1:1.0: USB hub found
<6>hub 2-1:1.0: 4 ports detected
[    3.646854] hub 2-1:1.0: 4 ports detected


And of course, the result from lsusb -t shows that port 4 is the rtl8192cu

Code Select Expand

root@a13-OLinuXino:~# lsusb -t
/:  Bus 03.Port 1: Dev 1, Class=root_hub, Driver=sw-ohci/1p, 12M
/:  Bus 02.Port 1: Dev 1, Class=root_hub, Driver=sw-ehci/1p, 480M
    |__ Port 1: Dev 7, If 0, Class=hub, Driver=hub/4p, 480M
        |__ Port 4: Dev 8, If 0, Class=vend., Driver=rtl8192cu, 480M
/:  Bus 01.Port 1: Dev 1, Class=root_hub, Driver=sw_hcd_host0/1p, 480M

[Michel De Meester]

Hello,

In another post in this forum, this problem was magically "solved" by using
the Release 9 official image from Olimex (according to Michel De Meester),
but that does not say what was wrong in the first place.

<bad English mode>
After updating to the latest debian release I also took the board at home where it seemed to be working fine, didn't try it at work anymore.

I'm not sure it's the new release or other wifi network that solved the problem, I'm even not sure the problem was solved at all, but I was able to do the apt-get update at home.

Because all my spare time goes to taming the saxophone, I didn't have the chance to check everything thoroughly, and probably I won't have time before July.

I'll keep you posted

Michel
</bad English mode>

[diracsbracket]

Hi Michel,
Thanks for the update.
I am certain it is not related to my WiFi router, because as I mentioned,
this is my 2nd board. The first board worked alright, well, until the
USB subsystem gave up that is.

[diracsbracket]

A quick update, to inform you about how the problem with my WiFi downloads seems to
have been resolved.

I re-did everything from the start to create an SD image as , as I did before,
according to:

http://olimex.wordpress.com/2014/01/15/building-debian-linux-bootable-sd-card-with-hardware-accelerated-video-decoding-and-kernel-3-4-for-a13-olinuxino/

Note that the syntax for building u-boot has changed somehow in the latest github Linux-sunxi:

You must now use:

  make CROSS_COMPILE=arm-linux-gnueabihf- A13-OLinuXino_config
  make CROSS_COMPILE=arm-linux-gnueabihf-

For the rest, I did exactly the same as always, except that this time, I checked the WiFi support in the menuconfig settings, after checking out the nice Vinifr tutorial:

http://linuxembeded.blogspot.com.br/2014/04/configuring-wifi-module-rtl8188cu.html

Device Drivers -->

• Network device support -->
• Wireless LAN --> <M> Realtek RTL8192CU/RTL8188CU USB Wireless Network Adapter
  
  My .config settings apparently did not have the above setting somehow ...? So I enabled it as described above.
  
  After completing the whole SD card build process and booting the target, I launched the same apt-get install that resulted in my WiFi interface failure yesterday, and this time, it succeeded.
  Fingers crossed...
  The module still gets quite hot though, even despite the fact that my router is 2m away from the board...
  
  In retrospect, the problems of interrupted downloads may also have been due to SD card corruption and that's maybe why during the download, the system crashed. Really restarting from the beginning (delete all partitions, recreating them and reformatting using dd) may have solved the problem.
  

[Michel De Meester]

Hello,

A quick update, to inform you about how the problem with my WiFi downloads seems to
have been resolved.

I re-did everything from the start to create an SD image as , as I did before,
according to:

As I already mentioned, I don't have any spare time for the moment........

Is there any change you could put the new working SDcard image somewhere so I can download and test it...?   

Tanks in advance
Michel

[vik]

Hi all,

Just a quick note about my findings regarding the issue. I also experienced the same symptoms as many other users described in their posts:

- The WiFi would be unreliable and disconnect after a while;
- The WiFi would disconnect right away when the connection is heavily loaded (e.g. FTPing large files);
- The chip on the miniboard would become very hot.

I experimented for a while. Cut out the miniboard and connected it with wires, so I can have easier access to it. Put a little heatsink on the chip. That made the connection work a little longer, but eventually even the heatsink would become too hot to hold and the connection would be lost. Then I started keeping the heatsink cold by spraying liquid air on it. For as long as I could keep it cold the connection would hold. Then it would go away...

Then I soldered an USB connector to the wires coming from the board and connected an USB dongle which I had bought long ago from eBay (Olimex sells the same under the code MOD-WIFI-R5370). Et voila! My WiFi worked perfectly. And is still working.

So I think that the few lucky us just got a defective WiFi chip on the miniboard. Not very unusual. I have had problems with cheap Chinese WiFi dongles many times before. Generally this is why I buy 2 or 3 every time I need one. From different vendors, of course. When buying from eBay I am paying for 3 devices the same price as for 1 from a local store. True, some are total junk, but still many of them work. So if they are good - then I get some extra. If these are bad - I am not sorry to throw them away...

Anyway, IMHO setting dedicated connector on the board just for the WiFi is a waste of a perfectly good general purpose USB port. It would have been much better if the board had 4 USB ports and any of us could plug-in whatever he pleases.

If Olimex guys bother to read the forum at all, I would note that it would have been a very good idea to sell some of their boards as separate products. For example they do sell the WiFi board, same as soldered to the so called A13-OLinuXino-WIFI (55 euro) under the code MOD-WIFI-RTL8188 for only 7.95. And it has USB connector to it (which could be unsoldered) and then connected to a 45 euro A13-OLinuXino-WIFI. Not that I would call that big savings, but it's just weird. If they could just put 4 male pins where the WiFi is soldered and then sell the WiFi module separately with a female connector - this would have made more sense.

On the same note - if anyone is reading this - I wouldn't mind purchasing some of the LCD adapter boards - the ones that come both with 4.3" and 7" displays. I have a bunch of 7" displays laying around and would be happy to put them to some use. The interface is the same, I already checked.

Well, that's it. Hope I helped some of you. Good luck!

[LubOlimex]

Hey vik,

Thank you for the input on the subject. We are following the forum closely and doing tests constantly. Each design and board has multiple board revisions and when we can we improve the design. However, we have sold a number of A13 boards and most of the people are satisfied with the wifi quality. The few cases that have had problems seem not related to the Olimex design.

It seems a lot of people tend to be surprised by temperatures in electronics. It is quite normal for wifi modules (among other parts) to run at a bit higher temperature.

Regarding the following part:

If Olimex guys bother to read the forum at all, I would note that it would heve been a very good idea to sell some of their boards as separate products. For example they do sell the WiFi board, same as soldered to the so called A13-OLinuXino-WIFI (55 euro) under the code MOD-WIFI-RTL8188 for only 7.95. And it has USB connector to it (which could be unsoldered) and then connected to a 45 euro A13-OLinuXino-WIFI. Not that I would call that big savings, but it's just weird. If they could just put 4 male pins where the WiFi is soldered and then sell the WiFi module separately with a female connector - this would have made more sense.

We have the A13-OLinuXino board without Wifi and with 4 USB ports (it also lacks NAND, but people can solder own one or skip Android at all). It is the base A13 board: https://www.olimex.com/Products/OLinuXino/A13/A13-OLinuXino/open-source-hardware. The A13 micro also might be a good choice for such a case: https://www.olimex.com/Products/OLinuXino/A13/A13-OLinuXino-MICRO/open-source-hardware.

That being said all feedback is welcome and we are happy that more people are participating in the forums and we try to listen to peoples advice and point of view.

Best regards,
Lub/OLIMEX

[vik]

Hi LubOlimex,

Thanks for the fast response. Yes, it seems like I will go with the base board, I got the WiFi board just so I can evaluate the high-end product first. I really don't need the NAND with Android. Nor do I need the on-board WiFi. For connectivity we can use a WiFi dongle or an USB to Ethernet adaptor. Both work fine.

I want to clarify that in my case the overheating was the real problem. I am not surprised at all at the heat generated by the WiFi modules as I have used quite a few of them. But the one that came with the board clearly had a faulty chip. I am not blaming Olimex for that. Shit happens. As I said, the moment I put a tested and working USB WiFi dongle it all worked perfectly. So for all those who, like me, have experienced a problem with the WiFi - my advice is - just replace the miniboard. Use a standard off-the-shelf USB dongle, or if you really want to have it onboard - buy an MOD-WIFI-RTL8188 and solder it back.

Generally I can only congratulate Olimex for the good job. I got my order delivered in 4 days overseas and the packaging and the overall quality were very good. The prices are also very competitive, as I have a wide range of similar boards and I am able to compare. So I consider the problem with the WiFi to be very minor.

[Andrew Smalley]

I will also confirm debian and 8192cu the chip gets VERY VERY hot and drops connection! I cant even install a single package from the internet without problems.

the only way i get it back is by a reboot into Android and e2fsck /dev/block/mmcblk0p2 and then reboot.

if i could find a metal heatsink small enough i would mount it on the chip to solve these problems,

I was wondering how USB works at 3.3v?

Ive had my A13 for around 28 hours now and like it apart from this really annoying problem...

also I have noted my ram is not reporting 512mb but

MemTotal:         135384 kB


I have semi resolved my heat issues with the wi-fi as i have used an alan-key and masking tape wrapped around the metal to prevent contact with other components, I think I will remove the wifi module and add a right angle usb socket and put other wifi dongle on the A13


Regards

[2001cpx]

i have same issue with wifi onbord module. run very hot and lost connection .  Power level too high?

[nickola]

I am also experiencing the same issues. I have just purchased A13-OLINUXINO-WIFI with SD card containing preinstalled Debian. The wifi module becomes very hot and stops working as long as you try transfer even a bit. No mater what you do over wifi, besides simple ssh browsing, triggers the problem. I tried the following:

* Cloning a git branch - wifi died
* Installing vim with apt-get - wifi died
* Connected through VNC, had a few clicks - wifi died
* Started the browser, googled random thing - wifi died

Olimex, in fact this is very disappointing. On your website you advertise this board as "industrial grade" and even mention temperature amplitudes that it will work into. In your presentation on Saturday you made me choose you over other manufacturers with promises for an easy start - well, facing basic issues such as internet connectivity is a great hassle and definitely far from the "easy start".

[vellamy]

Hi,

I'm trying to solve the problem of the Wifi connection with the A13-Olinuxino board.

The info on: http://linux-sunxi.org/Wifi#8192cu is useless. According to the web:
To fix the problem create a file /etc/modprobe.d/8192cu.conf with the following contents:

options 8192cu rtw_power_mgnt=0 rtw_enusbss=0
However, this configuration is already set Olinuxino.

Then, I tried to change the Debian image. I used your link:
https://drive.google.com/file/d/0B-bAEPML8fwlRHN4UzhrWVhzQms/edit?usp=sharing.

I downloaded the file. Then, unzipped the file with 7Zip. After this, I wrote the image on the SD memory.
Nevertheless, The problem persists.
On Android, the board is unable to connect with my wifi network.

I've probed Olinuxino with four different wifi networks, but the same problem.

I'm sure that I'm not the only one with wifi connection problems.So,What is the workaround for this problem?

[Michel De Meester]

Hello All,

It's been a while since I spent some time over here.

I have some spare time and am going to restart my A13 project, is there already a solution for the wifi module disconnecting from the network al the time?
Thanks in advance.

Michel

[asmecher]

One additional data point. I've recently received an A13 direct ordered from Olimex, put the Debian image downloaded directly from Olimex onto a MicroSD, and I'm experiencing the same problems: the wifi chip heats up very hot, and I can transfer very litte data (a few hundred KB using apt-get) before the connection stops working. Removing and re-adding the 8192cu module gets it working, briefly, again.

I like the board so far, but an untrustworthy wireless card is a bit of a black eye.