OpenWrt Forum Archive

Topic: A5-V11 3G/4G [external antenna][heat issue][power usage]

The content of this topic has been archived on 28 Mar 2018. Unfortunately there are posts – most likely complete pages – missing.

enderffx wrote:

Thank you for your help Tom, will try external soon and report !

Ender

I will wait to here your result.

A5-V11 places away from access point about 20M and 2 walls.
Before modding signal strenght is about 70-72 dBm. After modding signal strenght is about 65-67 dBm.
Before modding Iperf test is about 3-5 Mb/s. After modding iperf jump to 21-24 MBit/s.


To timeless. I have no problem with it but curious the difference between two model. Sorry if you don't understand what I wrote.

Something else:
Does anyone know (measured ?) how much mW this little thing outputs ?

Ender

bertc3p0 wrote:

Even if they are not physical connected they consume power - really.

I noticed the 5 PYH's in this SoC:

mmio 0x101100c0 0x08002000
mmio 0x101100c0 0x08002001
mmio 0x101100c0 0x08002002
mmio 0x101100c0 0x08002003
mmio 0x101100c0 0x08002004

[0x08002000] seems to be for eth0 acording to a post on that forum you noted:

rocksoft on 8devices.com wrote:

I only use a single Ethernet port so I periodiocally turn on the one I want (eth0) in my program
mmio 0x101100c0 0x00002000

But what about its WiFi? I guess this is placed in another register? or is that connected to one of the PYH's?

The power usage of this SoC (Carambola) is:

david on 8devices.com wrote:

Wi-Fi on, all ethernet on: 285mA
Wi-Fi on, all ethernet off: 194mA
Wi-Fi off, all ethernet on: 203mA
Wi-Fi off, all ethernet off: 112mA
Wi-Fi off, one ethernet on: 128mA

that 285mA is exactly my current power usage. so these devices are kinda similar.

Is it save to shutdown all the PHY's except for [0x08002000] which will keep the eth0 and WiFi functional?

(Last edited by Timeless on 20 Nov 2015, 15:29)

tomrev wrote:

I guess I found the right place for external connector. https://wiki.openwrt.org/_media/media/u … v11_03.jpg

I desolder the lead bead above rectangle (this is the ground) at left side near ethernet port which trace to the antenna. Below bead there are 2 copper pads. Desolder this lead will disconnect internal antenna from '+' source

The result modding this way is impressive. Gain with 3 dbi pigtail antenna is increase by 4-5 dBi.

Now I think the signal quality from this antenna mod is not good enough because error rate is quite high. The iperf test is about 22Mbit/s, not too bad to be unacceptable. However, my tablet iperf test at the same position is  better than this router. I placed the router near the AP but the error rate is still high. It shows high "tx excessive retries" and "Invalid misc" from iwconfig command. Here is the stat from iwconfig

$ iwconfig
          Link Quality=52/70  Signal level=-58 dBm
          Rx invalid nwid:0  Rx invalid crypt:0  Rx invalid frag:0
          Tx excessive retries:2845  Invalid misc:6109   Missed beacon:0
$ uptime
06:59:45 up  2:43,  load average: 0.06, 0.04, 0.04

I have tried with Chaos Calmer and trunk firmware but result is similar.


I'm not sure this problem is cause of software or hardware. I wish it could be better than current situation. Do anyone have some ideas to improve this device's signal quality?

(Last edited by tomrev on 22 Nov 2015, 15:35)

Timeless wrote:
bertc3p0 wrote:

Even if they are not physical connected they consume power - really.

I noticed the 5 PYH's in this SoC:

mmio 0x101100c0 0x08002000
mmio 0x101100c0 0x08002001
mmio 0x101100c0 0x08002002
mmio 0x101100c0 0x08002003
mmio 0x101100c0 0x08002004

[0x08002000] seems to be for eth0 acording to a post on that forum you noted:

rocksoft on 8devices.com wrote:

I only use a single Ethernet port so I periodiocally turn on the one I want (eth0) in my program
mmio 0x101100c0 0x00002000

But what about its WiFi? I guess this is placed in another register? or is that connected to one of the PYH's?

The power usage of this SoC (Carambola) is:

david on 8devices.com wrote:

Wi-Fi on, all ethernet on: 285mA
Wi-Fi on, all ethernet off: 194mA
Wi-Fi off, all ethernet on: 203mA
Wi-Fi off, all ethernet off: 112mA
Wi-Fi off, one ethernet on: 128mA

that 285mA is exactly my current power usage. so these devices are kinda similar.

Is it save to shutdown all the PHY's except for [0x08002000] which will keep the eth0 and WiFi functional?

These commands turn off port 1-4.
------------------------
swconfig dev switch0 port 1 set disable 1
swconfig dev switch0 port 2 set disable 1
swconfig dev switch0 port 3 set disable 1
swconfig dev switch0 port 4 set disable 1
swconfig dev switch0 set apply
------------------------

In additional information, A5-V11's CPU is too slow for wireless+NAT process (at least for CC). The wireless driver eat CPU to almost 80% when download at full speed(max at 8-9 MB/s) without NAT. The CPU almost has no power left for do NAT that cap download speed under 3.xx MB/s.

For antenna modding, after changed the rp-sma connector, signal quality improve the signal quality is acceptable now.

>iw dev wlan0 station dump
Station xx:xxxx:xx:xx:xx (on wlan0)
        inactive time:  140 ms
        rx bytes:       30129820
        rx packets:     291783
        tx bytes:       418522172
        tx packets:     312628
        tx retries:     15947
        tx failed:      6
        signal:         -50 dBm
        signal avg:     -51 dBm
        tx bitrate:     72.2 MBit/s MCS 7 short GI
        rx bitrate:     72.2 MBit/s MCS 7 short GI
        expected throughput:    34.84Mbps
        authorized:     yes
        authenticated:  yes
        preamble:       long
        WMM/WME:        yes
        MFP:            no
        TDLS peer:      no

(Last edited by tomrev on 26 Dec 2015, 10:50)

Hi tomrev, so your rp-sma connector was faulty or did you connect differently before ?

Ender

enderffx wrote:

Hi tomrev, so your rp-sma connector was faulty or did you connect differently before ?

Ender

The rp-sma connector is bad in the first attempt. The wireless connection is very stable now. Without NAT, @20HT transfer speed is 5.5-6.xx MB/s, @40HT download speed is max 7.xx which it capped by CPU bound. RT5350 wireless drivers is very CPU demand compare to MR3020 Atheros CPU speed is slightly higher but wireless driver is much faster.

tomrev wrote:

The rp-sma connector is bad in the first attempt. The wireless connection is very stable now. Without NAT, @20HT transfer speed is 5.5-6.xx MB/s, @40HT download speed is max 7.xx which it capped by CPU bound. RT5350 wireless drivers is very CPU demand compare to MR3020 Atheros CPU speed is slightly higher but wireless driver is much faster.

Are you performing your tests with the released Chaos Calmer 15.05 or a different version? Are you using Wi-Fi  only (AP+STA) or the Wi-Fi and Ethernet for your tests? Do you have Wi-Fi encryption enabled during your tests? Thanks.

vernonjvs wrote:
tomrev wrote:

The rp-sma connector is bad in the first attempt. The wireless connection is very stable now. Without NAT, @20HT transfer speed is 5.5-6.xx MB/s, @40HT download speed is max 7.xx which it capped by CPU bound. RT5350 wireless drivers is very CPU demand compare to MR3020 Atheros CPU speed is slightly higher but wireless driver is much faster.

Are you performing your tests with the released Chaos Calmer 15.05 or a different version? Are you using Wi-Fi  only (AP+STA) or the Wi-Fi and Ethernet for your tests? Do you have Wi-Fi encryption enabled during your tests? Thanks.

I tested with CC git released custom build and trunk with more than 10 different builds. Test result is barely different in  maximum performance but some version may have stability problem. Trunk changed ramips drivers from 3.xx to 4.xx which will soft bricked router but fixed with new release.

It used WPA2 WDS client + AP wireless and ethernet wired to internet gateway.

I've tested 15.05.1 and it works better than my previous trunk build (didn't remember which one it was) but still even after disabling all the switch ports (which is also possible in LuCi) the CPU is still hot. I can't hold my finger on it for more than 10sec.

Note: The temperature is ok when I flash the stock bootloader and firmware back.

(Last edited by Timeless on 11 Apr 2016, 19:09)

Yes you right. I have found the same on my router these two points are solder by default.

tomrev wrote:
hackru wrote:

if someone is in same situation as topic starter, check this image:
http://images.mysku.ru/uploads/images/03/39/01/2015/04/14/c26675.jpg
if it's looks like this, you need to solder it so it will look like this:
http://images.mysku.ru/uploads/images/01/87/22/2015/04/15/695f83009f.png
Then you will have good wifi coverage.

I notice the different antenna circuit layout. First pic has no part(inductor?) connect to ground (have 2 pads leave open). I have this version. Second pic has a part. I guess missing this part is a cause of poor wireless signal strength.

Temperature is become normal after flashing the following version with default/reset existing settings:
Powered by LuCI 15.05-187-gfe2a27c Release (git-16.018.33482-3201903) / GoldenOrb_2016-04-01 ( OpenWRT r48430 )

but Wireless frequently disappears while create second/ guest wifi hotspot.   

Timeless wrote:

I've tested 15.05.1 and it works better than my previous trunk build (didn't remember which one it was) but still even after disabling all the switch ports (which is also possible in LuCi) the CPU is still hot. I can't hold my finger on it for more than 10sec.

Note: The temperature is ok when I flash the stock bootloader and firmware back.

tomrev wrote:

These commands turn off port 1-4.
------------------------
swconfig dev switch0 port 1 set disable 1
swconfig dev switch0 port 2 set disable 1
swconfig dev switch0 port 3 set disable 1
swconfig dev switch0 port 4 set disable 1
swconfig dev switch0 set apply
------------------------

Hi Tomrev,
can you (or someone else) explain, how to switch OFF above ports 1-4 permanently. After reboot/poweroff-on they are ON again.
I suppose there must be made some changes in the /etc/config/network file.
I am newbie, please show me what to write there exactly.
Thanks!

May-be it's  useful for someone else, I've found a solution here on OPENWRT Forums
https://forum.openwrt.org/viewtopic.php … 62#p114062

"add the swconfig command to /etc/rc.local to make it permanent for now"

It's working for me, no overheat any-more.

Could someone repost the pictures to fix the antenna, please ? They have vanished. I'm a little bit disappointed about the WiFi performance of my A5-V11s I've got recently ...

The discussion might have continued from here.