OpenWrt Forum Archive

Topic: davidc502 1900ac 3200acm builds

The content of this topic has been archived between 26 Feb 2018 and 7 May 2018. Unfortunately there are posts – most likely complete pages – missing.

habakuk wrote:

@starcms
@davidc502

Hi!
I am running a WRT1900acsV2 - Rev: A00 - EU on reg code "FR" 20dbm on LEDE 17.01.4-r3560 shelby - mvebu.

Everthing is working fine sofar. We are running  a rather complex config with couple of package as Travelmate with AP and client as well as UMTS interface. completely stable.

Having trouble getting enough dBm (max. 20).
Tried reghack2 - went through - no changes as expected.

Is there a way to overcome the powertables and regtables in EPROM with/without sautering (reading/rewriting)?
(we are in search of a stable, powerful wifi solution for our cars, which has to be replicable)

Can you please help us?

wkr
habakuk

You cannot edit the powertables on the 1200ACv2, 1900ACSv2, or 3200ACM.  Full stop.

reghack2 doesn't even work on the other WRT models that don't have the power-tables in EEPROM anymore because of a change in lede/openwrt.  It would need to be recompiled with a few tiny modifications.

But you on your ACSv2, you cannot change the transmit power level or the country/region code.  It may appear to change in Luci, but it isn't actually changing.

Edit:  And I know you already knew this.  And you aren't even running @david's build.  So why are you posting in this thread?

(Last edited by starcms on 11 Apr 2018, 11:10)

davidc502 wrote:
rs.se wrote:

Just a friendly warning, today's code/snapshot for WRT3200ACM has corrupted my device leaving the power light blinking but wifi/ethernet ports working with no way to access the device.

I never did buy a JTag cable, looks like that is the only fix for this type of problem.

This  build is from the daily snapshots?  I'm guessing you have tried doing the power on/off method to recover the other partition?

By the way, if forced to use the serial method, it is a bit time consuming the first time, but over time gets much faster.

I have tried the on/off method many times at different timungs, also have tried flashing new image twice in failsafe mode.

The daily build for the 10th caused the issue, i am thinking of either recover via serial cable or linksys RMA if warranty is still good.

rs.se wrote:
davidc502 wrote:
rs.se wrote:

Just a friendly warning, today's code/snapshot for WRT3200ACM has corrupted my device leaving the power light blinking but wifi/ethernet ports working with no way to access the device.

I never did buy a JTag cable, looks like that is the only fix for this type of problem.

This  build is from the daily snapshots?  I'm guessing you have tried doing the power on/off method to recover the other partition?

By the way, if forced to use the serial method, it is a bit time consuming the first time, but over time gets much faster.

I have tried the on/off method many times at different timungs, also have tried flashing new image twice in failsafe mode.

The daily build for the 10th caused the issue, i am thinking of either recover via serial cable or linksys RMA if warranty is still good.

not using @david's build...why are you in this thread?

rs.se wrote:
davidc502 wrote:
rs.se wrote:

Just a friendly warning, today's code/snapshot for WRT3200ACM has corrupted my device leaving the power light blinking but wifi/ethernet ports working with no way to access the device.

I never did buy a JTag cable, looks like that is the only fix for this type of problem.

This  build is from the daily snapshots?  I'm guessing you have tried doing the power on/off method to recover the other partition?

By the way, if forced to use the serial method, it is a bit time consuming the first time, but over time gets much faster.

I have tried the on/off method many times at different timungs, also have tried flashing new image twice in failsafe mode.

The daily build for the 10th caused the issue, i am thinking of either recover via serial cable or linksys RMA if warranty is still good.

I recommend buying the cable and trying to recover because it is good experience. The 3200acm comes apart easily and the port is easily accessible, unlike the 1900ac smile....

starcms wrote:
habakuk wrote:

@starcms
@davidc502

Hi!
I am running a WRT1900acsV2 - Rev: A00 - EU on reg code "FR" 20dbm on LEDE 17.01.4-r3560 shelby - mvebu.

Everthing is working fine sofar. We are running  a rather complex config with couple of package as Travelmate with AP and client as well as UMTS interface. completely stable.

Having trouble getting enough dBm (max. 20).
Tried reghack2 - went through - no changes as expected.

Is there a way to overcome the powertables and regtables in EPROM with/without sautering (reading/rewriting)?
(we are in search of a stable, powerful wifi solution for our cars, which has to be replicable)

Can you please help us?

wkr
habakuk

You cannot edit the powertables on the 1200ACv2, 1900ACSv2, or 3200ACM.  Full stop.

reghack2 doesn't even work on the other WRT models that don't have the power-tables in EEPROM anymore because of a change in lede/openwrt.  It would need to be recompiled with a few tiny modifications.

But you on your ACSv2, you cannot change the transmit power level or the country/region code.  It may appear to change in Luci, but it isn't actually changing.

Edit:  And I know you already knew this.  And you aren't even running @david's build.  So why are you posting in this thread?

Thanks for your answer:
I am in this thread because here I finally found somewhat reliable information about the wrt1x00acxVx - confusion, which is most likely somewhat intended by Linksys to disguise their changes in HW (?)
After an endless search and finding of contradictive information I landed here.
So thanks once more for you answer.
I am convinced that you guys tried everything to overcome that burden in 1900acV2 an 3200 on a firmware basis and I appreciate this.
What about the approach to make an 1900acV2 out of an 1900acsV2, does it have a chance at all?

ASSUMPTION:
We believe that the mainboards are exactly the same. Unfortunately, I do not own a 1900acV2 board by now. (working on it...)
If one puts himself in the shoes of linksys, from an economic point of view, I would just flash the EEPROM differently right before distribution, to kill to birds with one stone: meeting FCCS reg codes and differentiate acV2 from acsV2 in a minute. Ifyou make more severe changes, you most like won't be able to flood the market with this devices around 150 bucks, huh!
This theory unfortunately opens up another box of questions: Which tool are they using via which interface (USB-eSATA?) to write to the EEPROM initially.

Well, whats your optinion on this? Is it worth a try?
Do you have any additional detailed information about the read only EEPROM?

(Last edited by habakuk on 11 Apr 2018, 13:51)

davidc502 wrote:
rs.se wrote:
davidc502 wrote:

This  build is from the daily snapshots?  I'm guessing you have tried doing the power on/off method to recover the other partition?

By the way, if forced to use the serial method, it is a bit time consuming the first time, but over time gets much faster.

I have tried the on/off method many times at different timungs, also have tried flashing new image twice in failsafe mode.

The daily build for the 10th caused the issue, i am thinking of either recover via serial cable or linksys RMA if warranty is still good.

I recommend buying the cable and trying to recover because it is good experience. The 3200acm comes apart easily and the port is easily accessible, unlike the 1900ac smile....

Thanks David, after 6hours the routers LED's have returned to normal but had to leave for work before trying Web UI. I suggest against trying latest code for a bit. I most definitely need to buy a cable after this adventure lol. We should connect sometime outside this forum to chat about your setup.


I have been in this thread on and off for sometime, my post was more so a warning to David and anyone else here who uses David's builds and tests their own.

I use David's builds, daily snapshots, and my own builds based on David's config to get bleed edge changes, sadly yesterday was absolutely untested code.

Probably a newb issue but am running the latest david release and have one device that wont connect on my 3200acm. It's a smart plug that only works on 2.4band. It tries to connect but never gets there. I have an old router connected lan to lan with the 3200 and if I turn on 2.4 on that it will connect fine. So it appears to be something about the 2.4 itself. I've tried wpa2, wpa2 mixed. N mode, legacy and all combos I can think of. Any suggestions? Every thing else that can connect by 2.4 seems ok.

Hi guys
Firstly, I’m very sorry if my question is noob-like.
I have a Linksys WRT3200ACM with LEDE SNAPSHOT r6565-fd588dbf6b / LuCI Master (git-18.088.49307-90ed423).
How can I install PPTP or IPSec server on it?
I tryed h_ttps://wiki.openwrt.org/doc/howto/vpn.server.pptpd - useless, no connect to router both inside or outside my LAN. When I start pptpd /etc/init.d/pptpd start I see an error “sh: out of range”. I googled the same case h_ttps://forum.lede-project.org/t/default-config-file-for-pptpd-lacks-logwtmp-option/4795
Is it mean poptop package is buggy?

2nd question - why this repo doesn't have luci-app-ipsec-vpnd.ipk package? smile Is it possible to use "built-in" Windows VPN connections for connection to this router?



If it is important - my ISP link uses PPPOE.

Great thanks for your advices!

P.S. Platform doesn't allow me to post URLs, just delete _ in h_ttps

makedir wrote:

I was able to solve the permanent dnsmasq restarts btw by deactivating ipv6.

@makedir

I have the same issue and thought I had disabled/ deactivated IPV6 but I must have missed something.

Can you give me your steps for disabling IPV6?

Sorry guys if i'm bothering you...
Is this the normal way a working adblock looks like in luci?
http://i199.photobucket.com/albums/aa309/komodo1980/th_adblock.jpg

Thank you very mutch

(Last edited by komodo_1 on 11 Apr 2018, 22:18)

Wasn't happy running at 80°C so I installed a blower fan to suck hot air out of my WRT1900ACS, here are the new temps;

root@OpenWrt:~# sensors
tmp421-i2c-0-4c
Adapter: mv64xxx_i2c adapter
temp1:        +35.8°C 
temp2:        +36.8°C 

armada_thermal-virtual-0
Adapter: Virtual device
temp1:        +63.2°C

This is the hottest things get now:)

Anyone know if there are any safety protocols in place be it hardware or software that auto power-off the router if temps hit 90°C for example? And if there is, can those settings be tweaked?

(Last edited by Cantenna on 11 Apr 2018, 22:05)

hmm...  i've noticed the wrt1900ac fan must be kicking on every 5 mins.  log shows:

cron.info crond[2304]: USER root pid 30148 cmd /sbin/fan_ctrl.sh

temperature seems acceptable.

 Lede SNAPSHOT, r6565-fd588dbf6b
 -----------------------------------------------------
root@OpenWrt:~# sensors
armada_thermal-virtual-0
Adapter: Virtual device
temp1:        +68.2°C

tmp421-i2c-0-4c
Adapter: mv64xxx_i2c adapter
temp1:        +52.9°C
temp2:        +53.9°C

root@OpenWrt:~#
wrtboy wrote:

hmm...  i've noticed the wrt1900ac fan must be kicking on every 5 mins.  log shows:

cron.info crond[2304]: USER root pid 30148 cmd /sbin/fan_ctrl.sh

temperature seems acceptable.

 Lede SNAPSHOT, r6565-fd588dbf6b
 -----------------------------------------------------
root@OpenWrt:~# sensors
armada_thermal-virtual-0
Adapter: Virtual device
temp1:        +68.2°C

tmp421-i2c-0-4c
Adapter: mv64xxx_i2c adapter
temp1:        +52.9°C
temp2:        +53.9°C

root@OpenWrt:~#

The WRT1900 certainly has more longevity than the ACS because of its fan, which is why i'm adding one:)

(Last edited by Cantenna on 12 Apr 2018, 12:28)

bigdave240 wrote:

Probably a newb issue but am running the latest david release and have one device that wont connect on my 3200acm. It's a smart plug that only works on 2.4band. It tries to connect but never gets there. I have an old router connected lan to lan with the 3200 and if I turn on 2.4 on that it will connect fine. So it appears to be something about the 2.4 itself. I've tried wpa2, wpa2 mixed. N mode, legacy and all combos I can think of. Any suggestions? Every thing else that can connect by 2.4 seems ok.

@bigdave240

For the smartplug, have it completely forget any part of the configuration for 5Ghz, and reboot it. Then try to connect again, and enter the pre-shared-key (password), and see if that doesn't fix the problem.

Cantenna wrote:

Wasn't happy running at 80°C so I installed a blower fan to suck hot air out of my WRT1900ACS, here are the new temps;

root@OpenWrt:~# sensors
tmp421-i2c-0-4c
Adapter: mv64xxx_i2c adapter
temp1:        +35.8°C 
temp2:        +36.8°C 

armada_thermal-virtual-0
Adapter: Virtual device
temp1:        +63.2°C

This is the hottest things get now:)

Anyone know if there are any safety protocols in place be it hardware or software that auto power-off the router if temps hit 90°C for example? And if there is, can those settings be tweaked?

Good news!! btu please if possible send a link for this cooler, a long time i'm want one solution for my ACS and ACM.

Thanks

fcs001fcs wrote:
makedir wrote:

I was able to solve the permanent dnsmasq restarts btw by deactivating ipv6.

@makedir

I have the same issue and thought I had disabled/ deactivated IPV6 but I must have missed something.

Can you give me your steps for disabling IPV6?

Yepp. Totally same issue. Why? Because the Lede morons havent still put a simple one button solution into OpenWrt to deactivate Ipv6. You have to hack yourself through it to fully deactivate it, do ALL these steps:

1. put net.ipv6.conf.all.disable_ipv6=1 into /etc/sysctl.conf
2. /etc/init.d/odhcpd disable
3. remove the network IPv6 ULA-Prefix (important), set it to blank (luci: network->interfaces https://i.imgur.com/wHS38sB.png)
4. deselect 'bring up at boot' for wan6 ( luci: https://i.imgur.com/yv3yEms.png )
5. select disabled/disabled/disabledfor ipv6 dhcp server for lan ( luci: https://i.imgur.com/XjPfA5c.png )

nasty, isnt it? THAT, and only THAT, should deactivate ipv6 on this build. the constant dnsmasq restart issues should now be fixed. the wan6 virtual interface still comes up afer this which also seems like a bug to me, but it doesnt matter.

@komodo_1 yes, seems so. just test it? look into /tmp/adb_list.overall and open one of the blocked entries on one of your clients in a browser which hasnt any adblock addon active.

(Last edited by makedir on 12 Apr 2018, 01:30)

Website (davidc502sis.dynamic-dns.net) Maintenance is scheduled tonight at 10pm CDT to patch for Spectre vulnerability. During this time the site will have interruptions and intermittent availability.

(Last edited by davidc502 on 12 Apr 2018, 01:36)

bigdave240 wrote:

Probably a newb issue but am running the latest david release and have one device that wont connect on my 3200acm. It's a smart plug that only works on 2.4band. It tries to connect but never gets there. I have an old router connected lan to lan with the 3200 and if I turn on 2.4 on that it will connect fine. So it appears to be something about the 2.4 itself. I've tried wpa2, wpa2 mixed. N mode, legacy and all combos I can think of. Any suggestions? Every thing else that can connect by 2.4 seems ok.

This is a bug in the mwlwifi driver that was closed for some reason see: (I can't post links) hxxps://github.com/kaloz/mwlwifi/issues/170

You can temporarily work around it by disabling WMM and N mode on the 2.4 interface.

It's a shame that there are still this many bugs with the wifi driver. I have connection issues with about half of my clients given that most are either android or "smart" iot devices, my network has a lot of bandaids to work around everything but at the end of the day I enjoy having control over where my information goes over the internet with wireguard and quad9 and also having flexible router level adblocking and firewall control

bigdave240 wrote:

Probably a newb issue but am running the latest david release and have one device that wont connect on my 3200acm. It's a smart plug that only works on 2.4band. It tries to connect but never gets there. I have an old router connected lan to lan with the 3200 and if I turn on 2.4 on that it will connect fine. So it appears to be something about the 2.4 itself. I've tried wpa2, wpa2 mixed. N mode, legacy and all combos I can think of. Any suggestions? Every thing else that can connect by 2.4 seems ok.

Here is my experience with the 3200acm. In order to connect with the smart plug on 2.4G, I have to just use legacy mode and uncheck the WMM (wireless multimedia). However with this configuration I can only get a few hundred KB download speed with other devices.
In DD-WRT, you can enable the third radio (BT/WiFi combo) and there is no issue for smart plug to connect to it while the main 8964 2.4G radio can be configured to be useful.

Cantenna wrote:

is there already a process in affect to auto power-off router if temperature becomes critical?

Read my message about it some pages ago, i have overheat problem too (live in tropics), so i patch the file for this.
For now, stable even in +40 outside temp.

Thank you for the ipv6 disabling steps, i forgotten the 1th and 2nd, so i just applyed them
1. put net.ipv6.conf.all.disable_ipv6=1 into /etc/sysctl.conf  (then run "sysctl -p")
2. /etc/init.d/odhcpd disable

makedir wrote:

@komodo_1 yes, seems so. just test it? look into /tmp/adb_list.overall and open one of the blocked entries on one of your clients in a browser which hasnt any adblock addon active.

On the adblock side, also to you luci or the "/etc/init.d/adblock status" command reports "Enabled" and not "Running"?

EDIT:
Using sshtunneling I just made the test on some "blocked url" from /tmp/adb_list.overall (i.e. http://addfreecounter.com/ or crownoption.com) but sadly i was able to surf them. sad 
Going to start a system in that network to check directly from there. Will report asap.

EDIT2: Same, tested from lan and it also browse the "blocked sites" from /tmp/adb_list.overall sad((

Planning to flash the firmware again, because i can't find a way to resolve the adblock issue and this make me mad :@

(Last edited by komodo_1 on 12 Apr 2018, 11:14)

gu3d3s wrote:

Good news!! btu please if possible send a link for this cooler, a long time i'm want one solution for my ACS and ACM.

Thanks


smile

1) Buy this (I plugged it into the esata/usb port on the back of the router)
https://www.ebay.com.au/itm/50mm-USB-Bl … SwTf9ZUhcV

This will suck hot air out and thereby keep dust out after you tape it up!

2) Buy Gloria clear tape to cover up all areas on top except for where the fan will be mounted.

3) Buy a couple of M3 blots and nuts to mount to unit.

4) Dissemble the router entirely; you'll definitely need to buy these to dissemble with minimal damage...
https://www.ebay.com.au/itm/10PCS-Black … SwAxNaxHS1

4) Create an overheat script to protect your router (because you have effectively handicapped passive cooling you'll need your router to auto power off at your per-defined threshold (I selected 80° C)

vi /etc/overheat.sh

#!/bin/sh

CPU_TEMP=`cut -c1-2 /sys/class/hwmon/hwmon1/temp1_input`

CPU_HIGH=80

if [ "$CPU_TEMP" -ge "$CPU_HIGH" ];then
        poweroff
fi


Then, add to luci gui > system > startup > local startup before exit 0
sleep 120 ; sh /etc/overheat.sh

5) *Bonus* use a tie to re-enforce cable strength of fan:)

http://i64.tinypic.com/k3tbps.jpg
http://i65.tinypic.com/dmxkz7.jpg
http://i68.tinypic.com/110hzci.png

(Last edited by Cantenna on 12 Apr 2018, 12:31)

komodo_1 wrote:

On the adblock side, also to you luci or the "/etc/init.d/adblock status" command reports "Enabled" and not "Running"?

You're are trying to use the new dnscrypt-proxy v2 with adblock? Did you reference the blocklist within your dnscrypt-proxy configuration? I don't know the syntax in v2 but in v1 you have to specify something like that:

list blacklist 'domains:/tmp/adb_list.overall'

Hope this helps.

Thank you very much didbot
Yes, I'm using dnscrypt v2 (2.0.9)
As you pointed me out,
I forgotten to reference the blacklist file on the dnscypt(v2) .toml config file.

Here's the v.2 syntax:
   blacklist_file = '/tmp/adb_list.overall'     

My first mistake was to believe that adblock's process had to stay running all time.
But as you all confirmed it just stay on Enabled, tha's perfect. smile

Thank you all for the help!!

dibdot wrote:

You're are trying to use the new dnscrypt-proxy v2 with adblock? Did you reference the blocklist within your dnscrypt-proxy configuration? I don't know the syntax in v2 but in v1 you have to specify something like that:

list blacklist 'domains:/tmp/adb_list.overall'

Hope this helps.

Cantenna wrote:

smile

1) Buy this (I plugged it into the esata/usb port on the back of the router)
https://www.ebay.com.au/itm/50mm-USB-Bl … SwTf9ZUhcV

This will suck hot air out and thereby keep dust out after you tape it up!

2) Buy Gloria clear tape to cover up all areas on top except for where the fan will be mounted.

3) Buy a couple of M3 blots and nuts to mount to unit.

4) Dissemble the router entirely; you'll definitely need to buy these to dissemble with minimal damage...
https://www.ebay.com.au/itm/10PCS-Black … SwAxNaxHS1

4) Create an overheat script to protect your router (because you have effectively handicapped passive cooling you'll need your router to auto power off at your per-defined threshold (I selected 80° C)

vi /etc/overheat.sh

#!/bin/sh

CPU_TEMP=`cut -c1-2 /sys/class/hwmon/hwmon1/temp1_input`

CPU_HIGH=80

if [ "$CPU_TEMP" -ge "$CPU_HIGH" ];then
        poweroff
fi


Then, add to luci gui > system > startup > local startup before exit 0
sleep 120 ; sh /etc/overheat.sh

5) *Bonus* use a tie to re-enforce cable strength of fan:)

http://i64.tinypic.com/k3tbps.jpg
http://i65.tinypic.com/dmxkz7.jpg
http://i68.tinypic.com/110hzci.png

thanks for posting all the information and steps that should be used, will provide for my two routers, because I live in tropiccal area and I already had peaks of 92C and this makes me very nervous: D

The photos as always are very useful, the only observation is that you lose one of the usb ports, but I thought putting a powered hub should solve the problem, since I use a PenDrive for / opt (yamon, colletd files, and saved LOG's files ), and one HD for my time machine and DLNa files.