BurninTurtles
DD-WRT Novice
Joined: 07 Jul 2018
Posts: 3
|
 Posted: Mon Jul 23, 2018 1:04 Post subject: TP-Link Archer C7v2 build log successful. Want opinions. |
 |
Hello all! this is my first real post so please bear with me. I am a long time user of DD-WRT and am not entirely ignorant with computers in general. I have a setup that I believe I have set up properly for my environment that I would like to share and get opinions on.
I purchased 3 TP-Link Archer c7v2's on Ebay recently and discovered that I needed to recover them via serial. I purchased the serial adapter and followed the directions in this thread ( https://forum.dd-wrt.com/phpBB2/viewtopic.php?p=1070103 ) and all went perfectly. I chose to flash build "DD-WRT v3.0-r35452 std (03/20/1 " as it was the most stable and up to date build to work on my other (decommissioned) linksys e1200v2, dlink dir601a1, and netgear wnr2000v2 routers. I flashed this image multiple times via many different methods (tftp, serial, webgui, etc) to be safe as well as throughout testing, a few times more, primarily through the webgui. I basically followed peacock thread procedures to be extra safe despite being on a different chipset.
Now for the setup. here are my router locations noted by rX:
(note all routers are in the same vertical plane, the ceiling of a basement or the ground of a garage.
_________________|-----------------------50ft------------------------------|
_________________|_________________________________________|
_________________|_______________________brick |__________r1_|
_________________|_____________________|chimney|____________|
________________40ft- - - - - - - -8in steel I-Beam - - - - - - - - - - - - -|
_________________|_________________________________________|
_________________|_________________________________________|
_________________|r2_______________________________________|
_________________|brick----------------------------------------------------|
_________________15ft space
--------------------brick|
_______________20ft|
__________________|------------------------yard
_______________r3_|
__________________|
----------30ft-----------|
_________________30ft space
__________________property line_______________________________
.
Network Layout.
Arris modem -> linksys ea6350v3 serving my moms network.
ea6350(with r1 plugged into LAN1 and prioritized in Stock firmware QOS)-> ArcherC7v2. All of this is in the immediate vicinity of r1.
ArcherC7v2 (My r1)
lan address 192.168.1.1
gateway, QOS set to 85mbps up(max average is 95mbps) and 20mbps down(max average is 25mbps), dhcp server pool is .101-.130.
2.4 (ath0) is set AP, g/n mixed, channel 11+7, ht40, lower, 24dbm tx, beacon 100, dtim 2, ack 450, WPA2 AES
5.0 (ath1) set to WDS AP, n/ac mixed, channel 161, vht80, LL, 30dbm tx, beacon 200, dtim 1, ack 450, WPA2 AES
ArcherC7v2 (r2)
lan address 192.168.1.2
router, no QOS, dhcp forward to r1
2.4 (ath0) is set to AP, g/n mixed, channel 6+2, ht40, lower, 15dbm tx, beacon 100, dtim 2, ack 450, WPA2 AES
5.0 (ath1) is set to WDS Station, n/ac mixed, 30dbm tx, beacon 200, dtim 1, ack 450, WPA2 AES
5.0 (ath1.1) WDS AP, dtim 1, WPA2 AES
ArcherC7v2 (r3)
lan address 192.168.1.3
router, no QOS, dhcp forward to r1
2.4 (ath0) is set to AP, g/n mixed, channel 11+7, ht40, lower, 15dbm tx, beacon 100, dtim 2, ack 450, WPA2 AES
5.0 (ath1) is set to WDS Station, n/ac mixed, 30dbm tx, beacon 200, dtim 1, ack 450, WPA2 AES
Results:
r1 has great coverage for its 2.4 zone. 5.0 has an average SNR of 38 to the WDS station r2 with a constant link speed of 650M or more. speedtests tests give me 50mbps down, 27 up.
r2 again has perfect 2.4 coverage for its zone. speedtests average on 35 down, 25 up. 5.0 is similar. minimum of 650m link to r1. averaging a 20snr to r3 with a linkspeed of 200-300m to r3.
r3 also has perfect 2.4 coverage for its zone. speedtests give me 25 down, 15 up. the 5.0 ghz link maintains a linkspeed of 200-300m.
If you made it this far, I appreciate your time first and foremost, but you will have also noticed that I'm using 5.0 ghz as a wds backbone to serve all of my clients in 2.4 mode. even in my noisy suburbia environment, where there virtually no free 2.4 channels, I manage to have multiple 1080p streams at either end of this chain running hiccup free simultaneously. I (believe) I did this by tweaking a very few settings to maximize data over the air as well as choosing the least populated channels, YMMY.
My purpose in this post is to document my experience and ask a few questions. so here are some questions.
How could i clean this post up better?
Is there anything that a veteran Guru would try differently?
Im happy to test those ideas, Procedure or link?
Is there any information that you believe I missed or may benefit from? Link?
to answer one question I believe people may have: I changed the beacon and dtim values on my 5.0 antennas for the purposes of maximizing time that data is in the air ad not beacon frames. the 5.0 environment in my area is virtually empty. This resulted in an extra 5-10mbps bump. tending towards the lower, but a boost nonetheless.
This was all in attempt to get better coverage in an environment not conducive to wireless signals or router placement, as well as to increase range by/for at least one hop, the second was just because i found a good deal xD.
Over all I'm quite happy with my setup and performance. being able to have 2 1080p streams 3 or 4 router hops away from the modem is satisfying and win in my book however I am open to ideas, texts, and respectful criticisms. I am very sarcasm friendly as well, so feel free to be a pr*ck  |
|
BurninTurtles
DD-WRT Novice
Joined: 07 Jul 2018
Posts: 3
|
| Posted: Wed Jul 25, 2018 2:46 Post subject: Re: TP-Link Archer C7v2 build log successful. Want opinions. |
|
| ramblin wrote: | | Maybe give up the QOS. Probably don`t need it and it is just sucking up valuable resources. |
Usually I would agree based on all past tests however I believe that this time it is actually doing it's job.
My typical speed test routine/average consists of a combination of dslreports and telnetting into a router and running some of these downloads ( https://linhost.info/2013/10/download-test-files/ ) while monitoring the bandwidth from one hop upstream. I noticed that my bandwidth is definitely leveled off and more robust against spikes or drops with QOS meanwhile my cpu load never gets above .30; even with either a 1080p stream or a BitTorrent client running at ~3MB/s.
The real benefit is the bandwidth leveling as far as I can tell. Dslreports used to give me a C for bufferbloat but now it's consistently an A. Im not much of a gamer, much less an online gamer, but I would like the network to be receptive to a gamer anyway. I mainly just consume media so as long as my bandwidth is above 30 I couldn't care less. Any extra I regard as a failsafe for the further-down-the-chain devices. It's observed, and common sense tells me rightfully so, that the further away from the wan connection down the chain I go, the buffer bloat does increase, but as noted in the first post, the speed test is fairly consistent and better than my expectations.
The only way I theorize that this works so well is because I'm using 5ghz vht as the backbone. I, unwittingly at first, planned
my positioning to always have double my wan speed as a minimum linkspeed for the wireless connections, especially the wds ones, and that seems to have allowed magic to happen lol.
As an aside, some events have transpired that has forced a change of the physical layout but somehow as a result of this, I will be getting my own internet hookup, so the documented positions will be restored and more tests or reports will follow for this thread. Some goals I have both for practicality and learnings sake are:
1. Getting ipv6 global addresses on lan with a good firewall.
2. An adblocker, ideally based on ublock origin or it's associated block lists.
3. Finding some fun uses for jffs.
4. Rather unassociated with dd wrt, but setting up a 2tb(raid 1:1 mirroring) linux server with a transmission client, Kodi client/library server, and a few web servers. (The only thing I used windows for in the last 5 years that wasn't tech support for family was the recovery of these Archer c7s lol)
5. A VPN server for remote access.
Any personal setups, accounts, stories, or informative links are welcome. For those, please assume I know a thing or two about Google, this wiki, and this forum. |
|
