freifunk-gluon / gluon

a modular framework for creating OpenWrt-based firmwares for wireless mesh nodes
https://gluon.readthedocs.io
Other
553 stars 325 forks source link

TP-Link 841n V10 + 940nV3 + 941ndV6 have less transmitting power (dbm) than specification? #555

Closed KK4 closed 8 years ago

KK4 commented 9 years ago

Ih tested three TP-Link 841n V10 + 940nV3 + 941ndV6, and all have 18 dbm only as max also, In LuCI the v9 has range "standard" , 19 dbm = 79 mW, 18 ... and so on the V10 has "standard", 18 dbm = 63 mW , 17 ... and so on (19 dbm is missing) the 940/941 have "standard", 20 dbm = 100 mw, 19 dbm ... and so on I think, "standard" should be the next step above always.

However, all these give effective 18 dbm = 63 mW only, even if I set "standard" or the next max

I tested this with both, the orginal firmware as well as with the last FF firmware, it is same. (the original firmware has three settings only, "high", "middle", "low".

I asked TP-Link technical support Matthias Korn Technical Support Engineer and got (for the 841n V10 only, for 940/041 they could nit provide the specifications ) the answer below:

"hier ist kein Fehler vorhanden sondern alles völlig normal"

"Laut den Spezifikationen des TL-WR841N(UN) Version 9.0 betragen die maximal gemessenen dbm Werte 20dbm für 11b, 21dbm für 11g und 20dbm für 11n.

Beim TL-WR841N(UN) Version 10.0 betragen diese Werte 19dbm für 11b, 21dbm für 11g und 21dbm für 11n. "

"die Spezifikationen für ihre Hardware Versionen von TL-WR940N und TL-WR941ND habe ich nicht vorliegen. "

This means, that these three have a clear worse transmitting power than the 841n V9 and than the 1043nd v2 divergent to the specifcations ?

pecifications I fond says for all hese three the same: 20dbm http://www.tp-link.com/lk/products/details/cat-9_TL-WR841N.html#specifications http://www.tp-link.com/lb/products/details/cat-9_TL-WR941ND.html

neocturne commented 9 years ago

This is correct. AFAIR, the limits are obtained by the WLAN driver from the ART partition (radio calibration data) from the flash.

The value "default" simply means that we don't set the limit at all, but use the default setting from the WLAN driver. This is usually the same as the maximum value displayed in the list, but we don't know for sure in the config mode, as the WLAN is not active then. I guess we'll look at the code to find out how the default value is determined eventually, so the "default" entry can be replaced by the correct value in the future, but it doesn't have very high priority as it is only a cosmetic issue.

Furthermore, it is unclear what these values mean at all. At which point in the transmission chain are these values measured? Do they contain the antenna gain? (Probably not) Wire loss? External amplifiers? (Probably don't exist on typical TP-LINK hardware).

It is probable that the limits are set in a way that ensures the regulations are observed with the default antennas. I think the default antennas have 3dBi, so a limit of 18dBm combined with some loss would yield a total of about 20dBm. In my opinion, this isn't really anything to care about. A difference of 2-3dB usually doesn't have a noticable effect in practice.

A better framework which could allow the user to set the antenna gain themselves if they have replaced the antennas, but still have meaningful defaults, would be nice. At the moment, the antenna gain is not used at all to determine the default TX power, leading to much too high TX power on some devices (see #94).

By the way, such questions are better fit for the Gluon mailing list (or some other related MLs like wlanware and OpenWrt).

Delta1977 commented 8 years ago

The same here on an TP-Link 841 v9 :

root@nodename:~# iwinfo mesh0 txpoewerlist …….. 15 dBm ( 31 mW)

greets

SenorKaffee commented 8 years ago

Checked with 841v8 (gluon-v2015.1-287-g6110d3d) and noticed the same thing, it's no longer running on max by default.

root@ff-re-Buddestr-Kuechenradio:~# iwinfo client0 txpowerlist [...] 14 dBm ( 25 mW)

15 dBm ( 31 mW)

39mW and 63mW seems to be a pretty large difference.

A-Kasper commented 8 years ago

Am 01.12.2015 um 08:08 schrieb Michael Herzog:

Checked with 841v8 (gluon-v2015.1-287-g6110d3d) and noticed the same thing, it's no longer running on max by default.

root@ff-re-Buddestr-Kuechenradio:~# iwinfo client0 txpowerlist [...] 14 dBm ( 25 mW) 15 dBm ( 31 mW)

  • 16 dBm ( 39 mW) 17 dBm ( 50 mW) 18 dBm ( 63 mW)

39mW and 63mW seems to be a pretty large difference.

in addition:

I've got two 1043 v2 with different max txpower values. I didn't handle to match them.

A-Kasper commented 8 years ago

https://dev.openwrt.org/browser/trunk/package/kernel/mac80211/patches I think the new patches should solve this. I'm not sure if gluon uses or ignores these patches. The most interesting ones are 402 and 404 402 forces the driver to accept the user settings instead of the ones in eeprom 404 disables analysing the neighbourhood for the correct regdomain I found these patches while searching for reasons, why the txpower under CC is restricted. The first attemp to fix it I found was: https://patchwork.ozlabs.org/patch/444666/ maybe this helps to fix the issues

neocturne commented 8 years ago

@A-Kasper, these patches are really old (I can't remember an OpenWrt version without them), and of course, Gluon contains them as well.

ATH_USER_REGD is already enabled by Gluon, so https://patchwork.ozlabs.org/patch/444666/ doesn't help either (without ATH_USER_REGD, we wouldn't be able to change the regdom from US to DE).

Looking at the OpenWrt code doesn't necessarily help much, the change might also be in the kernel. I'll try to bisect this, anything else is just a waste of time.

discsimo commented 8 years ago

You can solve this problem manually with the following steps (tried it on my TP-Link TL-WR841N v10, gluon-v2015.1-262-g77829cd):

• uci set wireless.radio0.country=00 • uci set wireless.radio0.txpower=21 • uci commit • reboot Afterwards iwinfo client0 txpowerlist looks like this (before it was set to 16dBm and uci set wireless.radio.txpower doesn't work): [...] 16 dBm ( 39 mW) 17 dBm ( 50 mW) *18 dBm ( 63 mW)

/etc/config/wireless contains also the correct settings. Seems not to be a generell driver problem, only with country code "DE" maximum txpower is set to 16dbm. Maybe this helps to fix the issue.

rotanid commented 8 years ago

@discsimo and you are sure you don't violate german law with that increased output power?

KK4 commented 8 years ago

I think it is according German law, what limits to 20dbm. There is no need to limit the txpower to 16 dbm

I compared the effective power output with original antennas for 841n v10, 1043nd v2 (current stabile firmware), 940nd v6 and then they all have same = 20dbm (with original tp-link firrmware it powers the same, exact 20dbm)

This is independent if the table goes until 18 only, or to 20, and what the asterix says or if there is no asterix.

Also, after that modification I checked country settings and as active with German figures and 20dbm. Looks for me like a bug causing confusion.

A-Kasper commented 8 years ago

you need to add 5db antenna gain. so the limit with standard antennas is correct. but we use some setups with long cable so the poer should be configurable as needed Am 12.12.2015 21:14 schrieb "KK4" notifications@github.com:

I think it is according German law, what limits to 20dbm. There is no need to limit the txpower to 16 dbm

I compared the effective power output with original antennas for 841n v10, 1043nd v2 (current stabile firmware), 940nd v6 and then they all have same = 20dbm (with original tp-link firrmware it powers the same, exact 20dbm)

This is independent if the table goes until 18 only, or to 20, and what the asterix says or if there is no asterix.

Also, after that modification I checked country settings and as active with German figures and 20dbm. Looks for me like a bug causing confusion.

— Reply to this email directly or view it on GitHub https://github.com/freifunk-gluon/gluon/issues/555#issuecomment-164186789 .

KK4 commented 8 years ago

don't think this is right so, because 1) they have different antenna gains, but restriction is same for all 2) they don't have 5dbm antenna gain with standard antennas, from where yoy have these figures? 3) If you are right, this would mean, that tp-links deliveres all routerss with wrong settings for DE, as the effective dbm is 20 with its antennas. 4) If" you are right, then all (also all installed 30.000 routers) are not according German law and need to get modifies, what means all will have about 50% less distance for meshes, 5) *If you are right, same is for Fritzbox and other routers. I simply cannot (will not) believe, that all these sold and installed routers are not correct working and until now no one complained that and forces a country wide update for all. This is not logic and against all life experience.

And that someone does use setup with long antgennas must not restrict all routers independent what antennas the use and also using standard antennas. Those users with own antennas should drop their output individually. What you say is, that all routers get the absolute minimun so that some can build big antennas it w/o need to modify any.

I think this restriction is made w/o need and I cannot see any sense, except you want to force stop working of meshing over more than 40 m

Marssl78 commented 8 years ago

I don't know if this has anything to mean in here, but I have an 841v10 running openwrt. I used a version of openwrt, somebody has built and put in the forums, which contains the tx-power-fix by neoraider. In the Luci tx-power menu, the pulldown list only offers 18db when i select channel 1 as wlan channel. when i switch to channel 7, the pulldown list offers even 20db. Does this give a hint to you or is it even more confusing? :-/

KK4 commented 8 years ago

I tested the experimental from PetaByteBoy made on this base, now as sysupgrade for the 841v9 to see what it will do with the old versions.

Result is terrible: 16dbm is hardcored, even with the workaround above it always is restrictedd to 16dbm = 39mW

root@841n_v9_a:~# iwinfo wlan0 txpowerlist 0 dBm ( 1 mW) 1 dBm ( 1 mW) 2 dBm ( 1 mW) 3 dBm ( 1 mW) 4 dBm ( 2 mW) 5 dBm ( 3 mW) 6 dBm ( 3 mW) 7 dBm ( 5 mW) 8 dBm ( 6 mW) 9 dBm ( 7 mW) 10 dBm ( 10 mW) 11 dBm ( 12 mW) 12 dBm ( 15 mW) 13 dBm ( 19 mW) 14 dBm ( 25 mW) 15 dBm ( 31 mW)

root@841n_v9_a:~# iw reg get country DE: DFS-ETSI (2400 - 2483 @ 40), (N/A, 20), (N/A) (5150 - 5250 @ 80), (N/A, 20), (N/A), NO-OUTDOOR (5250 - 5350 @ 80), (N/A, 20), (0 ms), NO-OUTDOOR, DFS (5470 - 5725 @ 160), (N/A, 27), (0 ms), DFS (57000 - 66000 @ 2160), (N/A, 40), (N/A)

step2a

KK4 commented 8 years ago

I come back to A-Kasper "you need to add 5db antenna gain."

I tolked with tp-link, and they confirmed, that the EIRP CE: <20dBm(2.4GHz) *inclusive standard antennas. Means, you have to add nothing. It is a common misundertanding, that the specification are w/o antenna if the antenna gain is told in additon.

I also refer also to http://www.com-magazin.de/tipps-tricks/wlan/wlan-strahlung-zulaessige-grenzwerte-920715.html (in German) "...Diese Angaben beziehen sich auf die Strahlungsleistung des gesamten Geräts inklusive Antennen. Die Strahlungsleistung als EIRP (Equivalent Isotropically Radiated Power) berechnet sich wie folgt: elektrische Sendeleistung des Access-Points (in dBm) plus Antennengewinn (in dBi) minus Dämpfung durch Antennenkabel, Stecker und so weiter (in dB)..."

Means, as all the tp-link models delivered in germany have CE: <20dBm(2.4GHz they are specified as 20dbm maximal incl. antenas as delivered.

Furthermore, I made a comparison with Fritzbox 7360, and they have exact same total power like tp-link 841n v10, 950n v3 in original, and same if take the current experimental firmware and lift the txpower=20

So I believe, those who are responsible for this restriction to 16dbm simply made a fallacy, and never tested and carry out measurements with the hardwarte in case and compared it.

If someone does use other antennas, it is his responsibility to lower the power, that is not the programmers responsibility..

A-Kasper commented 8 years ago

Hoe exactly, does the chip know which Antennas are connected? It's a common SoC which is used in a lot of devices. The driver can't detect the antenna gain. That's the problem. OpenWRT wants to stay legal, so I think they set the maxvalue for the controller txpower to 16 dB.

The relevant EIRP Power is: Controller Outputpower + Antenna Gain - Cable-Loss

In my mind we can just set the Controller Outputpower by Software. If you set it to 16 dB and add the 5 dB a-gain you'll get the legal 100mW.

If the value you set in the driver the antennagain have to be stored in the controller flash. That's possible, but I don't think so. I guess, that the txpower in the old firmware was to high and ignored the Antennagain.

Am 16.12.2015 00:47 schrieb "KK4" notifications@github.com:

I come back to A-Kasper "you need to add 5db antenna gain."

I tolked with tp-link, and they confirmed, that the EIRP CE: <20dBm(2.4GHz) *inclusive standard antennas. Means, you have to add nothing. It is a common misundertanding, that the specification are w/o antenna if the antenna gain is told in additon.

I also refer also to http://www.com-magazin.de/tipps-tricks/wlan/wlan-strahlung-zulaessige-grenzwerte-920715.html (in German) "...Diese Angaben beziehen sich auf die Strahlungsleistung des gesamten Geräts inklusive Antennen. Die Strahlungsleistung als EIRP (Equivalent Isotropically Radiated Power) berechnet sich wie folgt: elektrische Sendeleistung des Access-Points (in dBm) plus Antennengewinn (in dBi) minus Dämpfung durch Antennenkabel, Stecker und so weiter (in dB)..."

Means, as all the tp-link models delivered in germany have CE: <20dBm(2.4GHz they are specified as 20dbm maximal incl. antenas as delivered.

Furthermore, I made a comparison with Fritzbox 7360, and they have exact same total power like tp-link 841n v10, 950n v3 in original, and same if take the current experimental firmware and lift the txpower=20

So I believe, those who are responsible for this restriction to 16dbm simply made a fallacy, and never tested and carry out measurements with the hardwarte in case and compared it.

If someone does use other antennas, it is his responsibility to lower the power, that is not the programmers responsibility..

— Reply to this email directly or view it on GitHub https://github.com/freifunk-gluon/gluon/issues/555#issuecomment-164934444 .

KK4 commented 8 years ago

this is the fallacy, exactly.

does the chip know which Antennas are connected? It's a common SoC which is used in a lot of devices. The driver can't detect the antenna gain

Ther chip does not need to know it. The point is the software setting by country (and its limittions) This the chip reads. And as the specification for the routers say EIRP <20dbm with original antennas (wich are 2 x 5dbm for 841 and 3x3 dbm for 950) it sets the ampere by this to have just that limit given by the country definitions. In original the hardware gives with setting DE (by software), before antennas a physically output of 9dbm, plus antennas = 19 (for the 841 with 2x5dbm), resp. 11-12 for the 940 with 3x3 dbm plus antennas = 20 . As you say

The relevant EIRP Power is: Controller Outputpower + Antenna Gain - Cable-Loss

yes, this is inclusive antenna gain ! (for the original <20 dbm)

The versions up to v9 all have it correctly, and ther effective output is same as the output of the original routers..

If you don't believe it carry out measurements with the hardware. Routers with the new experimental firmwares have ~60% of power only compared with original factory hardware. Means, you run into a fallacy

neocturne commented 8 years ago

I think this discussion is not really getting forward, and there are some misunderstandings.

@KK4, you are right, the transmission power relevant for the regulations does include the antenna gain. I don't think there were any misunderstandings here. The basic question was if the TX power displayed by OpenWrt does include the antenna gain, and we're still not sure if it does. These are the possible explanations:

  1. If does -> The displayed value will only be correct with the stock antennas, as the router can't know that different antennas are used
  2. It doesn't -> The 20dBm that were shows until Barrier Breaker were much too high, and the 16dBm on Chaos Calmer still are (?). (3. It contains a random offset?)

The only way to know what the values displayed on OpenWrt actually mean is to measure the actual output with professional equipment and compare it with the stock firmware.

As someone has mentioned in this thread, the 20dBm that Barrier Breaker is showing seem to be the same as the maximum TX power of the stock firmware, and the 16dBm on Chaos Calmer are really 4dB less (and it is not just a display issue). If these measurements are correct, this would mean that explanation 1 is correct.

KK4 commented 8 years ago

you can do a simple comparison: original stock tp-link 841n v10 with setting DE gives same result as experimental with txpower=20 + country=00 and v1043nd v2 with stabile firmware, all with original antennas

01

02

CodeFetch commented 8 years ago

@KK4 Did you really have to open up this discussion?

I did measurements to compare the 1043ND v2 and 940N v3 a while ago with US-regdom. The 940N has a MUCH higher maximum EIRP than the 1043ND. The reason for that might be that the TP9343 built into the 940N was actually developed for the chinese mainland.

In original the hardware gives with setting DE (by software), before antennas a physically output of 9dbm, plus antennas = 19 (for the 841 with 2x5dbm), resp. 11-12 for the 940 with 3x3 dbm plus antennas = 20 .

That is wrong and please don't call antenna gain "dBm". It's "dBi".

I tolked with tp-link, and they confirmed, that the EIRP CE: <20dBm(2.4GHz) *inclusive standard antennas.

In their definition of "EIRP"...

Actually I have never seen a N wifi router that observes restrictions of the Bundesnetzagentur and I never heard of the Bundesnetzagentur measuring EIRP of wifi devices. Do I really have to go into detail?

I think the Gluon developers shouldn't need to worry about things that even the original distributors don't when it comes to regulations.

neocturne commented 8 years ago

If the Gluon master provides lower TX power than older versions, and than the stock firmware, we should fix that. I don't really care too much if TP-Link's measurements are accurate.

There's a patch now which might affect the issue: http://git.openwrt.org/?p=openwrt.git;a=commit;h=3550e5b2ec6bc49c2b73b00989a5e79fa9860c6c I'll check if it fixes the issue.

neocturne commented 8 years ago

I've bisected the change to http://git.openwrt.org/?p=openwrt.git;a=commit;h=37a90b43a1d8f9be05ca461e700a59c844715740 ; AFAICT, this is not a bug, but an intentional change.

I've also tried to measure the TX power on the reported devices. On the TL-WR841N v9 and v10, new TX power (which is reported as 16 dBm on OpenWrt) seems to be the same as used the stock firmware when set to German regdom; this leads me to the conclusion that the new value is correct, and the old one was too high.

For the TL-WR940N v3, the TX power seems to be a bit lower in OpenWrt than in the stock firmware, but only by about 2-3 dB. OpenWrt reports 14 dBm here.

That said, I have to admit that my test setup wasn't very professional, I just used my Android phone with the app Wifi Analyzer to measure the power. Occasional disturbances had a much higher impact than the differences I tried to measure (sometimes about 5 dB), so everything has to be taken with a grain of salt. I've tried my best to get reproducible results by measuring for 5-10 minutes and repeating the measurements a few times.

viisauksena commented 8 years ago

also ein paar Tests aus Freiburg dazu : setting: 2 Router beide mit 2016.1 (master, latest fw und der neuen Statuspage und neighbour)

ssh login auf testrouter und kontrolle über statuspage 2. am router! (2. Router 841N v9 und später mit gleichem Ergebnis noch ein 710Nv1)

Kontroll settings;
iw client0 set txpower fixed 100 ; sleep 1 ; iw client0 set txpower fixed 2000 ; sleep 1 ; iw client0 set txpower fixed 1000 ; sleep 1; iw client0 set txpower fi xed 2000 liefert schöne treppen und läst sich beliebig verfeinern in der Ausgabe der Statuspage auf den Kontrollroutern

Am Testrouter:

  1. wechsel der regd : Anzeige bei (iw reg set DE) db -> 16 und bei (iw reg set 00) db -> 19 kontrolle: iwinfo |grep Tx Wirkung auf den 2. Router in der Db Anzeige keiner Gleicher Test wie zuvor verlief gleichbleibend, also wechsel der txpower, liefert instant schöne treppchen.
  2. Test : reg BO und reg 00 und reg DE Ergebnis : zw. reg DE und 00 kein Unterschied (nur die Anzeige TX 16 und 19) zw. reg BO und den anderen deutlicher Unterschied im Empfang , obwohl Anzeige 00 und BO je 19DB sagen Befehl zum beobachten am Kontrollrouter; iw reg set BO ; sleep 2; iw reg set DE ; sleep 2 ; iw reg set BO; sleep 2; iw reg set 00; sleep 2 ; iw reg set DE ; sleep 2; iw reg set BO Kontrolle wieder iwinfo |grep Tx reg BO verhält sich hier wie TW ID QA DZ PK .

Bedeutet für mich, ja in DE (regd) steht da jetzt 16db, das sind aber sonst auch immerschon 16 db gewesen, nur das die da die Antennengewinn mit reingerechnet haben. Die 3 Db hätten schwach aber sichtbar sein müssen, empfang Router 1 -35 Db , Router 2 -50 Db

DISTRIB_REVISION='r47335' DISTRIB_CODENAME='chaos_calmer' DISTRIB_TARGET='ar71xx/generic' DISTRIB_DESCRIPTION='OpenWrt Chaos Calmer 15.05' TP 841 N v9 - normale Antennen am Testrouter.

hoffe das hilft mal zur klärung - mir hats gehoöfen, hab mir schon sorgen gemacht.

dominikhlbg commented 8 years ago

@NeoRaider It's not correct. The problem didn't already exists in your link. The lower tx power problem began in patch Bleeding Edge, r43032 at the same day of commit. http://git.openwrt.org/?p=openwrt.git;a=commit;h=0e82759adb84c5c5fc9c78243422446c550acd19

KK4 commented 8 years ago

ein anderer Vergleich wäre einfacher: das Orginalgerät ist sicherlich nicht hstärker, als die Bestimmugnen erlauben. Sonst wäre alle Geräte illegal. Ich hab noch mal mit TP-Link Support telefoniert, und die sagen:

Also einfach vergleichen: ein Originalgerät mit TP-Link Orginalfirmware und ein 841v10 mit neuem gluon. Wenn beide dieselbe Leistung haben, ist alles OK, ist das 841v10 mit neuem Guon schwächer, ist was unnötig gedrosselt. Vergleichergebnis: mit neuem Guon ist das 841v10 mehr als 50% schwächer, (ein db = Verdoppelung der Leistung) und mit Änderung auf Land = 00 ist die Leistung identisch.

Das alte 841v9 bringt incl. Antennengewinn 18, das neue 19, erlaubt wären 20 = 100 mW

Edit: ich vermute, dass deswegen die alten Firmwares auch alle bei Land = 00 eingstellt waren. Weil die Kanalwahl für Freifunk sowieso nicht darüber gemacht wird. Nehme ich nun aber das neue Gluon als Update für die alten (v9 + v8 etc.), werden die auch gedrosselt.

A-Kasper commented 8 years ago

(ein db = Verdoppelung der Leistung)

3dB sind ungefähr eine Verdopellung der Leistung. Ich sehe es aber auch so, dass die Bremse wieder raus muss. "Nur 3dB Unterschied" sind eben 50% Leistung.

Adorfer commented 8 years ago

Es hat für mich den Eindruck, dass die Ermittlung der zulässigen maximum-txpower als "kleinster gemeinsamer Nenner" gefahren wird. Wenn man erstmal die Regdomain, den Kanal und den HT-Mode richtig setzt, die Initialisierung abwartet und dann nochmal iwinfo radio0 txpowerlist holt, dann gibt es sowohl für den 841v9 als auch für den 841v10 (zumindest auf den unteren Kanälen) wieder 20dB als zulässigen Einstellwert.

viisauksena commented 8 years ago

@Adorfer ich dachte ht-mode support is dropped ( http://gluon.readthedocs.org/en/latest/releases/v2016.1.html ) Kanal und Regdomain über die site.conf sind nicht ausreichend, oder meinst du das man das trotzdem noch setzen soll?

Adorfer commented 8 years ago

@viisauksena Ich verstehe den Zusammenhang nicht. Ausserdem denke ich, dass die Frage der korrekten TX-Powerlist nicht Thema dieses Issues ist. Hier geht es ja um die fehlende Leistungstufen "generell" und nicht nur das Problem vom Setup-Mode.

viisauksena commented 8 years ago

@Adorfer .. ich beziehe mich auf deine Aussage "Wenn man erstmal die Regdomain, den Kanal und den HT-Mode richtig setzt, die Initialisierung abwartet" was / wie meinst denn du das das jeweils umgesetzt wird am einzelnen router wenn nicht durch die FW und damit die site.conf

Adorfer commented 8 years ago

@viisauksena Dann war meine Aussage falsch und ich bin hier im falschen Issue, da es hier offensichtich um Openwrt-Probleme geht und nicht um Gluon-Issues. Mein Anliegen ist daher (jetzt) in #641

Zu hier referenzierten 841v10 vermag ich nur zu vermelden: Der Fehler besteht nicht (mehr). Zumindest nicht, wenn man /etc/config/wireless richtig einstellt.

Ich gehe davon aus, dass das -wenn es mal bestanden hat- inzwischen upstream gefixt wurde.

@KK4 Ich sehe von Deinen Screenshots, dass da "192.168.1.1" steht als IP des Knotens. Falls der Node da noch im Setup-Mode war: Dann stimmt die TX-Powerlist sowieso nicht. Du musst "im Betrieb" schauen.

SimJoSt commented 8 years ago

After updating a TP-LINK WDR3600 I noticed, that it had reduced transmission power. My CPE210 with the new status page showed it with -85dbm - -90dbm before the update. After the update it wasn't shown at all. I cannot contribute anything to whether it was set on a too high value before or is now beneath the limit.

I will perform any requested tests, if possible.

RubenKelevra commented 8 years ago

We just performed a throughput test with two freifunk-nodes, we used one TL-WDR4300 and one TL-WDR3600 and placed them in the following three setups:

We used two VeEX VePAL TX300s to transmit 50 mbit/s in one direction and 1 mbit/s in the other via one continuous stream. We used L2 ethernet packets with 1518 bytes length.

The results are the following for gluon 2014.4 (IBSS):

The results for gluon with the latest stable (2016.1.x) with the following results (IBSS):

To push it to the limits, we increased the sending speed to 70 mbit/s instead of 50 mbit/s


Conclusion:

We learned that the newer firmware performed much better, in our setup. We are open for suggestions for further tests if desired.

img_8503

SimJoSt commented 8 years ago

Nice benchmark, interesting results. In order to adress the core of this issue, a maximum range test with both firmwares would be much appreciated. How far can the routers be apart until they don't link/mesh anymore, don't see each other anymore. My interest would be satisfied with results taken from the statuspage. The setup doesn't matter to me as well. Outside or with a lot of walls, I just want to know if the new firmware stops linking earlier.

RubenKelevra commented 8 years ago

@SimJoSt what do you think about an dBm-test on the antenna-connectors? That would be much more comparable and reproducible.

The overall range coverage is much reduced on our newer firmware, we have already the b-disable patch included for a year now: https://github.com/freifunk-gluon/gluon/pull/810

So our "old" firmware does not include it, but the new one does.

We don't think this change anything on this setup, since we stayed well above the basic_rates - in the extended rates - which are changed with the patch.

So in the new firmware does not get a that huge coverage but the links are much more usable if connected and the overall mesh is improved, because the hotspots does not try to send with very low end transmit rates to get some meters more range.

Best regards

Ruben

SimJoSt commented 8 years ago

Oh, I didn't know that. Thanks! Of course a dBm test would be much more perfect.

RubenKelevra commented 8 years ago

Actually a dBm test seems to be hard as well, cause we use TPC per packet now. But I think the beacons are still send with maximum txpower. So I think we can measure them with 1 Mbit on new/old firmware.

The txpower is usually different on different datarates so with actual data it's hard to measure. I think we'll borrow some hardware if available and give it a try. :)

//cc @bitnukl

neocturne commented 8 years ago

I have just verified that the maximum TX power of 16dBm seen on the TL-WR841 v10 (current LEDE, country DE) is the same power the stock firmware uses on level "high" when country is set to Germany.

Verified using devmem:

# devmem 0x181080e8
0x3F3F1818

The returned value is the same on the stock firmware and LEDE. I expect that tests on other devices will yield similar results.

Thus, I conclude that the new behaviour of OpenWrt/LEDE is correct and this is not a bug.

RubenKelevra commented 8 years ago

@NeoRaider thanks for pointing this out :)

jyotirajsharma commented 7 years ago

I have TP-Link 841n (V10) router. To run linux command, do I need to install OpenWrt firmware again or I can run linux command directly connecting router via serial console??