WHR-HP-G54 Flag 0x3758 instead of 0x2758/0x1758 ???

Post new topic   Reply to topic    DD-WRT Forum Forum Index -> Broadcom SoC based Hardware
Goto page 1, 2, 3, 4, 5, 6, 7  Next
Author Message
christoph78
DD-WRT Novice


Joined: 12 May 2007
Posts: 4

PostPosted: Sat May 12, 2007 9:06    Post subject: WHR-HP-G54 Flag 0x3758 instead of 0x2758/0x1758 ??? Reply with quote
Hi,
the discussion about WHR-G54S and WHR-HPG54 is quite confusing here on DDWRT Forums. In the following links it is recommended to use flag 0x3758 for the HP for keeping noise low. What do yout think about his opinion?

http://wiki.leipzig.freifunk.net/Buffalo_WHR-G54S

http://www.freifunk-halle.de/phpBB2/viewtopic.php?p=2807&sid=8430f6843894e5917226be877bda667f

Quote:

4huf:

Eine Änderung der Bootflags von 0x1... zu 0x2... erscheint unsinnig und eher falsch!
Wenn schon dann in 0x3...
Die Flags bedeuten (aus bcmdevs.h):
\#define BFL_EXTLNA 0x1000 /* This board has an external LNA */
\#define BFL_HGPA 0x2000 /* This board has a high gain PA */
Diese Flags werden bitweise addiert.
Eine Änderung von 0x1 auf 0x2 würde den "external LNA" aus und die "high gain PA" einschalten.
Auf meinen 2 WHR-HP-G54 steigt das Rauschen bei ausgeschaltetem "external LNA" mit "wl noise" von ca. -92db auf -83db an !
Das heißt das Signal-Rauschverhältniss verschlechtert sich um fast 10db !!
Das Signal geht durch eine Diode (vermutlich PIN-Diode zum Schutz oder zur Dämpfung (!) des Empfangs) und durch ein Teil mit der Bezeichung A832 (vermutlich npn-Transistor).
Die pin-Diode wird im Takt der Sendeimpulse angesteuert. Ev. Schutz des Transistors im Sendefall.
Beim Senden kann ich keinen Unterschied zwischen "high gain PA" on/off messen !
"wl txpwr 20" = 50mW
"wl txpwr 40" = 100mW
"wl txpwr 80" = 200mW
"wl txpwr 84" = 250mW
Es sind keine höheren Werte als "wl txpwr 84" sinnvoll. Man kann zwar mehr einstellen aber die gemessene HF-Leistung bleibt gleich.
Der PA-Chip "AWL6153" könnte sogar bis zu 630mW (802.11b-Modulation, 5V Spannung).
Im WHR-HP-G54 wird er mit 4,5V getrieben.
Offenbar reicht dann die Ansteuerleistung nicht mehr.
Die Anzeige des Override-Modus funktioniert scheinbar auch nicht (Anzeige immer "off").
Bei Benutzung des Befehle "wl txpwr1" ohne "-o" wird maximal 100mW gesendet, unabhängig von höheren Einstellungen.
Gemessen wurde mit einem AD8313 Taskkopf und 30db-Teiler.
Da WLAN nur schmale Impulse sendet erfolgte die Messung mit einem "cat /dev/zero" per ssh für eine Art Dauersendung und Spitzenwert-Messung.
Ich will nicht ausschließen das der Messfehler sehr groß sein kann.
Fakt ist: der WHR-HP-G54 verwendet eine andere CPU (BCM95352E) als der WHR-G54S (ohne E).
Es ist die CPU der 2. Generation, laut Broadcom mit "Signalprozessorfähigkeiten" für WLAN.
Damit soll bis zu 50% mehr Reichweite möglich sein.

Sponsor
GeeTek
DD-WRT Guru


Joined: 06 Jun 2006
Posts: 3768
Location: I'm the one on the plate.

PostPosted: Sat May 12, 2007 9:58    Post subject: Reply with quote
For those of us not familiar with the local lingo, could somebody summarize what was posted about 0x3758 ?
_________________
http://69.175.13.131:8015 Streaming Week-End Disco. Station Ripper V 1.1 will do.
Eko
DD-WRT Developer/Maintainer


Joined: 07 Jun 2006
Posts: 5772

PostPosted: Sat May 12, 2007 10:03    Post subject: Reply with quote
Could be true. Someone must test.
Eko
DD-WRT Developer/Maintainer


Joined: 07 Jun 2006
Posts: 5772

PostPosted: Sat May 12, 2007 10:08    Post subject: Reply with quote
Point is:
when we want to enable amp on HP, we should not disable internal now noise amp.
Jays
DD-WRT User


Joined: 02 May 2007
Posts: 91

PostPosted: Sat May 12, 2007 10:17    Post subject: Reply with quote
Very interesting..

The explanation sounds good, but this is strange:

"beim Senden kann ich keinen Unterschied zwischen "high gain PA" on/off messen !"

This means he isn't seeing a difference between 0x1 and 0x2 when transmitting, but a lot of people here noticed a stronger signal.

However if he's right and with other people noticing a difference between 0x1 and 0x2, 0x3 seems the way to go?
BS, any reason not to switch to 0x3?

Another question about the boardflags, when looking at http://forum.bsr-clan.de/viewtopic.php?p=42737#42737 , how can they be 0x?758
Eko
DD-WRT Developer/Maintainer


Joined: 07 Jun 2006
Posts: 5772

PostPosted: Sat May 12, 2007 10:20    Post subject: Reply with quote
Jays wrote:


Another question about the boardflags, when looking at http://forum.bsr-clan.de/viewtopic.php?p=42737#42737 , how can they be 0x?758


I know..... don't know!
GeeTek
DD-WRT Guru


Joined: 06 Jun 2006
Posts: 3768
Location: I'm the one on the plate.

PostPosted: Sat May 12, 2007 11:14    Post subject: Reply with quote
Jays wrote:
"beim Senden kann ich keinen Unterschied zwischen "high gain PA" on/off messen !"

This means he isn't seeing a difference between 0x1 and 0x2 when transmitting, but a lot of people here noticed a stronger signal.

Well since I am from Texas, I only speak a little English, Spanish and GWbasic. Anyway, I have noticed, and I thought that it was common knowledge, that 0x1758 and 0x2758 only control the pre-amp, and not the output power. I know that this turns on and off the pre-amp, and has no effect on the output amp, so I cannot imagine what 0x3875 might do, but I will give it a try. Thanks for the translation. Razz

_________________
http://69.175.13.131:8015 Streaming Week-End Disco. Station Ripper V 1.1 will do.
christoph78
DD-WRT Novice


Joined: 12 May 2007
Posts: 4

PostPosted: Sat May 12, 2007 11:30    Post subject: For ENLGISH speakers HP-G54 Flag 0x3758 Reply with quote
IN ENGLISH

Quote:

4huf:
a bootflag change from 0x1 to 0x2 is makes no sense and is wrong!
because changing 0x1 to 0x2 would disable external LNA and enable high gain PA
the flags mean (from bcmdevs.h)
\#define BFL_EXTLNA 0x1000 /* This board has an external LNA */
\#define BFL_HGPA 0x2000 /* This board has a high gain PA */
the flags are added bitwise
on his WHR-HP-G54 machines, the noise level rises from -92db to -83db when disabling external LNA (SNR -10db)
The way of signal goes through a diode (probably PIN Diode for protection or rx-attenuation) and through a part labeled A832 (probably npn transistor)

the pin diode is controled synchronously to tx power impulses.
He cant measure any difference between enabled and disabled "high gain PA"
"wl txpwr 20" = 50mW
"wl txpwr 40" = 100mW
"wl txpwr 80" = 200mW
"wl txpwr 84" = 250mW

higher levels than 84 are not useful. tx power won rise further.
The PA Chip AWL 6153 has acapability of 630mW in 11b modulation with Uin=5V
It is driven in WHR-HP-G54 with 4.5 V
if using the command wl txpwr1 without "-o" tx powers maximum level is 100mW, independent of any other commands

Measurement was carried out using a AD8313 probe head and 30dB splitter using a "cat/dev/zero command by ssh for continoous tx and maxmimum value measurement.
The measurement error may be quite high.
Fact is: WHR-HP-G54 has the CPU BCM95352E. Its a 2nd generation CPU with signal processor ability (WHR-G54S has only the old BCM95352)

GeeTek
DD-WRT Guru


Joined: 06 Jun 2006
Posts: 3768
Location: I'm the one on the plate.

PostPosted: Sat May 12, 2007 11:46    Post subject: Reply with quote
Thanks christoph, you translated "if using the command wl txpwr1 without "-o" tx powers maximum level is 100mW, independent of any other commands "
Transmit power conrol is something I have not heard of yet. txpwr1 is something I would like to find out more about.

Also from that thread is
"Boardflags original 0x1758
"wl noise" = -92db

Boardflags = 0x2758
"wl noise" = -83db !!

Boardflag = 0x3758
"wl noise" = -92db

Im DD-WRT-Wiki wird empfohlen für den HP von 0x1578 auf 0x2578 zu ändern.
Das ist definitiv falsch . Wenn schon dann auf 0x3578 !!
Denn mit der Änderung von 1 auf 2 würde ein vorhandene Rx-Vorverstärker abgeschaltet. "


This looks like the pre-amp control. I think only 0x1758 and 0x2758 are valid settings. 0x3758 or 0x1758, or any other setting, is not the valid 0x2758 setting for pre-amp "ON", so anything else, like 0x3758 is pre-amp off. I have 6 new HP radios in the box and 9 regular buffaloes in the box. If anybody can think of some testing, I will be happy to do what I can. I will test the 0x3758 boardflag on the HP and see if I can see anything different from the other 2 settings.

_________________
http://69.175.13.131:8015 Streaming Week-End Disco. Station Ripper V 1.1 will do.
u3gyxap
DD-WRT User


Joined: 07 Jun 2006
Posts: 298

PostPosted: Sat May 12, 2007 13:11    Post subject: Reply with quote
All these settings concern identifying Rx amplification only. Tx amp is always on, no matter how hard you try to play with boardflags.
Eko
DD-WRT Developer/Maintainer


Joined: 07 Jun 2006
Posts: 5772

PostPosted: Sat May 12, 2007 13:21    Post subject: Reply with quote
GeeTek: these are bit values:

Code:
PA  LNA    boarflag
0    1  =  1
1    0  =  2
1    1  =  3


So, to enable LNA, you use 1 or 3 (3 = 2^1 + 2^0)
Jays
DD-WRT User


Joined: 02 May 2007
Posts: 91

PostPosted: Sat May 12, 2007 14:47    Post subject: Reply with quote
a HP with Tomato as client of an DD-wrt hp ap:

1. 0x2758
RSSI -46-50
Quality 30

2. 0x1758
Rssi -63
quality 23

3. 0x3758
rssi -62
quality 32

I set the flag, reboot, watched about 10 minutes. Later i'll test it the other way round.


Last edited by Jays on Sat May 12, 2007 14:59; edited 1 time in total
shunt010
DD-WRT User


Joined: 17 Oct 2006
Posts: 147

PostPosted: Sat May 12, 2007 14:57    Post subject: Reply with quote
I've done some tests and can't see any difference between 1xxx and 3xxx either way around.

But would be interested in if someone does get this idea to work.

I've got a WHR-HP thing connected up a mast and I've set the boardflags to 0x1758 because with the preamp on, it appeared that the recieve was getting killed by interference.

I didn't notice any particular drop in signal strength at the other end either when I changed the boardflags, power out is set to a steady 200mw (and the things don't cook up!).

Be interested in people's results..
GeeTek
DD-WRT Guru


Joined: 06 Jun 2006
Posts: 3768
Location: I'm the one on the plate.

PostPosted: Sat May 12, 2007 15:05    Post subject: Reply with quote
Eko wrote:

(3 = 2^1 + 2^0)

I've been playing with it since posting last. Thanks for the responses. I need to put the radio on the tower to get better results. I have too much multi-path here in the metal trailor house. Measuring only RX signal from the tower radio and local radio shows only that 0x3758 has less gain on the pre-amp than 0x2758 or 0x1758. Again, my testing is faulty.

u3gyxap , Eko's chart seems to include TX settings. I'm still testing. I don't have a damned thing to do today, I have been awake since yesterday, and I'm drunk off my ass, so I'm fixing to get a pair of HP radios up on my towers (6 km distance) and give them hell. I'll post back.

Eko, your formula that I quoted shows 3 = two to the power of one plus two to the power of zero. This sums to 3=(4+2=6).I'm not sure I'm with you . I like your Bit Map showing LNA (Low Noise Amplifier = RX amp) and PA (Power amp = output amp). If you are telling me that boardflag settings for these 3 settings are 0x1758, 0x2758 and 0x3758 then please post back and say "Yes".

If you would like to hear a great radio station check out ;
http://85.93.129.4:8000/button6.ogg.m3u
(Russia, moon beat radio)

_________________
http://69.175.13.131:8015 Streaming Week-End Disco. Station Ripper V 1.1 will do.
Eko
DD-WRT Developer/Maintainer


Joined: 07 Jun 2006
Posts: 5772

PostPosted: Sat May 12, 2007 15:10    Post subject: Reply with quote
2^0 = 1
2^1 = 2
2^2 = 2*2 =4
2^3 = 2*2*2=8
etc
Goto page 1, 2, 3, 4, 5, 6, 7  Next Display posts from previous:    Page 1 of 7
Post new topic   Reply to topic    DD-WRT Forum Forum Index -> Broadcom SoC based Hardware All times are GMT

Navigation

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You can attach files in this forum
You can download files in this forum