PDA

View Full Version : Help with wireless network!? Periodic drops in performance...



NeilF
2007-09-17, 01:12
From day one my SB3 whad problems with my wireless network. I upgraded the bios of my Netgear DG834GT and that seemed to completely solve the problems.

However, out of the blue, this weekend I seem to be getting odd patched of poor performance.

Running a network test at times, even the 1000 or 2000 test are getting only 50-70%.

The SB3 (denoted 'S') is only about 10ft away from the router (denoted 'R'), but the router is in a study in an extension so sort of going through a double wall ('=')/single wall ('|'), but there are doors around etc.

The router is on my desk along with two PCs and a monitor etc. The SB3 is waist high on a shelf on a oak cupboard. But like I said, generally, the signal does seem 100% rock solid the majority of the time, with 100% on a 5000 test. So the question is what can I do to help with these odd periods of poor performance?


..........................S.....................
=======........===========
...R.........|.................................
..............|.................................
..............|.................................
.................................................
=====================


Running netstumbler shows no other network... What sort of db would you expect on the graph it shows?

Any advice/tips?

Could I plug this into the router for example? http://www.maplin.co.uk/module.aspx?ModuleNo=38211&doy=17m9#overview


ps: The poor performance doesn't seem connected to any other devices be on/used.

Ramage
2007-09-17, 03:28
The possibilities are, interference, reflection or absorbtion. You rule out other wifi networks in the vicinity, but cordless phones and microwaves can have a significant effect if operated near the router or SB3.

I have a DG834 and find that adjusting the antenna orientation can have a significant effect on signal level at the SB3.

To avoid reflection and absorbtion, the router should not be sited close to metal objects and other electronic kit. I have also found that the presence of people in the radio path can affect the signal levels.

Such is the nature of wireless.

You need to try various things, the Maplin antenna might solve your problem by increasing the radio path gain.

NeilF
2007-09-17, 03:40
The possibilities are, interference, reflection or absorbtion. You rule out other wifi networks in the vicinity, but cordless phones and microwaves can have a significant effect if operated near the router or SB3.

I have a DG834 and find that adjusting the antenna orientation can have a significant effect on signal level at the SB3.

To avoid reflection and absorbtion, the router should not be sited close to metal objects and other electronic kit. I have also found that the presence of people in the radio path can affect the signal levels.

Such is the nature of wireless.

You need to try various things, the Maplin antenna might solve your problem by increasing the radio path gain.


Well, what's odd is:-
- the microwave isn't in use at problem times.
- the phone isn't in use at problem times.
- the baby monitor isn't on at problem times.
- I've even tried turning off the TV etc.
- people aren't inbetween/near the router/SB3 at problem times.
- the antenna of the router isn't being moved.

I really just cannot see what is changing when I seemingly get poor(er) reception? Just seems to come out of the blue.

I could try that antenna from Maplins or a Tritium Flatenna (which will look pretty aweful!)? http://www.tritium.co.uk/

But with the problem being so wierd I'm not convinced I'll be solving it!

If I sit there with my laptop in NetStumbler would you expect me to see the signal/noise going down when my performance problem happens?

Ramage
2007-09-17, 03:51
I don't know about Netstumbler S/N, but I have found the results inconclusive.

Its worth trying reorienting the antenna anyway - what have you got to loose?

Moving the SB3 a few cm either way could also have a significant effect.

You could try this antenna, easy to build and may help - http://www.freeantennas.com/projects/template2/index.html

bpa
2007-09-17, 03:55
Before you try buying another antenna you could try this ( http://www.freeantennas.com/projects/template2/index.html )

Some PCs especially laptops alter the power output to Wifi - is your server wired or wireless ? if wireless - disable the wifi power mgmt.

Many house extension have added insulation which includes a foil lining so your antenna should be in line-of sight going through the doorway and not through a wall.

NeilF
2007-09-17, 04:01
Before you try buying another antenna you could try this ( http://www.freeantennas.com/projects/template2/index.html )

Some PCs especially laptops alter the power output to Wifi - is your server wired or wireless ? if wireless - disable the wifi power mgmt.

Many house extension have added insulation which includes a foil lining so your antenna should be in line-of sight going through the doorway and not through a wall.

All my machines are wired, except a laptop, which is typically off when the problem occurs...

I can't see it's anything to do with things like insulations etc etc as most of the time the network is rock solid with 100% on the 5000 test!

Line of sight is impossible! If you look at the diagram at the very least theres a single brick wall to go thru to the SB3 from the router...

I seen on ebay there are loads of upgrade aerials for just 4-5!

OldeSchule
2007-09-17, 07:41
If you're in an apartment/condo/flat/multi-family dwelling, perhaps a neighbor got a new phone?

Also, the problem isn't with _all_ wireless phones, only 2.4GHz ones, since they share the same general frequency as WiFi. Newer 5.8GHz phones don't cause any interference (unless you're running 802.11a, but that's a whole different conversation).

NeilF
2007-09-17, 07:53
If you're in an apartment/condo/flat/multi-family dwelling, perhaps a neighbor got a new phone?

Also, the problem isn't with _all_ wireless phones, only 2.4GHz ones, since they share the same general frequency as WiFi. Newer 5.8GHz phones don't cause any interference (unless you're running 802.11a, but that's a whole different conversation).

Detached house...

NeilF
2007-09-17, 10:43
Over the course of the past hour its been getting worse and worse.

The 5000 network test is now averaging about 10-30%!

I've rebooted the router. Changed its channel etc etc... All to no avail.

If I put a book with a sheet of tin foil to one side of the SB3 and the rooter with little if any effect!

Usually this thing just sits there at 100%!! This is doing my head in!!!



EDIT -- It's now down to about 11%!!!!!! Netstumbler is showing (on a latop near the sb3) about -60db??? Is that normal or bad? NExt to the router it reads about -35db?!?

EDIT2 -- It's now returned to 100% with me doing nothing!!! Netstumbler shows no change in the signal near the sb3! FFS!

bpa
2007-09-17, 10:48
Could be a dying router - have you checked TCP network stats when doing using wireless connection with laptop.

NeilF
2007-09-17, 11:14
Could be a dying router - have you checked TCP network stats when doing using wireless connection with laptop.

Sorry, how do I do that?

peterw
2007-09-17, 11:17
Netstumbler is showing (on a latop near the sb3) about -60db??? Is that normal or bad? NExt to the router it reads about -35db?!?


-35 next to the router sounds pretty bad.

NeilF
2007-09-17, 11:23
-35 next to the router sounds pretty bad.

Back down to 12% again on the network test!! I've not touched anything!

Little/no difference on the netstumbler log!

EDIT - Rebooted the PC and router and that helped... Rate is now flicking around 70-100%!

EDIT2 - It dropped down again. Turned off the router for about 3-4 mins. On turning it back on... SB3 was at 100%. Also I'll swear the graph on Netstumbler, although it still goes up and down, is smoother. ie: Rather than it jumping up and down almost every pixel, it sits at one level for 5-10 pixels at a time, before going up/down! Why isn't the graph flat? Why is it jumping up and down? Is this all pointing to a bad router? With intermittent/poor wireless? Surely wireless would either work or not work?


How do we know this isn't the Squeezebox playing up? Surely it's half of the equation here. Both the router and the SB are both sending and receiving? How can we prove if it's the router or the SB at fault? PS: It's sitting at 100% rock solid again now! :/

NeilF
2007-09-18, 00:59
Here's some test this morning at 3000!

< 10 : 0 : 0%
< 20 : 0 : 0%
< 30 : 0 : 0%
< 40 : 0 : 0%
< 50 : 0 : 0%
< 60 : 1 : 33% ################
< 70 : 2 : 67% #################################
< 75 : 0 : 0%
< 80 : 0 : 0%
< 85 : 0 : 0%
< 90 : 0 : 0%
< 95 : 0 : 0%
< 100 : 0 : 0%
>=100 : 0 : 0%
max : 66.666667
min : 57.000000
avg : 61.760943


< 10 : 0 : 0%
< 20 : 0 : 0%
< 30 : 0 : 0%
< 40 : 0 : 0%
< 50 : 0 : 0%
< 60 : 0 : 0%
< 70 : 0 : 0%
< 75 : 0 : 0%
< 80 : 0 : 0%
< 85 : 0 : 0%
< 90 : 0 : 0%
< 95 : 0 : 0%
< 100 : 0 : 0%
>=100 : 2 :100% ##################################################
max : 100.000000
min : 100.000000
avg : 100.000000


< 10 : 0 : 0%
< 20 : 0 : 0%
< 30 : 0 : 0%
< 40 : 0 : 0%
< 50 : 0 : 0%
< 60 : 0 : 0%
< 70 : 0 : 0%
< 75 : 0 : 0%
< 80 : 1 : 50% #########################
< 85 : 0 : 0%
< 90 : 0 : 0%
< 95 : 0 : 0%
< 100 : 0 : 0%
>=100 : 1 : 50% #########################
max : 100.000000
min : 78.823529
avg : 89.411765


< 10 : 0 : 0%
< 20 : 0 : 0%
< 30 : 0 : 0%
< 40 : 1 : 50% #########################
< 50 : 0 : 0%
< 60 : 0 : 0%
< 70 : 0 : 0%
< 75 : 0 : 0%
< 80 : 0 : 0%
< 85 : 0 : 0%
< 90 : 0 : 0%
< 95 : 0 : 0%
< 100 : 0 : 0%
>=100 : 1 : 50% #########################
max : 100.000000
min : 35.000000
avg : 67.500000


Better than last night, but still surely not consistant? The majority (say 8/10) are giving 100!

bpa
2007-09-18, 02:22
TO get TCP network stats on Windows. Open a command prompt window and run the command "netstat -s -p tcp".

The statistics are a running count off all TCP traffic so to see if there are errors you need to
1. run the netstat command
2. Note the "segments retransmitted" and "reset connection"
3. run a test - the best test would be long in duration and a reasonable volume traffic.
4. run netstat and calculate change in "segments retransmitted" and "reset connection"

For a good local connection the change in "retransmitted" will be zero or low single digits. Connection reset should be zero. Keep internet traffic to a minimum as Windows bundles ALL TCP traffic into the counts.

Run the test between wireless laptop and wired server and look at counts on both PCs. If the error counts are not big then it points to a SB issue.

Look at statistics between server PC and SB and see if anything unusual shows up.

Poor signal quality usually result in a good but slow wireless link. Inteference (e..g. microwave) usually results in errors and link resets or packets retransmitted.

NeilF
2007-09-18, 02:41
TO get TCP network stats on Windows. Open a command prompt window and run the command "netstat -s -p tcp".


Thanks for that!! Next time I get poor performance I'll do this on the laptop and see if it points to anything!

The only problem is I suspect it'll show nothing and I'll still be left wondering if the router or SB (or something else) is to blame.


I have another plan/option as well - A friend of mine has a Netgear DG8834G router. I could swap our routers over and live with theres for a few weeks. If I get no problems that would sure imply it's my router. The only problem is, if I do still get spuradic problems that means it could be the SB or indeed something else affecting my network :/

At least that could help!?

What I find frustrating is I'd expect a wireless connection to either work or not work, not mostly work. Unless it's possible for the maybe the circuit in the router to periodically start gettings its frequencies slightly wrong which is what we're seeing every now and then for a couple of hours...


By the way, on a test at the moment I'm getting this on the SB. Do these figures look normal/abnormal to you? (Especially that signal strength one?):-

-ON A 3000 TEST from SlimServer-
< 10 : 0 : 0%
< 20 : 0 : 0%
< 30 : 0 : 0%
< 40 : 0 : 0%
< 50 : 0 : 0%
< 60 : 3 : 4% #
< 70 : 8 : 10% #####
< 75 : 3 : 4% #
< 80 : 1 : 1%
< 85 : 2 : 3% #
< 90 : 2 : 3% #
< 95 : 1 : 1%
< 100 : 1 : 1%
>=100 : 59 : 74% ####################################
max : 100.000000
min : 56.190476
avg : 92.709236



And then this test as well while playing a FLAC:-



Home / Server & Network Health / Player Statistics

The server is currently collecting performance statistics for this player.

You may set warning thresholds for each measurement. This will record in the server log whenever the threshold is exceeded. The most recent log entries can be viewed here.
Signal Strength
This graph shows the strength of the wireless signal received by your player. Higher signal strength is better. The player reports signal strength while it is playing.

< 10 : 0 : 0%
< 20 : 0 : 0%
< 30 : 0 : 0%
< 40 : 0 : 0%
< 50 : 0 : 0%
< 60 : 0 : 0%
< 70 : 10 : 2%
< 80 : 558 : 98% #################################################
< 90 : 0 : 0%
< 100 : 0 : 0%
>=100 : 0 : 0%
max : 78.000000
min : 66.000000
avg : 73.795775

Warning Thresholds : Low High
Buffer Fullness
This graph shows the fill of the player's buffer. Higher buffer fullness is better. Note the buffer is only filled while the player is playing tracks.

Squeezebox1 uses a small buffer and it is expected to stay full while playing. If this value drops to 0 it will result in audio dropouts. This is likely to be due to network problems.

Squeezebox2/3 uses a large buffer. This drains to 0 at the end of each track and then refills for the next track. You should only be concerned if the buffer fill is not high for the majority of the time a track is playing.

Playing remote streams can lead to low buffer fill as the player needs to wait for data from the remote server. This is not a cause for concern.

< 10 : 552 :100% ##################################################
< 20 : 0 : 0%
< 30 : 0 : 0%
< 40 : 0 : 0%
< 50 : 0 : 0%
< 60 : 0 : 0%
< 70 : 0 : 0%
< 80 : 0 : 0%
< 90 : 0 : 0%
< 100 : 0 : 0%
>=100 : 0 : 0%
max : 8.346939
min : 0.000000
avg : 0.037636

Warning Thresholds : Low High
Control Connection
This graph shows the number of messages queued up to send to the player over the control connection. A measurement is taken every time a new message is sent to the player. Values above 1-2 indicate potential network congestion or that the player has become disconnected.

< 1 : 15 : 79% #######################################
< 2 : 4 : 21% ##########
< 5 : 0 : 0%
< 10 : 0 : 0%
< 20 : 0 : 0%
>=20 : 0 : 0%
max : 1.000000
min : 0.000000
avg : 0.210526

bpa
2007-09-18, 03:25
You can run netstat on your server PC - it may show errors. I know of one user who had a dodgy ethernet interface in the PC and the problem showed as breaks in SB audio. File transfer and Web browsing on the PC just did normal recovery so the problem was not noticeable on data transfers which were not time critical.

NeilF
2007-09-18, 03:36
You can run netstat on your server PC - it may show errors. I know of one user who had a dodgy ethernet interface in the PC and the problem showed as breaks in SB audio. File transfer and Web browsing on the PC just did normal recovery so the problem was not noticeable on data transfers which were not time critical.

OK... So I'll run it on the server and my wireless laptop and see if I get any of the issues mentioned above... (I'll copy a directory of files to/from the laptop)

Don't forget the major odd thing is the problem is spuradic, with the SB being OK most of the time... :/

NeilF
2007-09-18, 11:22
Here's the result of a netstat test. I copied a folder containing just 3x13meg files from my server to my laptop wirelessly, about 5ft from the router through a wall/around corner through open door.

Netstat before:-
Active Opens: 98
Passive Opens: 3
Failed Connection Attempts: 12
Reset Connections: 38
Current Connections: 1
Segments Received: 37734
Segments Sent: 22050
Segments Retransmitted: 14

Netstat after:-
Active Opens: 127
Passive Opens: 3
Failed COnnection Attempts: 12
Reset Connections: 55
Current Connections: 2
Segments Received: 66226
Segments Sent: 37951
Segments Retransmitted: 17

Then did it wired in...

Netstat before:-
Active Opens: 162
Passive Opens: 5
Failed COnnection Attempts: 12
Reset Connections: 62
Current Connections: 22
Segments Received: 69952
Segments Sent: 40190
Segments Retransmitted: 17

Netstat after:-
Active Opens: 164
Passive Opens: 5
Failed COnnection Attempts: 12
Reset Connections: 80
Current Connections: 1
Segments Received: 97966
Segments Sent: 55274
Segments Retransmitted: 17


Point to anything?



Then did the wireless copy again with my laptop 1ft away from the router!

Netstat before:-
Active Opens: 167
Passive Opens: 5
Failed Connection Attempts: 13
Reset Connections: 82
Current Connections: 1
Segments Received: 100303
Segments Sent: 57133
Segments Retransmitted: 20


Netstat after:-
Active Opens: 167
Passive Opens: 5
Failed Connection Attempts: 13
Reset Connections: 82
Current Connections: 1
Segments Received: 127364
Segments Sent: 71272
Segments Retransmitted: 20

(still 1ft away, on a far larger copy the Segments Retransmitted went up by 1 but Failed Connection Attempts remained unchanged)

ps: At the moment if I try my 5000 network test on the SB it's getting 100%!


EDIT!!! Five minutes after posting this my SB is down to about 10-20% with the 5000 test... Trying the netstat check twice, copying the files mentioned above, didn't increase the retransmitted or reset connections... *SIGH* What is going on!!!

< 10 : 166 : 17% ########
< 20 : 710 : 74% #####################################
< 30 : 36 : 4% #
< 40 : 14 : 1%
< 50 : 14 : 1%
< 60 : 8 : 1%
< 70 : 3 : 0%
< 75 : 2 : 0%
< 80 : 6 : 1%
< 85 : 0 : 0%
< 90 : 0 : 0%
< 95 : 0 : 0%
< 100 : 0 : 0%
>=100 : 0 : 0%
max : 78.691983
min : 0.206186
avg : 13.787081

bpa
2007-09-18, 12:29
The number of resets for the wired and 1st wireless is quite high (17) indicating that the far-end system probably lost packets and had to reset the link to continue the transfer.

If the same thing happened from server to SB (SB as the receiver) - each reset would result in either a time delay or possibly loss of data - both of which could sound as a drop-out.

It is possible that it is normal for the protocol being used by Windows to have a number of resets since it occurred both wired and wireless but your 2nd wireless run has no resets.

The evidence is not conclusive but I think you should examine the cable between server and router and server network interface. Check if there is a upgrade to the driver for server ethernet interface.

NeilF
2007-09-18, 12:36
The number of resets for the wired and 1st wireless is quite high (17) indicating that the far-end system probably lost packets and had to reset the link to continue the transfer.

If the same thing happened from server to SB (SB as the receiver) - each reset would result in either a time delay or possibly loss of data - both of which could sound as a drop-out.

It is possible that it is normal for the protocol being used by Windows to have a number of resets since it occurred both wired and wireless but your 2nd wireless run has no resets.

The evidence is not conclusive but I think you should examine the cable between server and router and server network interface. Check if there is a upgrade to the driver for server ethernet interface.

If it's a driver or a cable, why has it been fine during the day, and is now at 10% (on the 5000) test :/

Compare this to my figures/graph from earlier...

< 10 : 0 : 0%
< 20 : 9 : 18% #########
< 30 : 35 : 70% ###################################
< 40 : 4 : 8% ####
< 50 : 2 : 4% ##
< 60 : 0 : 0%
< 70 : 0 : 0%
< 75 : 0 : 0%
< 80 : 0 : 0%
< 85 : 0 : 0%
< 90 : 0 : 0%
< 95 : 0 : 0%
< 100 : 0 : 0%
>=100 : 0 : 0%
max : 47.540984
min : 18.442623
avg : 24.435732

Here's the one from earlier...

< 10 : 0 : 0%
< 20 : 0 : 0%
< 30 : 0 : 0%
< 40 : 0 : 0%
< 50 : 0 : 0%
< 60 : 1 : 33% ################
< 70 : 2 : 67% #################################
< 75 : 0 : 0%
< 80 : 0 : 0%
< 85 : 0 : 0%
< 90 : 0 : 0%
< 95 : 0 : 0%
< 100 : 0 : 0%
>=100 : 0 : 0%
max : 66.666667
min : 57.000000
avg : 61.760943

bpa
2007-09-18, 13:47
If it's a driver or a cable, why has it been fine during the day, and is now at 10% (on the 5000) test :/


The same question can be asked of the wireless or router - why does it vary ?

Re the PC - is there is something run at night in the background (e.g. virus check, games) on your PC that may lock out the CPU/bus and hence affect the Ethernet driver.

I am not suggesting that the PC/cable/Ethernet driver is definitely the problem - all I am saying is do not exclude the possibility that some of the problem may lie on the server side.

NeilF
2007-09-18, 14:08
The same question can be asked of the wireless or router - why does it vary ?

Re the PC - is there is something run at night in the background (e.g. virus check, games) on your PC that may lock out the CPU/bus and hence affect the Ethernet driver.

I am not suggesting that the PC/cable/Ethernet driver is definitely the problem - all I am saying is do not exclude the possibility that some of the problem may lie on the server side.

This evening seemed somewhat like last night again. A dip in performance for a couple of hours...

What I noticed in netstumbler was another network...

Here are the figures from NetStumbler for the two channels 1 & 6 (6 being mine):-
1 - SNR=14 , Signal+=-68 , SNR+=32, Signal=-83
6 - SNR=39 to 41(flucating every second or so) , Signal+=-32 , SNR+=68, Signal -56 to -61 (flucating every second or so)

Is it odd for my SNR and Signal to be skipping up and down constantly?

bpa
2007-09-18, 14:25
If time-of-day is a factor, then you should consider what is used a few hours in the evening that could cause Wifi interference.

For example, - the ballasts of fluorescent lights and motion detectors (e.g. burglar alarm sensors, outside security lights) have been noted as sources of Wifi interference in a paper by Cognio.

NeilF
2007-09-18, 14:59
If time-of-day is a factor, then you should consider what is used a few hours in the evening that could cause Wifi interference.

For example, - the ballasts of fluorescent lights and motion detectors (e.g. burglar alarm sensors, outside security lights) have been noted as sources of Wifi interference in a paper by Cognio.

Been thru that a number of times in my head, and I've even gone around turning things off/on (eg: heating) trying to deduce that.

As I type this my performance isn't back up to 100% but it is far far better than it was 2hrs ago! Odd!