WiFi mysteries resolved once and for all — Power saving is clearly the culprit!

[PowerGod]

I have 5 networks around, I had a lot of problems when my router was using the same channel of another network, but after changing it all became better.

I also remember that the Pandora had problems with my router if it was using channels over the 11th, for example with channel 13 it wasn't seeing the access point at all.

 

[klapse]

Of those, the MeeGo version seemed to work best (relatively spoken) - iperf test over 10 minutes averaged about 2.7Mbit/s instead of 1.8Mbit/s with the Pandora firmware.

I gave it a try on a rebirth pandora, putting it in /lib/firmware, ad-hoc wlan to N900 hotspot. Got same speeds as before with PNDmanager - 60-90 kB/s. Downloads on N900 ranged 130-150kB/s.

 

[porg]

As mentioned, changing channels had no effect.


What I'll try is a test in a rural area with the Pandora being the only WiFi device connecting to a router with a wireled webserver.

 

[ShleeDragon]

As mentioned, changing channels had no effect.


What I'll try is a test in a rural area with the Pandora being the only WiFi device connecting to a router with a wireled webserver.

there is a awesome android app that scans the network and tells you which channels are currently being flooded..

cool fact: Each channel covers at least three channels.


1-3 4-6 7-9 10-12 so you need to find the range that isn't used. Slipping into the middle normally gets you interferance from both sides XD

 

[Galaxis]

No useful conclusions here. I did a couple of test runs with different firmware versions and various work loads, but there's not really anything that sticks out as a particular problem. During continuous transmission, the current firmware sometimes manages to stall traffic for several seconds though, which doesn't seem to happen so drastically with the older wireless firmware. In the end there's not much difference in overall transmission rate between the versions I used.

One thing I noticed was that the bit rate shown by an iwconfig wlan0 drops to 1 Mb/s seemingly in conjunction with the the Invalid misc and Tx excessive retries counters going up. While transmitting data, the wireless bit rate stays at 1 Mb/s, and it's usually slowly fluctuating upwards (2M, 5.5M, 9M, 11M) when nothing is happening on the interface. But that, too, may just have happened by chance while I was looking at it.

Unfortunately, setting a fixed bit rate with iwconfig doesn't seem to work.

 

[mosschops]

there is a awesome android app that scans the network and tells you which channels are currently being flooded..

any chance of a name for this awesome app?

 

[TrashyMG]

there is a awesome android app that scans the network and tells you which channels are currently being flooded..

any chance of a name for this awesome app?

I have this on my phone.

https://play.google.com/store/apps/details?id=com.farproc.wifi.analyzer


There is a view that shows the channel spread of all the networks around you and also one that shows specific APs and what channel they're on.

 

[PowerGod]

I just noted this, while copying a big file from the PC to a samba share on the pandora, the speed was stable at 614 KB/s with the lid open, but then I closed it and suddenly reached 768 KB/s stable... so I tried to open and close it many times and it worked every time the same way...

 

[Franko]

I just noted this, while copying a big file from the PC to a samba share on the pandora, the speed was stable at 614 KB/s with the lid open, but then I closed it and suddenly reached 768 KB/s stable... so I tried to open and close it many times and it worked every time the same way...

Wow.  That is interesting...like power usage (lcd backlighting in this case?) is DIRECTLY effecting wifi speeds and throttles data speeds accordingly?  Good to know.  I guess since turning "power saving mode" off also seems to have a (positive) effect on wifi data speeds this sort of makes sense...   :blink:

 

[Franko]

Ooooh, and I just thought of another possible cause here.  Perhaps the lcd screen being on does indeed interfere with the data rate and the wifi antenna's effectiveness?  Maybe...

...and sorry for the double-post.  Was just thinking ou loud here...

 

[Yoyobuae]

Maybe test the effect of turning the display off without closing the lid:

• Turn down screen brightness until it goes black
• OR from a terminal run:

  sudo /usr/pandora/scripts/op_power.sh 1 lid

  To simulate lid close effect

 

[PowerGod]

Maybe test the effect of turning the display off without closing the lid:

• Turn down screen brightness until it goes black
• OR from a terminal run:sudo /usr/pandora/scripts/op_power.sh 1 lidTo simulate lid close effect

 Just to point out, I was using since the beginning "power off".

Reducing brightness changed nothing, but using the script did the trick !!

EDIT: I had a look at that script and found the command that made it

sudo sh -c "echo 1 > /sys/class/graphics/fb0/blank"
 if I launch this, the speed increases almost immediately

 

[WizardStan]

That line shuts off the LCD. Not just the backlight, the entire screen is shut right off. That's kind of fascinating.

 

[Franko]

Still could either be a power-related correlation OR an interference one.  Maybe someone can test the wifi data rates with varying power-intensive things switched on and off (like usb-host and bluetooth, etc).  If those things don't have a noticeable effect then we're probably looking at an interference issue with the lcd and the antenna...

Edit 1:  Although, now that I think about it, using bluetooth might even cause some signal interference.  So I'd probably drop that usage from the testing parameters.

Edit 2:  If it's an interference issue than this is really a goof-up.  I think it was Mr. Weston's initial idea to tape it to the back of the lcd screen in the first place (?) and this was presumably to enhance the signal and not hinder it...BUT as we all know now, it took the OP Team years to realize that the (magnetic?) tape was masking the signal as well.   :unsure:

 

[Neelix]

The issue wasn't the the tape itself,  It was the placement.   In some units the tape wasn't placed very carefully, and the misalignment is what caused the issues.   I was having dreadful connection drop-out issues with my current primary unit,   until I opened the lid and adjusted the tape.  (adjusted, not removed)

- Neelix

 

[WizardStan]

I can confirm this. With the lid open, transfering a 100MB file takes 9 minutes, with the lid closed it took 4. Power saving is on.

 

[Franko]

The issue wasn't the the tape itself,  It was the placement.   In some units the tape wasn't placed very carefully, and the misalignment is what caused the issues.   I was having dreadful connection drop-out issues with my current primary unit,   until I opened the lid and adjusted the tape.  (adjusted, not removed)

- Neelix

Ahh, okay.  Thanks for the clarification, Neelix.  I remember it was "tape related" somehow, but didn't exactly know the specifics.  

Edit 1:  That being said, then there presumably was an interference issue with the lcd/antenna/placement of tape aligning said antenna.  So it sounds like there is a "sweet spot" for the taping the antenna to the back of the lcd.  I can't find that old thread (a year ago?) right now regarding the tape issues and if someone had kept the lcd on and had strung the antenna outside of the case (i.e. away from the screen).  One wonders if there would be any signal strength improvement...it'd be an interesting test as well - unless I'm beating a dead horse here and someone has already tried this.

 

[Yoyobuae]

Edit 2:  If it's an interference issue than this is really a goof-up.  I think it was Mr. Weston's initial idea to tape it to the back of the lcd screen in the first place (?) and this was presumably to enhance the signal and not hinder it...BUT as we all know now, it took the OP Team years to realize that the (magnetic?) tape was masking the signal as well.   :unsure:

I kinda understand the OP team. Wireless signals can be so mysterious at times, almost like black magic.
I mean, even Apple got it wrong with iPhone 4.

 

[Franko]

Forgive my musings/ramblings again on the subject, but my old laptop had decent wifi capabilities and the antenna was strung around the lcd screen and not taped to the back of it.  Now I can totally understand why the OP Team wouldn't want to do that - the case is hard enough to assemble and snap in place without having to string an extra wire around the screen or inside of the case...

...also, don't get me wrong - I totally love my new 1GHz Pandora.  My wifi seems decent enough for my uses (I'd love better range at times), and I may eventually buy an external usb wifi stick at some point...but it'd be great to finally get to the bottom of this and see how great we can improve this already great product.  Time has afforded the OP Team experience and many lessons have already been learnt (stronger cases, improved firmware updates/fixes, etc).  I'm all for improvement and any perceived "complaining" isn't my intention.  

 

[Neelix]

The antenna, in this case, is a track on a PCB, not just a length of cable.  The tape holds the PCB in place.

- Neelix