10.6.5 and wifi issues

Hi, I've held off posting until I had systematically acquired data but I think I'm there now. My system is a 2007 White 13" Macbook with the Atheros 5416 chip. I'll provide my RSSI (received signal strength indicator) measurements then speculate on the cause a bit. All measurements were taken at the exact same locations +/- 1 foot.

Location 10.6.4 10.6.5
Next to access point -34 -78
One room away -42 -85
Two rooms away -52 -90
Basement -63 -92

I didn't take as systematic a set of measurements for 10.6.6 but it looked to be in the same ballpark as 10.6.5 +/- 3dB which makes sense since effective performance was equally bad. The only difference I saw in 10.6.6 was that the # of segments in the wifi indicator on the menubar stayed at 4 even when the signal sucked, until the connection dropped at which point it went to 0. 10.6.5 was more honest in its portrayal of signal strength, showing 2 segments at the same conditions. Let's be charitable and assume this wasn't an attempt to whitewash the problem.

I'm guessing that the RSSI measurement is in dBm since it matches the dBm signal measurement in System Profiler even though I think standards only require RSSI to be a relative indicator and not an absolute measurement like dBm. Since dBm is a logarithmic scale, a drop of 30 dB indicates a received signal strength in 10.6.5 that is 1/1000 as strong as that in 10.6.4. That's a darned impressive effect to achieve purely with software 🙂

The way wireless systems work (from my own professional experience and consultation with a former co-worker who was one of my go-to experts) is that there's a Low Noise Amplifier (LNA) or similar element at the analog front-end of the radio chain that has a firmware-controllable gain. This gain should be set to provide an optimal signal to the transceiver and A/D conversion blocks so that the digital signal they feed to the downstream digital baseband processor is optimally scaled, i.e. so that for example the highest received voltage seen on the analog side maps to a parallel digital signal that's mostly 1's. If this gain is set too low, the digital stage rapidly loses the ability to process signals as you move away from the access point since weak signals look like a flat digital 0. If it's set too high, you saturate the digital logic when you're close to the access point resulting in failure to demodulate signals stronger than X dBm.

What I think is going on is that there's a problem with the latest Atheros firmware setting the gain on the analog front end too low so that the receive logic perceives the same signal as much weaker going from one firmware version to another. Another reason I suspect this is that when I upgraded bootcamp drivers a year ago, the windows side of my laptop started exhibiting the same behavior but since I only boot windows to run a non-networked app, I ignored it. A few days ago, I downgraded the wifi driver in bootcamp to the one on my original bootcamp drivers disk and performance increased drastically. My suspicion is that the firmware is packaged with the driver so that it is reloaded whenever that driver initially loads. That I believe is why downgrading to 10.6.4 from either 10.6.5 or 10.6.6 has restored performance for me. A little googling has also indicated to me that Atheros uses (or used in the past) a different way of reporting RSSI from their firmware as other chip vendors. It's possible that there are some mismatched assumptions between the driver and firmware or firmware and hardware resulting in incorrect setting of the front-end gain for a given set of channel conditions.

I've reported my measurements and speculation to Apple through a bug report. For completeness, I tried every other suggestion I found on the web and this thread with no luck. This has eaten up days but I was curious and figured I could learn something in the process. At this point, I don't expect rapid response from Apple because I think this weak-signal problem only affects people using computers from a particular manufacturing stretch when Atheros was the wifi supplier of choice. I suspect that there are many other issues affecting a broader subset of the population that will get the attention.

Still, Apple if you are listening and care, I can't use the App Store until I get a working 10.6.x and I won't give this Macbook to my wife as a kitchen computer and get a new Air/Pro as I had planned until I can give her a machine capable of running the latest 10.6 for security and reliability purposes. I don't expect the machine to support Lion though I think it should. Either way, my purchasing decisions are now based on last possible moment rather than opportunistic, not that it will affect their bottom line in the least.

Whining aside, I hope this information/approach helps some of you to be able to distinguish OS issues like forgetting your authentication credentials from system issues like decreased WiFi radio functionality.

-c8

Now we're talkin'. This is the first answer that actually makes sense to me. Certainly jives with the behavior I've experienced. Makes more sense that some have experienced more/worse troubles than others if their wifi chip are from a different vendor. Let's get a survey of WiFi chip brand from users out there. That could help. Crowdsource the dang problem.

To confirm Atheros chip here too:

Card Type: AirPort Extreme (0x168C, 0x86)
Firmware Version: Atheros 5424: 2.1.14.5

10.6.6 improved things marginally for me but nothing like it used to be. Laptop killa....

-ashley

In an earlier post I speculated that my wifi problems stem from a compatibility bug with the 10.6.5 update and some (but not all) Mac/Airport Card configurations. While my MacBook Pro has had major connection issues since the install and restart, my GFs MacBook has had no issues at all after the update, even though we share the same wireless network.

chromatic8, a BIG thx to you for confirming this with your hard work and research and finally giving me something to bring to Apple (Over the past 2 months, Apple keeps giving me the run-around saying that my wifi problem is purely a router issue and has nothing to do with my wifi card that is working correctly, despite not being able to connect to any network that is farther than 15 feet away). Your insights are incredibly valuable and greatly appreciated.

A Shepherd, a great idea... here is my Airport Card info to add to your informal poll:

*Software Versions:*
*Menu Extra: 6.2.1 (621.1)*
*configd plug-in: 6.2.3 (623.2)*
*System Profiler: 6.0 (600.9)*
*Network Preference: 6.2.1 (621.1)*
*AirPort Utility: 5.5.2 (552.11)*
*IO80211 Family: 3.1.2 (312)*
Interfaces:
en1:
*Card Type: AirPort Extreme (0x168C, 0x86)*
*Firmware Version: Atheros 5424: 2.1.14.5*
*Locale: FCC*
*Country Code: US*
*Supported PHY Modes: 802.11 a/b/g*
*Supported Channels: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 36, 40, 44, 48, 52, 56, 60, 64, 149, 153, 157, 161, 165*
*Status: Off*

What is your GF's airport chipset type and firmware? That would also clue us in to the issue.

Everyone, tell us what computers ARE WORKING and what the firmware and chipset details are so we can see if there is a definite pattern here. Ahhh, trouble-shooing at its finest.

For example, my iMac has had no issues and....... un huh....... Broadcom chipset. BCM43xx

We might be on to something here.

thanks,
ashley

I have the problem and I have:

*Card Type: AirPort Extreme (0x14E4, 0x93)*
*Firmware Version: Broadcom BCM43xx 1.0*

Ashley,

My GFs Macbook specs for you. As I mentioned, unlike my MacBook Pro, her wireless connection has worked flawlessly with many different routers before and after the 10.6.5 update:

*Software Versions:*
*Menu Extra: 6.2.1 (621.1)*
*configd plug-in: 6.2.3 (623.1)*
*System Profiler: 6.0 (600.9)*
*Network Preference: 6.2.1 (621.1)*
*AirPort Utility: 5.5.2 (552.11)*
*IO80211 Family: 3.1.2 (312)*
Interfaces:
en1:
*Card Type: AirPort Extreme (0x14E4, 0x8D)*
*Firmware Version: Broadcom BCM43xx 1.0 (5.10.131.36.1)*
*Locale: FCC*
*Country Code: US*
*Supported PHY Modes: 802.11 a/b/g/n*
*Supported Channels: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 36, 40, 44, 48, 52, 56, 60, 64, 100, 104, 108, 112, 116, 120, 124, 128, 132, 136, 140, 149, 153, 157, 161, 165*
*Wake On Wireless: Supported*
*Status: Off*

...interesting.

pookeyblow wrote:
I have the problem and I have:

*Card Type: AirPort Extreme (0x14E4, 0x93)*
*Firmware Version: Broadcom BCM43xx 1.0*

Hi pookeyblow, are you specifically having greatly reduced RSSI in 10.6.5 compared to 10.6.4 or are you having general connectivity issues? You can check your RSSI by holding down the option key I believe while clicking on the wi-fi symbol in your start menu top right or by running system profiler and finding the signal strength value under the Network->Airport->Signal/Noise section, first dBm value. Can you please post your observed RSSI value when you're in the same room as your access point? Thanks - c8

Chromatic8

The network I am having problems with - has RSSI value of -36.
My Mini is 43cm from the router's arial.

Other networks with much weaker strength - neighbours and others in my house are much weaker.
I am connected to my Airpot Express network with RSSI = -69.

The network I am having problem with is the strongest in the vicinity, but will not register
in System Profiler/Airport or respond reliably in iStumbler = it is the only network with green bars in iStumbler for example.

Other networks with RSSI of -91 are clearly reported.

*Card Type: AirPort Extreme (0x168C, 0x86)*
*Firmware Version: Atheros 5424: 2.1.14.5*

Home network:
PHY Mode: 802.11g
Signal / Noise: -37 dBm / -96 dBm

In contrast my iMac that has no problems and is connected to my Home Network has this config
*Card Type: AirPort Extreme (0x14E4, 0x8E)*
*Firmware Version: Broadcom BCM43xx 1.0 (5.10.131.36.1)*

Home network:
PHY Mode: 802.11g
Signal / Noise: -76 dBm / -93 dBm

Message was edited by: MrARH

Message was edited by: MrARH

Interesting. -36 dBm is a hot signal so your connection issues are probably due to a different problem, especially if you can connect to a -69 dBm signal from one access point but not to a -36 dBm signal from another. I don't know your level of expertise so don't want to insult you with obvious suggestions but this feels like a misconfiguration between laptop & network. Hope you're able to chase it down.

OK, learned that new trick to see the signal to noise stats, perfect. Numbers I can understand as I'm a recording engineer. Signal to noise ratios are right up my alley. So being at two locations today, I gathered some stats:

studio:
channel 7, RSSI -78
trans rate 24
Airport Express V6.3
was not working in 10.6.5 but works in 6.6

home:
CH 11 RSSI signal -83 noise -96
trans rate 2-5
linksys WRT150N Firmware 1.01.9
did not work in 10.6.5, does somewhat in 6.6
cannot move more than 15' away from router which is under wood floor beneath me.

Card Type: AirPort Extreme (0x168C, 0x86)
Firmware Version: Atheros 5424: 2.1.14.5

chromatic8 wrote:
Hi pookeyblow, are you specifically having greatly reduced RSSI in 10.6.5 compared to 10.6.4 or are you having general connectivity issues? You can check your RSSI by holding down the option key I believe while clicking on the wi-fi symbol in your start menu top right or by running system profiler and finding the signal strength value under the Network->Airport->Signal/Noise section, first dBm value. Can you please post your observed RSSI value when you're in the same room as your access point? Thanks - c8

It says:
*PHY Mode: 802.11g*
*BSSID: 0:1d:7e:fb:7f:15*
*Channel: 11*
*Network Type: Infrastructure*
*Security: WPA2 Personal*
*Signal / Noise: -25 dBm / -92 dBm*

I don't think I ever had 10.6.4. Got my MacBook early this month. Worked great then suddenly when I did some updates, the problems started. Got the "Airport does not have an ip address and cannot connect..", "Airport has a self assigned IP and cannot connect..", "Connection timeout" and so on.. sometimes it would work for a very limited time, but really slow compared to the ethernet. I upgraded from 10.6.5. to 10.6.6. but it didn't help. One occasion I even had the exact same problems with the ethernet, but it works fine now.

I have a MacBook, a PC, an iPod, an iPhone and they all work fine. The problem only occurs on my 17" MacBook Pro.

chromatic8 wrote:
Interesting. -36 dBm is a hot signal so your connection issues are probably due to a different problem, especially if you can connect to a -69 dBm signal from one access point but not to a -36 dBm signal from another. I don't know your level of expertise so don't want to insult you with obvious suggestions but this feels like a misconfiguration between laptop & network. Hope you're able to chase it down.

Insult away - I am pretty technical.
See earlier posts that show all remedies I tried and failed.

But can you believe it has just connected after weeks of not doing so - at that hot signal strength.
It did this before xmas week - lasted a couple of hours - but then stopped for another week.
I have made no changes over the last few days ... i have given up and am waiting for the Apple fix.

C8 keep up the good work...I am sure you are on to something.

This morning after a few changes to Imacs & resetting my router/modem, I switched 2 x iMacs and a Netbook on at the same time (within seconds of each other). At about an hour, they started to disappear. Then I noted something in my modem/router called GROUP KEY UPDATE INTERVAL that was set at 3600 secs (1 hour). Found this page. http://forums.whirlpool.net.au/archive/1089217. Set my key to 0 and my network has now been stable for over 1.5 hours. Not holding my breath, but live in hope.

By the way, my two iMac cards are:

Card Type: AirPort Extreme (0x168C, 0x8F)
Firmware Version: Atheros 9280: 2.1.14.5
Locale: APAC
Country Code: AU

I've been monitoring this thread for 2 days, trying to find an answer to the WIFI issue after upgrade to 10.6.6.

I own 1 MBP Core 2 Duo, 1 MBP Core duo 2.0, 2 POWERBOOKS G4 1.67 -
I upgraded the 2 MBP to 10.6 and updated both of them to 10.6.6 - while the MBP Core 2 Duo didn't show any sign of erratic behavior, the Core Duo was dropping the WIFI randomly and "looking for Network " every 12 seconds - So I reverted the upgrade of the Core Duo to Leopard 10.4 and all symptoms of WIFI dropping disappeared and I even noticed that my connection is faster than before. I've been online for at least 2 hrs and haven't had any problem -

To make the story short There is definitely an issue with OSX SNOW LEOPARD -
I am running on Motorola 68329 ROUTER / WIFI with AIRPORT EXPRESS as WIFI Extender.

Network stable for 4 hours now. Looking back at all the notes I’ve been taking with the WiFi issue, I’m reasonably confident that the group update key is the issue.

A few days ago I switched just one wireless iMac on at 4:50pm then checked at 5:00 and every 15 minutes later. At 6:00pm, my connection was gone. The next day repeated same and after about 50 minutes switched on the second wireless iMac which then lost connection after 10 minutes. This indicates that the update interval starts counting when the first wireless connects. Hence the random dropouts within first hour. Also appears that my Dell NetBook was suffering same problem but was hidden by my obsession with the 2 x iMacs. The laptop and PC on my network worked fine. The culprit would appear to be the firmware in the wireless cards (some, but not all). Cannot believe Apple have let this go on for so long, assuming that the firmware is the problem.

Footnote: Suspect that the Apple Call Centre (located in Brisbane) has become of victim of the floods. A call I was promised at 9:00 am this morning didn’t happen.