I don`t know. Lots of speculation here and other places as to the source of the problem. I have been through all levels-had my 1st macbook "captured", and then got 2 replacements. The last contact was from an engineer who was leading me to believe he is still stuck on blaming my home router, even though I explained that it happens everywhere including at the Apple store......had all the patches too...I am an orphan now, because there is nothing left to do except pay up to a pro upgrade.
Well this turned out to be interesting. Router is a Linksys E4200 with WPA/WPA2.
I downloaded a freebie ping program - "Super Ping" and ran it on three machines (10 pings each) with results as follows
Numbers are Min/Max/ Avg. in ms for each machine. Loss rate was 0% in all cases
2013 MBA 13"/i5/4G/256G - 1.5 / 54.5 / 36.5
6 year old MBP 2.2 GHz C2D/10.8.4 - 1.0 / 3.0 / 1.7
3 year old 27" iMac 2.8 GHz i7/10.8.4 - 0.9 / 1.4 / 1.1
Out of curiosity I also ran the ping test on my iPad2 - 2.2 / 12.6 / 7.6
So the MBA clearly had the slowest pings. Surprised me a bit as I have had virtually no WiFi problems with it except for one instance of a router that had known problems. Don't know what to make of it.
FWIW, I'm on the Broadcom BCM43xx 1.0 (126.96.36.199.35) as well (MBA 2013 11" i7/8GB/512GB). Pings range from 1ms to 300ms randomly. I'm getting poor stability with AirPlay Mirroing to an AppleTV and I'm sure this is one of the problems. AirPort Extreme AC 5GHz with few obstructions. All other wireless devices get very low pings.
Yeah, the pings could certainly be better - my old MBP is way faster. But since I have seen no functional or operational issues I'm okay for now.
As I posted earlier, I did apply the first update and now I wish I had done these tests before the update just to see if anything changed (retired engineer, I like this kind of stuff).
Will have to see what 10.8.5 and then latger on Mavericks brings.
I played around with some numbers, did a "back of the envelope" to take a SWAG at how widespread the problem might be. Without going into the gory details, a guess at the number of new Haswell machines sold since June might be on the order of 800,000 MBAs. There have been around 1200 postings in this thread, and if we guess - agan - that one third of them are troulbe machines and the rest are various comments, etc., then this gives us 400 "bad" machines.
So if all that is reasonable, then this thread represents something like 400/800,000 = 0.05% of all the MBAs sold.
And how about the 99.95% other machines? How many folks with bad machines haven't posted anything? Don't know on that - just playing with numbers.
I think only Apple really knows the number of bad machines out there.I got 3 in a row, and who really knows how many are silently suffering through, not being forum posters, and holding out hope that Apple will come through with a fix. Still others are stuck in the "blame your router or ISP" mindset and wasting time working through that. I`ll bet the numbers are alot higher, though we will probably never know. I like this machine and am still hoping there will be a fix. I waited years to buy one, figuring the bugs would all be worked out by now. All of my other Apple stuff is bulletproof and has been ultra reliable for years.
Could it be that the update updated the firmware from .22 to .35? Only problem machines would have .35.
Leo did you apply the update and to which machines?
Can more people please tell us if they have problems and what the wifi firmware version is? We need firmware versions for good working machines as well as bad.
Also did they install the update?
For the average user, sorry if I am going geeky here...
Russets, you are correct... my wife's machine (the base one with firmware .22) never had the update applied. I played around with the update, by doing a time machine backup before and after, and can make the pings go from consistant (~3ms, before the update) to inconsistent (1-100ms, after the update). Yes, the update is what turns the firmware from .22 to .35.
Ultimately, I realized that the update in fact changes the AirPortBrcm4360 kext (Kernel Extension, like a driver) for the Broadcom 4360 chip, responsible for the Wi-Fi and BT, which is the chipset shown in the iFixit teardown of the mid-13 MBA.
You can find the kerts in the /System/Library/Extension folder. This particular kext is inside the generic, IO80211Family kext, which has the overall Wi-Fi drivers.
In my particular case, the update did helped me in eliminate the drops when I connected to a particular Wi-Fi network (small branch office of the company I work for that I had to visit a couple of times last month). That is the only place I noticed the drops, where I had to either turn off the Wi-Fi or run the diagnostics to get the connection back on, but the drops happened every 5-20 mins... extremely annoying when trying to get work done.
Back then, when I called Apple tech support, I couldn't afford wasting my time with the usual crap: "it's the router, reset config, reset the machine"... None of them collected any log. One of them (supposedly an Wi-Fi escalation engineer) didn't even really know what an IP address and Subnet Mask really means. Being an Computer Engineer with about 15 years of experience in networking technologies, I started to wonder what level of engineers is Apple hiring for their "escalation" staff...
Looks like we are going deep here and doing Apple's job in troubleshooting this...
Yes, I have been through all levels of support at Apple including "engineering" and wonder just how qualified some of these people are.No one has any more ideas to try other than my buying a pro.
Someone else posted on another forum that the power saving archetecture of the haswell puts the wifi board to sleep after 250 ms of inactivity.....I guess to get that 12 hour battery life. I wonder if that sleep mode can be turned off even if it means less battery?
I am one of those who are currently doing well with a June (early) MBA, which is an upgraded model with the higher processor, 8 ram and 250 GB harddrive. I was told the machine was put together in China (no big surprise). I have very good battery performance (albeit the battery meter is not reliable for the first hour or so and rockets up and down, depending upon the software being used at the time, or other factors).
I have not installed the 1.0 patch. My WIFI has the following original firmware (not ~35):
Card Type: AirPort Extreme (0x14E4, 0x117) Firmware Version: Broadcom BCM43xx 1.0 (188.8.131.52.22)
I have posted before. I have used many different routers (about a dozen) during travel. I have never had the opportunity to try AC OR an Apple router. This particular MBA has been cooperative, connecting to everything fast and not ever disconnecting, even over several days when I was using a 'busy' connection in an urban apartment.
Websites load fast enough not to worry me, and downloads are extremely fast. Compared to my still functioning 2006 MBP with a two year old Airport card, the MBA's wifi radio has a longer range (by far). It is impressive. For now, I have an ancient "original" router (Linksys G); I fully realize that this is not going to boost my performance!
In case it is helpful, I called to ask Apple about a week ago to ask about the recent 1.0 (recommended) patch for the new (2013) MBAs. The rep, who may well have been new on the job, was not familiar with WIFI patches for the MBA and said I was his first call on WIFI and this particular patch. I ended up telling him a little about it, and he looked it up to read the details. Together we decided NOT to install the patch, given my lack of manifest problems. Let sleeping dogs lie?
I hope these fragments of information will help a little.
It's pure speculation on my part, but I would venture a guess that the ping timing issues in the newer firmware are a by-product of the fix that corrects dropped connections on some mid-2013 MBAs. Could be something like delays added to ensure that the connection doesn't timeout.
As posted here a few weeks ago, I had a loaded 13" MBA that froze a few seconds after connecting via WiFi. Apple replaced it, and the new machine has worked flawlessly with every router I've tried (even during a week-long trip to the NYC area.) I installed the 1.0 update and didn't notice any issues with Wifi thereafter.
However, I just discovered that my MBA has the random fluctuations in ping times to/from my local router that have been reported here. Ping times are all over the place, from 1-300ms. I tested the same ping from my Windows 7 system and they're all 1ms (probably under.) I also tried a couple of Wifi dongles on the MBA that I picked up back when I had the faulty MBA). They don't have the problem, either, though one averages about 1.6ms and the other averages about 2.7 ms.
So, does anyone know how to revert to the previous version of the Wifi firmware? I probably have it somewhere on my Time Capsule, but don't have the slightest idea how to find or install it. I'd rather not revert my entire machine to its pre-update state just for a test.
It gets curiouser and curiouser.
I reported few posts ago about my ping data with the well-performing MBA and the updated .35 firmware patch.
Results were along these lines (min / max / avg ping in ms)
2013 MBA 13"/i5/4G/256G - 1.5 / 54.5 / 36.5
As it turns out I have a backup clone of my MBA system that still has the .22 WiFi firmware before the patch. I booted the MBA from the backup via USB3 and reran the ping test. Results now are typically like:
1.1 / 3.0 / 2.7
So my conclusion is that the .35 patch mangled my system big time. I need to get rid of it.
So how do I get the .22 firmware back on my system? Options are to reinstall 10.8.4 or to wipe the drive and reload from the clone backup? Any easier way to do it?