WiFi Won't Stay Connected or Randomly Becomes "Limited"

[GoodBytes]

My laptop wireless (Intel 51000) didn't work with my university new Cisco router Wireless N when they did the switch. Worked at home (Linksys router wireless N), worked anywhere else. Coffee shop, friends house (wireless N), home, not at school. And no one else had any problem at school. Just me. Drivers update "solved" the problem, with intermittent connection after trying to connect for hours and always getting limited.

Called Dell, and they shipped me a NEW revision of the wireless card that apparently just came out, despite having my laptop made to order directly from them, with the wireless card being back ordered (as I wanted wireless N, and not G option).
New wireless card in, and since then everything worked perfectly at school and continues to work... All fine. Then I switch a computer we have form Vista to Win7 as it was time, so I had to use new drivers. This resulted in not working, and the fix: Update the firmware router. I did that. Now my laptop wireless which worked before, and worked now everywhere... was wonky. Called Dell as it was under warranty, got a tech (as I took Dell's in home tech (came for 3 years with the system as it's a business class system, flag ship product), the tech tried the new and old card. New -> same problem. Old funny enough worked.. but I knew it would not worked at university. Didn't take teh chance.

Replaced my router with another Linksys (new model, and well better - 'cause why not). And now everything worked.

When I say something, I have some level of knowledge and/or experience to back it up. Else I mention it.

So instead of being offensive. Take suggestion in consideration. Even a senior engineer at Linksys, and used to worked at Belkin, D-Link, and Netgear, and you live and breath routers, and wireless communication, and assuming that, let's just assume, because I doubt it's true, that you are right and that the above makes no sense, he or she would consider it, and run tests. Why? Because that is how you discover problems and solve them, as an engineer. Practically any university paper (assuming it's not brand new), you can find a number of other paper contradicting it or finding faults in some theory presented.

It's life. I also got surprised on situations. For example:
I worked as IT. We have received new computers (at the time), from Lenovo, SSD, 3rd generation Core i5 CPU's (Ivy Bridge), Windows 7 64-bit, 4GB of RAM.
Nice and fast system. We upgraded one of the computer ourself with 8GB of RAM. Your cheap standard entry level RAM, with the same timings and speed as the 4GB ones, as this is what the Lenovo system supported.
The system was used for office work, running Microsoft Office 2010. So clearly we are not reaching anything remotely close to filling up the RAM. We put 8GB (because we wanted to test if 8GB would work, because we were gonna have a spare soon (someone was using it), to give it to someone would need more than 4GB (not general office work). Anyway, the results is that we saw a performance increase in Windows boot time and is more responsive. Went back to 4GB, and it's slower. Why?
Makes no sense. We tried it on other computer.. obviously no difference. You can see from the Surface Pro 2 4Gb and 8GB, no difference. Why this computer goes faster with 8GB? Scratch your head on that one.

My guess is the switch between using 2 RAM sticks and 4. A fault in the motherboard 'causes that a performance loss with 2 stick of RAM. Maybe, just maybe, 2 stick of RAM, for some reason a bug form the system firmware or circuit design doesn't make it go dual channel, but with 4 it does.

 

[jrapdx]

My laptop wireless (Intel 51000) didn't work with my university new Cisco router Wireless N when they did the switch. Worked at home (Linksys router wireless N), worked anywhere else. Coffee shop, friends house (wireless N), home, not at school. And no one else had any problem at school. Just me. Drivers update "solved" the problem, with intermittent connection after trying to connect for hours and always getting limited.

Called Dell, and they shipped me a NEW revision of the wireless card that apparently just came out, despite having my laptop made to order directly from them, with the wireless card being back ordered (as I wanted wireless N, and not G option).
New wireless card in, and since then everything worked perfectly at school and continues to work... All fine. Then I switch a computer we have form Vista to Win7 as it was time, so I had to use new drivers. This resulted in not working, and the fix: Update the firmware router. I did that. Now my laptop wireless which worked before, and worked now everywhere... was wonky. Called Dell as it was under warranty, got a tech (as I took Dell's in home tech (came for 3 years with the system as it's a business class system, flag ship product), the tech tried the new and old card. New -> same problem. Old funny enough worked.. but I knew it would not worked at university. Didn't take teh chance.

Replaced my router with another Linksys (new model, and well better - 'cause why not). And now everything worked.

Of course, incompatibilities happen, but that does it mean we should shrug it off as "no big deal"? Isn't it at least somewhat troubling that every software tweak potentially subjects users to network "incompatibility"? How often should I be forced to buy another $200 router just because an update makes the SP2's network adapter incompatible with my router? As in the present case, how much time and energy will have to be devoted to "debugging" WiFi connections when incompatibilities arise? And come to think of it, isn't the term "incompatible" effectively equivalent to "failure to function within standards"?

And consider, I'm a fairly sophisticated user, and after a while, can understand well enough what you are talking about. The average user at best admits knowing even less than me; there must be plenty of users to whom network connections are unalloyed magic. What are the odds such "incompatibility" has greater impact on typical users than it has on you?

Considering the burden in time, money, stress, why should we take such incompatibility lightly? Shouldn't we be telling equipment vendors incompatibility is not acceptable? To quote a friend, isn't it important to "hold their feet to the fire"?

The error is not bad devices it is passivity.

When I say something, I have some level of knowledge and/or experience to back it up. Else I mention it.

So instead of being offensive. Take suggestion in consideration. Even a senior engineer at Linksys, and used to worked at Belkin, D-Link, and Netgear, and you live and breath routers, and wireless communication, and assuming that, let's just assume, because I doubt it's true, that you are right and that the above makes no sense, he or she would consider it, and run tests. Why? Because that is how you discover problems and solve them, as an engineer. Practically any university paper (assuming it's not brand new), you can find a number of other paper contradicting it or finding faults in some theory presented.

Absolutely, engineers solve problems. But the modal user, like my wife, highly intelligent as she is, in no way can "engineer" subtleties of WiFi connections. Which is near-literally what she says when she takes pity on me struggling with such problems: she knows nothing about it and can't help me solve it.

I have no doubt you know more than me, and you should, given youth and proximity to school and experts being rich knowledge resources. Believe me, I know from long experience, observation and study, glitches are inevitable, screw-ups always happen. But now I am pointing at unnecessary error, mistakes beyond the random gaussian/bayesian noise. Engineers, the good ones, that's precisely what they do, they strive to minimize unnecessary defects, bugs, and incompatibilities.

The the error rate surrounding the SP2 vs. other "devices" I've owned, is among the highest as I experience it. I an completely aware that doesn't constitute "proof" of any kind. However, when hearing many comments of that nature, it raises flags, and may signal events that could be associated with adverse outcomes. Statistically rare occurrences are difficult to evaluate and easy to dismiss or overlook exactly because they are rare. A 1/10000 defect rate may not be obvious vs. 1/5000, though 2 to 1 could easily turn out to be non-trivial. I've heard it said that good engineers can smell such small differences--that's what makes them "good".

Keeping an open mind is a challenge, how well I know. Truly I do strive to achieve it, and I commend it to you. "Whiny" SP2 owners aren't lame engineers, or engineers at all but genuinely perplexed, disappointed or irritated by their misfortune. Solutions rely on real engineers, ears to the ground, hearing the distant rumbling as a metric of user satisfaction vs. discontent. Listening is an art form, the hardest to perfect while of all the most essential to success.

It's life. I also got surprised on situations. For example:
I worked as IT. We have received new computers (at the time), from Lenovo, SSD, 3rd generation Core i5 CPU's (Ivy Bridge), Windows 7 64-bit, 4GB of RAM.
Nice and fast system. We upgraded one of the computer ourself with 8GB of RAM. Your cheap standard entry level RAM, with the same timings and speed as the 4GB ones, as this is what the Lenovo system supported.
The system was used for office work, running Microsoft Office 2010. So clearly we are not reaching anything remotely close to filling up the RAM. We put 8GB (because we wanted to test if 8GB would work, because we were gonna have a spare soon (someone was using it), to give it to someone would need more than 4GB (not general office work). Anyway, the results is that we saw a performance increase in Windows boot time and is more responsive. Went back to 4GB, and it's slower. Why?
Makes no sense. We tried it on other computer.. obviously no difference. You can see from the Surface Pro 2 4Gb and 8GB, no difference. Why this computer goes faster with 8GB? Scratch your head on that one.

My guess is the switch between using 2 RAM sticks and 4. A fault in the motherboard 'causes that a performance loss with 2 stick of RAM. Maybe, just maybe, 2 stick of RAM, for some reason a bug form the system firmware or circuit design doesn't make it go dual channel, but with 4 it does.

What you describe fits the thrust of this quotation:

Let's face it: software is crap. Feature-laden and bloated, written under
tremendous time-pressure, often by incapable coders, using dangerous
languages and inadequate tools, trying to connect to heaps of broken or
obsolete protocols, implemented equally insufficiently, running on
unpredictable hardware -- we are all more than used to brokenness.
--Felix Winkelmann

And aren't we all burdened by it and tired of it? Not limited to software,
in any case, we will have more wholeness iff we insist on it and work to
produce it.

 

[be77solo]

Well, for what it's worth, I too had my wifi drop out along with my Bluetooth at the same time yesterday while using a Surface Arc Touch Bluetooth mouse.... a reboot fixed WiFi, but had to remove and re pair the mouse to get it recognized after reboot again. Has worked ok since, but this is after the March updates, and didn't happen before. I too fully believe they still have serious gremlins in this Marvell setup.

This occurred while sitting stationary 7' away from my Linksys EA4500 on 5.0ghz that is rock solid with all other devices.

 

[cali]

Members are reminded to address each other with courtesy in compliance with site rules for posting.

 

[jrapdx]

Here's the report of tests I made concerning my SP2/router problems.

I placed a new dual-band, "high-end" WiFi router in a location near the computer, setting up WiFi on the 5 GHz signal, and tracked performance over an approximate 40 hour period.

As expected, since the router was literally only an arm's length away, signal strength was excellent, . However, the internet connection remained intermittent. The pattern varied from less than a minute on or off (at 03:00-04:00H), to connections lasting up to 20 minutes before lapsing for 2 or 3 minutes (18:00-22:00). Description of the fluctuation is hazy because the need at times to manage settings manually made it very difficult for me to accurately measure such highly variable and unpredictable intervals.

Though the internet connection was unreliable, I noticed that the connection between SP2 and router was continuous even when the WiFi status of the SP2 became "Limited". IOW the problem did not reflect connectivity between the computer and router, but between router and the greater network. Somewhere, from the cable modem and beyond, the connection was frequently breaking.

To test this idea, a second device was connected by WiFi with the router (on a 2.4 GHz channel) and showed connection/disconnection in sync with the SP2. Similarly, a third (WiFi disabled) device was ethernet-linked to the router's integrated switch and also showed the same connection pattern as the SP2. It was evident that in this environment, the SP2 did not demonstrate adapter/router connection problems.

The possibility of problems in the ISP network was discussed in detail with the provider's technical support. At their recommendation, trying direct ethernet computer-to-modem connection revealed nothing different than found with the LAN computer, or the WiFi-linked devices above. In any case, it was reassuring when the technician agreed with my reasoning--the problem was not my equipment but somewhere in the larger network. (The ISP confirmed they'd received a number of reports of "connection problems" in this part of the city.) "Resetting" the cable modem restored connection for >3 hours, but of course it's too soon to tell if it's an enduring solution.

I'm not sure what conclusions to extract from this set of tests. Extending the test to a number of routers, and at varied distances, might show differences among combinations of SP2/router models. The necessity of including "controls" among variables is certainly a lesson learned once again. Having non-SP2 device performance and wired connection comparisons helped narrow down the "problem space". Now I have a better idea of solutions to pursue, and more powerfully, a clearer view of followup tests worthwhile to do.

 

[ArcticDX]

Jrapdx, I have the same router. DIR-655. My SP2 had a problem with wifi and bluetooth disconnect, usually after a few hours on idle. This would happen quite regularly, at least once a day. After the March update, it hasn't happened again. My internet connection is "relatively" slow, 15/15 Mbps, but I have the same speed on my SP2 (2.4 GHz wifi + 2 bluetooth periferals connected) as my desktop which is connected via ethernet.

 

[jrapdx]

Jrapdx, I have the same router. DIR-655. My SP2 had a problem with wifi and bluetooth disconnect, usually after a few hours on idle. This would happen quite regularly, at least once a day. After the March update, it hasn't happened again. My internet connection is "relatively" slow, 15/15 Mbps, but I have the same speed on my SP2 (2.4 GHz wifi + 2 bluetooth periferals connected) as my desktop which is connected via ethernet.

One of the main things I've learned is how much "mileage varies" in the WiFi world. The range of factors affecting the speed and stability of connections is enormous. We all know that distance between SP2 and router, RF-absorbing or blocking materials in the environment and the quality of the ISP's signal itself, all play a role in the quality of connections.

In my case, I have to consider if the onset of difficulty after the March updates was just coincidental. Could be just by chance the ISP's network became flaky at just the right time to fool me. Maybe it's human nature, once we get an idea in our thoughts it's hard to dislodge. I see a "Limited" connection and think "what's wrong with the SP2?". Maybe the SP2 is misbehaving, or maybe not. I've got to check it out. I think it's easy to mistake speculations for actual fact.

Currently I'm using a DIR-868L router, advertised as "top-of-the-line" for home use. It's supposed to be capable of 867 mpbs (5GHz) and 300 mpbs (2.4GHz). The SP2 says I'm getting 270 mbps on the 5GHz band, and 130 mbps at 2.4GHz. With BT in use, the 2.4 GHz speed is reduced to 65 mbps. Across locations, these numbers are as good as it gets, that is, they seem to be the maximum the Marvell can achieve.

(OTOH, an inexpensive USB adapter used with the SP2 gives 433 mbps at 5GHz, but curiously only 72 mbps at 2.4GHz. When I used the DIR-655 the outboard adapter connected at 150 mbps, with or without BT active. I don't know what accounts for the differences.)

The DIR-655 has worked well for 6 years and provides the same SP2 130/65 mbps at 2.4GHz. But it is a single-band router, therefore doesn't offer the same capabilities as the new router. I plan to test it in the same environment as was done with the 868L. In my "spare time" I volunteer for a non-profit group assisting with IT tasks. I'm planning to donate the DIR-655 to the org and anticipate the router will serve splendidly for years to come.

BTW, if you haven't done it already, you should update the 655's firmware. Amazingly, DLink is still supporting their old hardware. The most recent FW (1.37, I think) was released 8/2013. If the 655 works, why not keep using it?

 

[fsphill]

Microsoft has told me that updates are not reversible. The only way to reverse is to do a reset and reinstall what you want.

I, too, have had problems with limited connectivity since the new firmware update. However, I found out Friday that it had nothing to do with the update:

1. I was not able to connect at my favorite coffee shop that has excellent connectivity. Turns out their router was failing at the same time and they didn't know it.
2. I was unable to connect at work, but that was because the hotspot for the wifi service for "unofficial" computers that I use was dead and I was trying to connect to the next weak hotspot.

So I can't blame it on Microsoft or the update. Just a couple of weird coincidences. At home I have noticed no problem, but I have a top of the line Netgear dual band router that seems to work with every device without any problems. I replaced my older one a year ago when I started having connection problems with newer devices. A new router made all the difference for me.

 

[jrapdx]

Microsoft has told me that updates are not reversible. The only way to reverse is to do a reset and reinstall what you want.

I, too, have had problems with limited connectivity since the new firmware update. However, I found out Friday that it had nothing to do with the update:

1. I was not able to connect at my favorite coffee shop that has excellent connectivity. Turns out their router was failing at the same time and they didn't know it.
2. I was unable to connect at work, but that was because the hotspot for the wifi service for "unofficial" computers that I use was dead and I was trying to connect to the next weak hotspot.

So I can't blame it on Microsoft or the update. Just a couple of weird coincidences. At home I have noticed no problem, but I have a top of the line Netgear dual band router that seems to work with every device without any problems. I replaced my older one a year ago when I started having connection problems with newer devices. A new router made all the difference for me.

It's evening here, and internet connection is getting flaky again. I have my old router hooked up to the modem at a distant location, down in a corner of the basement. Even so, I can easily log on to the router/gateway despite the weak signal strength. Instructing the router to ping an internet site shows the same on and off pattern as I have from the SP2 desktop, no surprise--logically it's the same thing.

As before, there's trouble in the network "out there", wherever that's happening. I can't blame either of the routers I have here at home, both perform their role at least as well as I'd expect. Maybe the cable modem is faulty, but all it's LEDs light up and twinkle normally. We'll probably have to get the ISP repair service involved, the problem isn't going away.

So turns out the SP2 isn't the bad guy in this scenario. Like you say, it's not the fault of MS, not this time anyway, but I'm sure another opportunity to cast stones will soon come along.

 

[ecorstor]

I have to jump in here, because I've consistently had the same problem JRAPDX has reported. There is nothing worse than having a really annoying problem and all you get are non-solutions or suppositions that it's not really an issue with your unit at all, more environmental than hardware. In this case, I can definitively say the problem JRAPDX is reporting is not environment. A little background to set the scene...

a) I have both Meraki MR16 and Engenius access points in my home, all of them are dual-band. All of the AP's have the same SSID so we can roam through the house.
b) The Meraki has the ability to do "band steering" to 5GHz, which I employ, as 2.4Ghz in my area is very congested.
c) All other devices with dual-band capability obey the Meraki, and connect on 5Ghz. Not my Surface Pro 2 -- it insists on connecting to 2.4GHz on Channel 1.
d) The other devices I'm referring to are, A HP/Compaq 8510w workstation, a Dell Inspiron laptop with dual-band card, and iPad 2, an iPad 3, an iPad Air, a MacBook Pro, a MacBook Air, and 4 iPhone 5S's. Every single one of these devices obey the rules and connect to 5GHz first when on my network. The only one that refuses is the Surface Pro 2. I am able to verify this through the Meraki control panel, Client's section, which indicates which specific channel each connected device is using.

It is <not> a viable option, nor should it even be necessary in any way shape or form, to have to either separate my SSID's to accommodate the Surface, or to employ an external dual-band USB to get around these problems.

What have I done about it? My friend is an MS engineer and a fellow Surface owner, and my son also owns a SP2 bought around the same time as mine -- but his works properly. Anyway, what we have done is as follows.

1) Before I got him involved, I updated *everything* on the Surface and made sure I was on the most current software and firmware from Microsoft, in case they had identified and patched the problems I was having. No-go.
2) We wiped my SP2 and reloaded the OS from the ground up. When it was rebuilt I had high hopes, but with a half hour I had re-experienced the same 'drop' of connectivity, requiring me to either change SSID's and change back, or to turn off the wireless and turn it back on to regain connectivity (albeit on Channel 1).
3) For what it's worth, I CAN connect to 5GHz, but only if the SSID I connect to is ONLY broadcasting 5Ghz. If it allows a choice, this thing will choose 2.4Ghz every time.

What have I left out? I don't think there is anything left to try. At this point, I think I'm trying to solve a problem which is out of my reach, and really between Marvell and Microsoft. And yes, why didn't they choose someone like Intel or Broadcom for this instead of taking a risk on Marvell????

So, yesterday I documented all of the problems I've been having (I haven't even mentioned the power problems and it not starting when it's fully charged.., or the spontaneous reboots..) and brought the device down to the Microsoft store in Mountain View and made it their problem. The person who I worked with was very nice, and said it would be back in a couple weeks, probably replaced. I chose to do it in person because I have had a pretty bad time working with the MS Phone Support, which tends to be offshore most of the time, and is heavy on platitudes and faux politeness, but very short on action and empowerment. Hopefully I get back a working Surface Pro 2. I desperately want this thing to work right and for MS to just get this recipe right. These problems shouldn't be happening. I have to say, so far it's been pretty frustrating. -- EC

 

[Paullaberge]

I am having the same problem and am becoming very discouraged with this Microsoft product. Many technicians have spent hours troubleshooting router & Surface Pro 2 wifi disconnection to no avail. My problem does not appear at home but constantly disconnects at the school that it is used in. So I can use it at home but not at work.

 

[revmike]

Maybe you need to contact your IT person at work to have them ascertain why the issue occurs only there.