Subscribe via RSS Feed

iPad WiFi Woes – Myth or Reality?

By | April 21, 2010 | 27 comments
Share |

Over the last couple weeks, we received a lot of great feedback from our readers about the iPad, but we also received quite a few complaints, especially when it comes to WiFi. To check whether these issues are isolated or widespread, we decided to submit 3 of our iPads to a series of tests in order to assess their wireless performance.

To keep it simple, we mainly focused on signal quality, as well as the average download speed the devices could sustain. All tests were performed using 2 different routers, to make sure the results were not tied to a specific brand/model. We tested both 2.4GHz and 5.0GHz bands, and all tests were performed in an urban area with a decent amount of interference (Dallas-Fort-Worth metroplex), around noon on April 20th, 2010. We also used a MacBook Pro and a Windows Tablet PC using WiFi chipsets from the same family as the iPad’s, for comparison purposes. The results are in.

Testbed

Devices

  • iPad (WiFi-only model): iPhone OS 3.2, 16GB Flash, WiFi via Broadcom BCM4329XKUBG (basically a BCM4322AG with Bluetooth support, driver version unknown)
  • MacBook Pro (late 2009): OS X Snow Leopard 10.6.3, 2.8GHz Core 2 Duo, 4GB RAM, WiFi via Broadcom BCM4322AG (driver version 5.10.91.27)
  • HP Tablet PC (tx2500): Windows XP SP3, 2.0GHz Turion X2, 2GB RAM, WiFi via Broadcom BCM4322AG (driver version 5.10.38.26)

Routers

  • NetGear WNDR3700, firmware 1.0.4.55NA
  • D-Link: DGL-4300, firmware 1.22NA

Tests

  • Test 1: Open field, devices 20ft away from the routers
  • Test 2: Open field, devices 40ft away from the routers
  • Test 3: Open field, devices 60ft away from the routers
  • Test 4: Routers inside a house, devices 20ft away from the router (1 wall between the routers and the devices)
  • Test 5: Routers inside a house, devices 40ft away from the router (2 walls between the routers and the devices)
  • Test 6: Routers inside a house, devices 60ft away from the router (backyard, more than 3 walls between the routers and the devices)

Measurements

  • Signal strength: Level received at the router (information provided by the DGL-4500)
  • Speed test: Mix of content download as well as raw http performance tests via speedtest.net, repeated 10 times for each test, and averaged out (normalized to a scale going from 0 to 100, 100 representing the best throughput we could measure)

Results

Signal Strength

Let’s be clear, the iPad performed decently, but performance was not on par with the MacBook and the HP Tablet PC. First things first, range. It is obvious that the iPad’s WiFi range is limited compared to the other devices, very likely due to the fact that the Wifi antenna is hidden behind a thick piece of aluminum and plastic (behind the Apple logo, on the back). And within the over-crowded 2.4GHz band, the iPad had a harder time to stand out of the crowd, compared to the Mac and the PC.

When we used the much less crowded 5.0GHz band, the iPad’s signal strength was much higher, and actually beat both PC and Mac, at least for close to medium range tests. However, in a normal environment such as a house, the results were less convincing, and the iPad eventually lost its connectivity when we went too far from the router, while the PC and the Mac managed to remain connected.

Throughput

When it comes to throughput, the iPad was on par with the Mac and the PC for the open field tests while in a close range, but as soon as we used it in a enclosed environment, the WiFi throughput performance on the 2.4GHz band was still decent, but was in most cases not higher than 2/3 of the throughput we managed to get from the Mac and the PC.

As for the 5.0GHz band, the iPad did very well in open air tests, and fared better than on the 2.4GHZ band, which confirms the high signal strength measured by the DGL-4500. On the other hand, as soon as the iPad was introduced to a more hostile environment, the results we obtained were again about 2/3 of what we could see with the Mac and the PC. Note that the last test was really tough on our iPads, as the throughput was extremely low, which is in line with the poor signal strength/quality assessed by the DGL-4500 for this specific test.

Conclusion

Overall, the iPad’s WiFi performance was ok, but clearly not in line with what we measured with the Mac or the PC.  When you add this to the fact that the iPad also has other network issues such as a buggy DHCP implementation, it feels like the iPad was a bit rushed out of the door when it comes to networking.

Hopefully these figures will get better over time as Apple and Broadcom fine tune their WiFi drivers for the iPad, but meanwhile:

  • Try to stay fairly close to your WiFi router. The farther you go, the more likely the iPad will lose its connection, even in locations where your laptop can still connect to your WiFi network. Based on our tests, trying to go beyond 2 walls between the router and the iPad is too much.
  • Switching the iPad’s WiFi on and off can help if you experience problems, especially if your iPad was in sleep mode for a long time.
  • Try to switch your router from 2.4GHz to 5.0GHz or vice-versa, to see if things get better. 2.4GHz offers longer range but is overcrowded. 5.0GHz experiences much less interference, but range can be severely limited.

Editor’s note: Your mileage may vary, depending on your location, the configuration of your house, the location of your router, as well as the amount of interference you may experience from other WiFi networks and cordless phones.

email

About dag: Certified geek

  • Chad

    Interesting tests- thanks for taking the time.

    Although the WiFi concerns some are experiencing are more than likely valid, I do not seem to be having any of the iPad WiFi frustrations being reported. Currently, I’m on an older 801.11b/g router and have wondered if most/any of the dropped DHCP and poor distance performance are due the newer 801.11n single/dual band protocol.

    Does anyone know if the IEEE 801.11n standard has been finalized, or is it still in draft?

    • http://www.padgadget.com dag

      Hi Chad – 802.11n was published on October 29, 2009, after 6 years of work (and many arguments!). I did run some tests with 802.11g as well, but the results ended up being very close to the 802.11n 2.4GHz results above, so I decided to stick with 802.11n in order to try the 5.0GHz band as well and keep the results consistent. Also note that we used WPA2-PSK with AES encoding for the tests.

      • Rich Adams

        I’ll add my thanks for taking the time to do this “reality check”. I loved using the iPad in the Apple store but am really on the fence at the moment since WiFi is at the heart of the iPad’s functionality. I’m going to wait for a fix and/or see if the WiFi/3G model’s performance is markedly better (or worse).

        You should include the fact that WPA2 security was used during the tests as that seems to make a difference (WEP Vs WPA Vs WPA2) according to numerous posts on the Apple forums and others.

        Keep up the great work!

  • Chad

    Thanks for your reply.

    Question: Are suggesting the G router configuration was the same download speeds as the N router configuration for the iPad, or in general?

    I suppose I’m trying to decide if I should invest in upgrading to a new N router or stick with my older B/G router to increase dowload sppeds to my iPad. Any thoughts on this would be greatly appreciated.

    • http://www.padgadget.com dag

      Chad – Correct, overall the download speeds were inferior for all the devices when using 802.11G (not just the iPad), but the results were very similar, the PC came out on top, and the iPad was behind. I can send you the raw data if you’re interested, let me know (offline).

  • Eric Hofstetter

    The ON/OFF method works every time for me. I am getting the less bars like a lot of others. Perhaps the next iPad will have the glass back like the recently outer iPhone 4G.

  • Pinky

    I had wretched wi-fi coverage and called Apple twice about it. After resetting the network settings twice on the first call and then on the second call, doing a restore under a different name, I get ‘ok’ performance but I still get the ‘Connecting…’ message and I’m sitting in the same room from the Apple Airport Extreme, and less than 10 feet away from it. Since finding out about the dodgy DHCP experiences at Princeton, I see where that is perhaps *some* of my problem.

    Yes, the iPad seems rushed. There were issues with accessories. I ordered mine for delivery April 3rd, and the case was scheduled for delivery the end of this month yet, I found a case on a shelf at the local Best Buy. Surprising that they could get one (many) and I would still be waiting for mine from Apple. Yes, their hardware is great and all, but the iPad has been a messy mashup of issues from day one… Now I don’t know what to do with my nearly functional iPad. Swap it, return it, sell it, smash it, ignore it…

  • Lloyd

    I to have not had any issues with WiFi on the Ipad. I connect to 2 Linksys routers, one T-Mobile/Linksys Hi-Port router at home and another Linksys at work. At home I only have one wall between the Ipad and the router. At work there are probably 4 walls, but as mentioned no problems encountered.

  • Pingback: iPad has Wi-Fi problems, some users say - MacNN Forums

  • Gerard Boehm

    Chad, thanks for all the work. Your results seem consistent with the experience I had with the iPad in terms of weak signals (especially with 5GHz) while other devices continued to work. Since I didn’t know if this would be addressed I returned my iPad for a refund- I’ll wait until this is sorted before buying again.
    One thing I noticed with my Airport extreme base station (newest model, current firmware), sitting 5 feet away was that download speeds with the iPad were consistently almost 2 times faster with the 2.4GHz band compared to the 5GHz (no difference using a MacBook Pro at the same distance). Your test 4 seems to support this finding (although I had no walls between me and the router).
    Thanks again.

  • Rich

    Add me to the list of those that have a wi-fi issue. Your test results are interesting. What I see in my home when in the vicinity of the router within 30 ft. and 1 or 2 direct walls… very good performance. Once I am a little further away my two laptops still connect and work properly. The ipad however constantly pauses and at times disconnects all together. I have even resorted to “reserving” an IP address on my home network but I have not seen any change. My router is a DIR-655 with firmware 1.33NA and I am using WPA2 for security.
    Thanks,
    RA

  • Doug Petrosky

    You stated that you tested 3 iPads and two routers. The fact that I don’t see any break down of that data suggests that you saw no real difference between each iPad and no real differences from one brand of router vs the other?

    Also, just wondering if Bluetooth was on or off and if it had any effect on connectivity.

    Finally, did iPad orientation make a difference? With that metal back facing the router you might have more issues than if the router was at your back or if the device was lying down. Just wanted to know if you noticed anything.

    • http://www.padgadget.com dag

      Doug – Correct, the 3 iPads gave very similar results (we took the average for each test). We didn’t try to play with the bluetooth settings, can’t tell you if that would have changed anything. As for the iPad’s orientation, you bring up a good point. It made a difference, especially for the tests where the iPad didn’t perform well. We plan to cover this in a later article, stay tuned!

  • http://www.roughlydrafted.com Daniel Dilger

    Your results don’t seem to reflect an understanding that most desktop computers that support 802.11n have multiple antennas, while iPad does not.

    802.11n supports the use of multiple antennas to boost signal range and to use wider bandwidth. Configured to support two bands, a MacBook Pro can support theoretical speeds of 300 mbps, where iPad tops out at 150 mbps.

    Expecting a $500 device to perform the same as a full notebook computer is not really realistic. How about comparing netbooks, if you can find one that supports 802.11n at all.

    Next time you do a shootout, you should consult someone with networking knowledge to help you set up tests that are meaningful.

    • http://www.padgadget.com dag

      Daniel,

      You do bring up some good points that need clarification, couple notes:
      - This was not a shootout, I am not sure where you got that from. The point was not to say that the MacBook or the Tablet PC are better machines, the point was to set the right expectation wrt the iPad’s WiFi capabilities vs. mobile machines we use everyday, as many of our readers ask us why their iPad doesn’t offer the same speed/range compared to their laptops.
      - Apple chose a WiFi chipset (BCM4329) primarily built for phones. Call it what you will, but the iPad is not a phone in my opinion.
      - MIMO / antenna diversity offered in the MacBook doesn’t mean double throughput. Moreover, our tests were done at rates lower than 150MBps. Finally, since you bring up a good point, I just re-ran a new batch of tests using an older 802.11g router (Linksys, non-MIMO capable), and I also played with an old Dell laptop (802.11g) to see if the iPad would look better. Results are in line, the iPad still falls short when it comes to range. I can send them to you of you’re interested.
      - About 802.11n support on netbooks, I suggest you do your own research.
      - I won’t comment on your last point, not a big fan of trolling!

    • Jake R

      Sorry for being a bit rude, but your rant is a bit too fanboy-ish for you to be taken seriously (since these guys “should consult someone with networking knowledge to help [them] set up tests that are meaningful”, well, here I am, will be more than happy to discuss my credentials with you if you’re interested).

      You’re entitled to your opinion, but I actually think they did a pretty decent job.

      1) If you look at their tests, in some cases the ipad offered a higher throughput than the macbook (e.g. tests 1/2/3 @ 5.0GHz), so your point about 300Mbps support on the macbook vs. 150Mbps-only on the ipad makes very little sense imo. These theoretical speeds are marketing gimmicks, and mimo is most of the time so poorly implemented and so prone to external factors such as interferences that real-word tests show only a marginal improvement, certainly not 2x. Plus, I seriously doubt they managed to do tests at 300Mbps using speedtest.net, so given that they must have run their tests using DSL or Cable, we can safely assume that the max speed they got from their ISP was below 30Mbps.

      2) Since you seem to be so verse about the point of using multiple antennas, can you then explain what prevented Apple from using more than one? If you look at the teardown pics, there’s a lot of room on the back of the ipad. Plus they’re about to launch a 3G model, so clearly it’s ok for them to add as many antennas as they want… as long as they charge you extra!

      3) I’m getting both better range and better throughput with my sub-$400 netbook – and yes, the netbook supports 802.11n, like many others!

      Don’t get me wrong, I love my ipad, and in many ways I think it’s vastly superior to my netbook. I just think Apple could have done a much better job in some areas, and WiFi is one of them. My 2 cents.

      • http://www.padgadget.com dag

        Jake – Correct, we used Cable for the speedtest.net tests (15 Mbps). We repeated each test at least 10 times to make sure there was no temporary bottleneck on the ISP side.

  • Chad

    Well,

    I’d like to say that this test encouraged me to replace my old 801.11g router to a new 801.11n 2.4GHz router ($59 at bestbuy)…. I’m glad I did,… my download speeds, using my iPad and measured with an online WiFi tests, are now ~10-13Mbps whereas before I measured about 3Mbps… I wasn’t expecting that much improvement, but the increased dl speeds on my iPad with the N is definitely faster. Thx again.

  • Pingback: iPad has Wi-Fi problems, some users say - MacNN Forums

  • romills

    I had an iPad networking issue, that so far seems unreported(at least I haven’t seen it).

    My iPad was working great for about 2 weeks, then it starting having wifi connectivity problems, with no change in the use. I at first thought it was a particular program, but it actually seemed to be all programs as I dug into it deeper.

    From what I can tell iPad was losing the IP assigned by the router. It would show up in the IPAD as 169.xxx.xxx.xxx. I did not record the other numbers, but if my understanding is correct, any 169 ip is an error, or a redirect to an internal source. I verified that the router was seeing the ipad, as it had an IP on the routers tables. I then tried to statically assign it on the Ipad, and this appeared to work ok, but seemed flakey at times. It seemed a little slower than standard, but that is subjective, as I can’t think of a technical reason for that.

    I used that for a couple days, and decided I really wanted to figure out if it was the router or the Ipad. I went through redoing my whole router setup and tried to get it to properly assign an ip to the ipad through DHCP. No luck. Next step was trying to change the Ipad settings. A long story short, the only thing that worked was a restore through itunes. After the restore, I could not recreate this issue (I tried). This leads me to believe that it is some sort of internal software/firmware issue. It’s been working good now for 5 days, just like when I first bought it. Love the device, but I suspect we will be seeing more updates to the network code.

    My setup
    Windows 7(64bit)
    32GB Wifi Ipad
    Uverse Wireless G router

    Any ideas?
    I am speculating that the gain/power settings on the ipad were somehow adjusted, and by restore, they were returned to normal. Low gain setting my not allow for DHCP to be properly recieved from router, and static IP allowed for some communication, but slower due to errors??

    • http://www.padgadget.com dag

      Romills – 169.xxx.xxx.xxx usually means the iPad managed to get connected to the WiFi network, but did not manage to get an IP address from the DHCP server. Have you tried a different router to see if the behavior is reproducible? Also, static or dynamic IP should not show any difference in speed.

      • romills

        I agree, it is seeing the network, it shows up on the list in the ipad when I select it. I did not change the router, but I did mess with many of the settings in the router. The router was seeing the ipad, and had assigned it an ip, it was just the ipad had internally said the IP was 169.xxx.xxx.xxx (x being some number). The router had it in the valid range for the network 192.168.xxx.xxx. I reset the router a couple times, just to make sure it was a current ip, and not one from the past. It would always recreate one for the ipad.

        I tried using different channels, going from wireless B/G to just G, I re-did my table of ip’s. I was configuring the router to reserve some ip’s for particular hardware, this was removed. When I went to static IP, I could get to the internet, but it just seemed slower than normal. I am guessing it could be some sort of retry issue, that is why I was speculating it could be a broadcast power issue. I did not really test distances, but I was having issues both right next to the router, and 20+ feet away. Since I restored the iPad using iTunes I could not recreate the issue at these distances. It is now working as good as new(3+ days). I’m just waiting for it to happen again.

    • http://amillionbetterthings.com Steve

      I’m having the same problem with U-Verse and iPad, and it spills over to our iPhones, too. I rebooted the U-Verse modem this morning, and it *seems* to have solved the problem, at least for now.

    • http://twitter.com/oisee oisee

      Looks like I figured it out why iPad is so unpredictably stubborn sometimes. And I have found a real solution (not such pseudo-solutions as “turn off autobright” or “autofill”).

      The reason of getting 169.x.x.x ip is that router’s DHCP-server is too slow for the jumpy iPad, and the server cannot give the iPad a correct ip (something like 192.168.0.x) quickly. So the jumpy iPad makes decision to get itself a 169.x.x.x.

      The reason why router can’t manage to do it is: it’s busy with other tasks.

      And that is why the full router-reset helps for a while. When the router’s RAM gets busy with junk tasks results in DHCP low performance, so it’s too slow for iPad’s DHCP client. And that’s why getting new modern AirPort Extreme helps too.

      So, the real solution is: speed-up the DHCP-server on your good old router (installing latest firmware), OR just set up another DHCP-server in your subnet.

      For example, I have a NAS (DNS-323), and i just turned on its DHCP-server and the 169-problem was just gone. It (DHCP-server) runs so fast, even for the jumpy iPad.

      The subnet config is: good old router D-Link DI-642 192.168.0.1 (with DHCP-server’s range: 192.168.0.100-199), NAS D-Link DNS-323 192.168.0.100 (with DHCP-server’s range: 192.168.0.200-254).

      So, now iPad just gets its ip from the fastest (second) DHCP server: 192.168.0.200.

      You can set up second DHCP-server on your macBook or somewhere/somewhat else.

  • Alison

    It has to be the antenna. Cover the antenna and receiver or transmitter ise suffers.

  • Bill

    Interesting tests but the WiFi problems being reported by a lot of users were the slow speeds and connectivity issues following an upgrade to IOS 4.2.1. Be interesting to see your results if you did this comparison.

    • http://Jocaetanophoto.com John C

      I agree, complaints were after the upgrade and just noriced your tests were on the older ios…unfortunately I purchased mine with iOS 4 so I can’t tell if it was slower but it def. Is slow