Test (and Fix) Your Broadband Connection

By  |  Wednesday, March 24, 2010 at 8:05 pm

There’s been a big kerfuffle since the FCC recently proposed to give broadband a goose.

If the National Broadband Plan goes as it should (and no, I’m not counting on it, either), almost everyone in the U.S. will get high-speed Internet access (the goal is 90% coverage); average speeds will increase by 20%; and everyone will get malware and viruses lots more quickly.

As it is, the U.S. is near the bottom of the broadband pile, with speeds averaging 2.5- to 10-megabits per second (Mbps); Japan, France, and Korea lead the pack at warp speeds ranging from 160 to 100-Mbps. Read BusinessWeek’s World’s Fastest Broadband; check some numbers from 2007; and if you’re a hard-core techie, dive into some OECD specs from 2009.

The dust-up happened when the FCC said most of us are paying for an “up to” speed, but not always getting nearly that advertised rate. The FCC offered to test your bandwidth speed. That’s nice, except it uses two testing services and the results at the sites vary, sometimes substantially. (Read How fast is your broadband? Even the feds can’t be sure in the L.A. Times. (Thanks for this, Barry.)

For comparison, I downloaded a 700MB file using JDownloader at 7:00 am, noon, and 5:00 pm. I clocked my Charter Communication cable connection and throughput was consistent: It stayed mostly at 350 kilobits per second (Kbps) with lots of bursts as high as 460 Kbps. In megabits per second, that’s 2.8 Mbps with bursts as high as 3.7 Mbps. (Techies: That’s based on 350 kilobits per second = 350 Kbps = 8*350 = 2800 Kbps = 2.8 Mbps.)

With any speed test, resist the temptation to check e-mail or browse the Web while testing so as not to alter the results.

If any of the online tests I’ve listed this week won’t work, you might need the most current version of Flash Player and Java. (Start on the Java site, click the “Free Java Download link,” and then click the “Download Java Now” button for Windows. And be careful: During the installation, make sure to deselect the Yahoo toolbar option. [top]

Broadband Tests: Pedal to the Internet Metal

BroadBand DSLReports: The granddaddy of online testing, DSLReports gives you a stack of tests to try. The Speed Test (choose the “Flash 8 plugin based speed test”) has a cool speedometer that displays your upload and download speeds. You can test your speed by using any of six servers in different locations. You’ll get a good average by testing with two: Try one close to home and another across the country.

Line quality test: If you’re worried about your connection, use the Line Quality test to look for packet loss and excessive latency. If you register and log in when testing, and then come back for another test, you can compare results between tests

SpeedTest: The site is powered by Ookla, one of the two tests that the feds use–and which the L.A. Times article implies overestimates results. But Speedtest is quicker to get to than DSLReports and is has an oh-so-nifty auto dashboard-like interface, the kind you want to show off when friends are in your office.

Fix Your Broadband’s Sluggishness

If your Internet connection is really sluggish, like emulating dialup, consider doing a couple of the tests I mentioned this week every hour or two for a few days–and keep a log. If you’re getting nowhere near advertised rates, send the results to your ISP — and raise a high-speed ruckus.

If you complain vigorously, and can substantiate your complaint with test results, chances are good, or at least better, that you’ll get some help.

And it could be something simple causing the problem. Here are a few typical issues (and solutions) to high-speed hassles:

  • If you have DSL and your phone lines are old and cracked (ditto for cable modem); or if any of your lines have been chewed by squirrels or other vermin, performance can degrade substantially. Ask that they be replaced. You may get lucky and they’ll do it. My Copyeditor, Kim, kvetched and said it ain’t easy. “Will Charter replace in-home phone lines? Speakeasy certainly won’t… we had this problem for years and were told that we could pay big bucks for the ISP’s contractor to do this, or big bucks to AT&T for them to do it. Ultimately the problem went away when we had everything rewired during our remodel.” Good point.
  • Use the cables that came with your modem. I know, you really like the 40-foot replacement cable you’re using; unfortunately, it could be degrading performance.
  • Don’t throw away the ferrite chokes (they look like plastic cubes) that wrap around the modem’s cable and power line. They keep RF interference out of your LAN. If you didn’t get any with your modem, sometimes a simple filter on the lines can reduce interference from other electronic devices. Your ISP can supply them.
  • If there are devices in your office that can emit RFI, such as halogen lamps or AM radios, they could slow down your connection. Isolate your modem from those items.

[This post is excerpted from Steve’s TechBite newsletter. If you liked it, head here to sign up–it’s delivered on Wednesdays to your inbox, and it’s free.]


Read more: 

6 Comments For This Post

  1. Christian Says:

    “It stayed mostly at 350 kilobits per second (Kbps) with lots of bursts as high as 460 Kbps. In megabits per second, that’s 2.8 Mbps with bursts as high as 3.7 Mbps. (Techies: That’s based on 350 kilobits per second = 350 Kbps = 8*350 = 2800 Kbps = 2.8 Mbps.)”

    Think what you meant to say was, “350 KiloBYTES per second…. as high as 460 KiloBYTES”

    And when you abbreviate kilobytes per second, you use a big B (KBps).

    so your last line should read:

    (Techies: That’s based on 350 kilobytes per second = 350 KBps = 8*350 = 2800 Kbps = 2.8 Mbps.)

  2. dale Says:

    the program does nothing to determine what kind of connection you are using. I have about 4 different types of conenction on the pole outside my house, how do they know which one im using in the test?

  3. George Warren Says:

    Shouldn’t we ask, what is the problem we are trying to solve with faster broadband?

    This seems to me like an answer in search of a question.

    Has anyone done a cost benefit analysis?

    Does faster always equal better?

  4. George Warren Says:

    And by this I meant the money being spent in national broadband, not fixing your connection.

  5. dholyer Says:

    My Qwest DSL line first started out a 147KBs on it’s 1.5Mbit line. Now 2 years later and still paying the same price I’m now peaking at 163KBs. This free increse in the size of the data pipe could be due to the fact that most of the trunk phone lines are now fiber not copper. I’d like or hope Denver will win the Gigabit test plan Google will start this fall. If not figure in 10 years most metro areas will have moved up to the gigabit fiber optic speed. This may offer the chance for everyone to have a HD picture phone and 50 HDTV channels. Who knows in a century Earth may no longer be a RF polluter in the Milky Way. And could that be why SETI is having a hard time finding RF Polluters to prove there is life out there.

  6. dholyer Says:

    I have a question, when I first got into computers in 1976 in high school. The 110baud modems used 11 bits per byte. Now that most things are in digital form does serial data transfer no longer have or need start and stop bits or a parity bit to validate the byte that was sent. You would think that if these aspects of serial data being sent over a one bit data path will greatly change the KB and Kb data rates. Is this factor not being used in the data rate computations.

    So either serial data transmission has progressed past the start/stop/parity bits or someone has forgotten to include this and only doing simple math.

    Along the same point are they thinking 1Kbtye is 1000 not the true binary to decimal numeric conversion of 1024.

    If they forgot this fact do they not need to correct their math figures. For those that used data rate computations please show the correct figures or give proof that the star/stop/parity bits are no longer needed and serial data is now the equivocal equal to parallel data transition.

    I first started reading about this kind of stuff in 1972 when I found a box of 1968-1971 Radio Electronics Magazines and then some Popular Electronics ones. And started to read Byte Mag in the library in 1977.

    If I’m wrong please correct me or correct the other guys that have made errors in their serial data rate computations.

    Either way I will ask my brother who is a head Electrical Engineer at the Denver Lockheed Martin plant, and was building the Orion Command Capsule for the Ares Rocket program to the Moon. Thank you President O’buma for erasing our future moon bases and free fusion fuel on the moons surface.

    P.S. I did spell his name correctly because he is making Americans become bums or at least Euro members.

3 Trackbacks For This Post

  1. This is a test | Stop Panic Now Says:

    […] Test (and Fix) Your Broadband Connection […]

  2. Howdy! | Webneur.com Says:

    […] Test (and Fix) Your Broadband Connection […]

  3. My weekend with COMCAST Support « Authentic Voice Says:

    […] Test (and Fix) Your Broadband Connection (technologizer.com) […]