Feb 162011

My home ADSL Internet access has been flaky for more than a year, with daily service interruptions that wouldn’t disappear without rebooting my wireless router. Naturally, I thought that was the culprit, so early last year I replaced the older 802.11g Netgear with the great spanking new dual band WNDR3700 802.11 a/b/g/n. However, the interruptions continued, although not as often, so after talking to AT&T support again and not getting anywhere, I did my research online (this actually took most time) and selected Actiontec GT701D as the best candidate. I was very happy to finally get my internet to its regular slow but reliable performance. My decade-old SpeedStream 5360 DSL modem was dying.

Actiontec works great and it was very easy to configure. I discovered that if you leave it on for some time, it auto-detects the type of the ISP connection and auto-configures itself for ppoe or no authentication. Of course, you have the option of reconfiguring it, which is what I’ve done in the first place and then several times more just to figure out which one works better. I didn’t see much of a difference between using it as a bridge, a router, or for direct connection to a PC, or in my case, PS3. It wouldn’t be me without doing a test, so I’m posting the results here. Although there’s no perceivable difference, since I game very rarely these days with two little daughters, when I do, nobody else is around, so I connect the PS3 directly to it, which is why it’s most convenient for me to keep it configured as a router. I’m a big Battlefield 1943 fan (or used to be) and in online games miliseconds can make the difference between (virtual) life or death. In any case, when I bought it, I couldn’t resist but testing my bandwidth with it configured as a bridge, as a router, and connected directly.

Broadband Tests

Here’s the spreadsheet with measurements in kbps, in the same order as listed above. I converted the ones in Mbps to kbps by using decimal instead of binary 2^10 numbers (1,000 instead of 1024), the main reason for that approximation being the fact that these numbers are compared to themselves in three different configurations, so the math didn’t have to be exact. Also, when I had results that were much different than from other web sites, I repeated the test and took the average of the two or three passes and used that number. You can see that there was virtually no difference in upload speed and in latency or jitter in all three cases, hence I omitted them from the graphs.

Speed test hosts in order of appearance:

1. Broadband.gov (OOKLA) – dark blue
2. Speedtest.net – orange
3. chicago.speedtest.frontier.com – yellow
4. bandwidthplace.com – green
5. Dslreports.com/stest – Sprint (Chicago) – dark red
6. Dslreports.com/stest – Speakeasy (NY) – light blue
7. Speakeasy.net/speedtest (Chicago) – dark green
8. Foxvalley.net/speedtest – light green
9. CNET Bandwidth Meter – purple
10. Average of all tests – black

The line across all bars is showing the achieved average download speed.


The average achieved download speed for all three configurations was almost identical, around 4,000 Mbps. What surprised me most was virtually no difference in latency when I eliminated the DSL router and connected the laptop directly to the Actiontec modem. I expected at least several milliseconds improvement but observed none, so this set is good for VOIP and online gaming in any of the tested configurations, and connecting directly from PS3 to Actiontec during online gaming is most likely just an overkill, unless I’m missing something (please let me know if you think so). The only difference that makes a wire bypass to PS3 worth is that my usual WiFi connection has about 50 visible access points in this building to compete with, so that may impair my PS3 network gaming experience. Update May 22, 2011: Although my tests didn’t show much difference, my gaming performance is much better when my PS3 is wired directly to this modem configured as a router.

And Now about Bandwidth Tests Themselves

There are big differences when using different test providers, but their results are compared to themselves, so that’s not skewing the comparison. There were some individual variations which might have been caused by a bad link anywhere in between my ISP and the server where the test was coming from, by the test web site being slow, or skewing results by caching on my ISP (I often wonder whether they do cache bandwidth tests more than anything else, so they come out faster than usual). However, the average stayed almost the same. One speed test that is consistently showing lowest results is speakeasy.net/speedtest, but that one has been showing lower results than the others for more than a year, possibly in an attempt to attract more customers to their ISP. In any case, I should have 6 Mbps, but my highest results don’t go much over 5 – smells like another one of those decimal vs binary measurements where manufacturers and ISPs use the one that suits them best. I however can’t blame this modem for that, as I was getting very similar results with the old one when it worked well.

I Recommend It

As many users who reviewed it on Amazon.com, I also recommend Actiontec GT701D DSL modem. We didn’t have a single outage since several months ago when I bought it, the setup was easy, it works flawlessly, and it has more configuration options than your usual DSL modem, so it’s thumbs up. My test was on AT&T (ex SBC Yahoo) 6 Mbps DSL service. Your results may vary.

 Leave a Reply

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>