fey42

I live outside Lebanon, OR and communicate through a Remote Central Office (RCO) to the main office at which point I pick up the backbone (I think) through CenturyLink (CL). CL bought out Qwest and that is the my access to the world. I have my second DSL-2740B and it has the same issue as the first one. My download speed slowly decays over time. When I reset my modem I get 3Mbps, after 24 hours I get 2.5Mbps another 24 hours and I get 2Mbps, another 24 hours and I get 1.5Mbps, etc. Reset the modem and it is back to 3Mbps Below are some modem stats, that BTW agree with the equipment at CL Downstream Upstream Line Coding(Trellis): On On SNR Margin (0.1 dB): 151 170 Attenuation (0.1 dB): 520 290 Output Power (0.1 dBm): 192 123 Attainable Rate (Kbps): 5216 956 Path 0 Downstream Upstream Rate (Kbps): 3520 576 The reason my Attenuation is so high is that I live 10,000' from the RCO. I have tested to San Francisco and Los Angeles and the same thing happens. Been fighting this for 3 moths now. Everything was fine with my OLD Westell 7500 modem, but it died. Anybody have any clues? Thanks, Fred

I am "conversing" with CenturyLink about my congestion problem during the evening rush hours (7PM - 11PM), that occurs almost without fail every night. I live outside Lebanon, OR up in the foothills of the Cascades and CL bought the Qwest backbone in this neck of the country. I am not the only one that has this problem up here. Folks that live closer to the remote central office we use with 8 Mbps also drop down to 60Kbps during rush hour. My first personal problem is that I live at the extreme edge of the the DSL reach so my max is only 3 Mbps on a good day, downhill, with a tailwind. But it is beautiful. They have sent out a technician to "look" at the line and he found that my error rate is almost 0 from my location through the remote central office to the real central office. They have me on a port that is not in overload from here to there. I am convinced the congestion is on the other side of the real central office. The tech agrees. Soooo, IF I could get CL to agree, what and how do I setup a temporary test server? Also how do I "move" it further and further away until I find where the congestion starts to occur? Please be gentle with me, since I don't really understand the intricacies of doing this task. If this has already been answered I am sorry I did not find it in the forums. Thanks, Fred (fey42)

Like most things obsolete, Global Warming, Tenured Professors, NWS, etc., it is called publish or perish. Adobe, a long time ago, served the greater common good, now they can't program their way out of a wet paper bag. Witness the constant updates to their software that works most of the time or the switch from selling software to renting software. The almighty $ rules.

Your problem might not be your local ISP. Since they have to connect across the country you are at the mercy of every connection made and every node you pass through. They have control over everything until they hand it off to the Internet backbone and then all bets are off, It can go anywhere. I live 70 miles south of Portland, but to get to the Portland Netfix server, I go though Seattle and sometimes San Jose. Also you have time of day that affects speed. Imagine as everyone comes home, has supper and kicks back to stream their favorite show. It starts first on the East Coast and the load picks up as you move West. 7PM is the worst, the whole country is streaming and unless the server is in your back yard you are in trouble. Try your test against multiple test sites and time of day. You will see a wide variance the further away you are from "home" and the time of day.

If I look at Skunkdog results logged as him I see these results If look at Skunkdog results logged in as me I see these results What am I doing wrong? Thanks, Fred

That's not fair, you expect me to read everything golly gee whiz.. Hard to make it stand out more unless it was BOLD and italics also. That would be my problem, not the program. Having made the mistake once, I kept on repeating it. I was trying to get a good window around prime time, that starts on the East Coast at 7PM and ends at 11PM West Coast. All cleared up.

That is not the problem. Look below and you will see in the URL I selected 100 tests at 5 minutes, yet the info in red shows only 50 tests will be run. I think we need to wait for a"higher power" to weigh in.

First I am being brand spanking new to all of this. Second I am trying to resolve a Netflix streaming issue for a neighbor who doesn't even know where to begin (I'm just slightly smarter). His dataset is hjb44. Third I have been using TMN for three days now and love it. It showed the problem the very first run. His download was all over the map from 8Mb to 750Kb. Instead of testing to San Francisco I would like to test to someplace closer, in fact I know the IP address of Netfix in Portland, but tracert quits after the third hop. I live just 70 miles south, I know that has zero bearing on the issue; I am at the mercy of routing. Did a tracert to oregonlive.com (Portland newspaper) and it went to Seattle, to San Jose, then to Portland. How do I go about using one of the Mercury Test Site? Is that portion of code even working? Remember I warned you I am brand new at this. Thanks, Fred

My 2 cents. Mostly everybody treats their computers the same, regardless of whether they are at home or at work. It's working don't touch it. Instead, let's keep up on the PM, like something as simple as using a vacuum cleaner and just sucking at all the fan and vent locations. NOT compressed air and blowing it all in.

I would like it to run 100 times at 5minute intervals or at 10 minute intervals. I am testing my download speed over an extended period to prove to my ISP that they have bandwidth problems during peak time periods. If you look at the URL in the picture (right above your post) you can see that I requested 10 minute intervals and 100 each. It only ran 50. The one above that shows 5 minutes 20 times and it stopped after 18. It failed to complete the sequence. Thanks, Fred

Here is the photo showing a request of 100 in the command line and the screen showing only 50 will be run

Nothing seems to be working quite right. I selected run 5 min tests 25 ea and it quit before completing. Seems to be able to do a short series, but not a long series. See attached file. Tried 5 min at 100 on hjb44 and it stated it would run 50, not the 100 and even that quit after about 15 or so. Meanwhile, if you scroll down to read the results you lose the top part of the screen that tells where you are in the sequence and what time the next test will be. Everything is being run under Chrome as you suggested. Both are running Windows 7, one is 32 bit, the other 64 bit.

Successfully ran all 5 of the 5 min tests on Chrome. I will try a longer run and see if it holds. New test run start 2:36 on Chrome. I reported earlier that Firefox shut down during one test run. Seems that might be the problem. Any clue?

This machine is an almost fresh install of W7. Latest Firefox. I first found and started using your program yesterday and 3:45PM PST. It worked fine the first time and hasn't since. Off to load Chrome, be back.. Running Chrome, start at 2:04PM PDT

Both my laptops are set not to sleep. So far I have no joy, I am back on the original computer that I started with. Yesterday I also had a case of FireFox shutting down. T+8min It tested 2 times and then quit. Now what?