Problems with Internet Disconnects - Router Issue?

[I'm Murrin]

Since about midnight on Monday, my internet has been disconnecting with a very high frequency. It varies, with most disconnects happening after about 5-10 minutes, but sometimes it'll stay on for an hour, and others it'll disconnect every minute for a while.

(NB: my wireless connection to the router, and the router itself, stay up indefinately with no issues. It's the DSL line from my phone socket into the router I'm losing contact with (when it happens, I see "DSL line is disconnected" on the router home page).)

I've noticed a couple of times when I've come home from work that the router shows the connection as having been active for hours - 16 hours a couple of times, about as lnog as I was not using the computer for. It's then disconnected as soon as I've started using it.

I've spent a while trying to find some ISP or exchange-related reason for the disconnects, but today I noticed some threads online mentioning similar problems and blaming them on power supply to the router, saying that dips in supply can cause this issue. Because this seems, from the 16 hour stable connection, to only happen when the computers (and, of course, TVs and so on) in the house are in use, I'm wondering if this might be my problem.

My router is connected to a twin socket on the wall right next to the electricity meter and circuit breakers (because that is the location of our master phone socket), and is the only thing connected through that socket. I believe it's powered through the same breaker as other sockets in the house, however. This issue has also only been happening for a week, and nothing has changed in our power use in this time.


Anyone have any insight into these kind of problems, what might be causing it, what I could try to solve it from my end?

I was originally doubting it could be a router problem as everything else worked and the only issue was the line showing as dead, but I'm not sure any more.

[High Lord Tolkien]

Dumb question time....do you own the router or are you paying a monthly fee for it?
If you don't own it they should replace it for you free of charge.
I do with my cable modem and usually get an increase in performance each time.

[I'm Murrin]

Own it. Was a cheap one, but it's worked pretty well until now. This problem was sudden, gone from perfect to four days of DCing every few minutes...

[Avatar]

Surge protection plug for the router?

--A

[I'm Murrin]

Not even sure it's a power issue yet, since nothing in the house has changed. How likely is it it could have developed a problem on its own?

[Menolly]

Have you asked your ISP to check the DSL line itself, Murrin? I had a similar issue about a week ago during some heavy rain here, and the phone company did a check on the lines remotely and wound up sending a technician up a pole to work on a box on the street outside the house. No major problems since then except once, and that fixed itself when I unplugged and replugged in the modem.

[wayfriend]

I would think about the possibility that it isn't the power at all.

DSL, IIRC, is tricky in that quality degrades with the distance from the station. If the station is too far away from your house, you simply can't use DSL. (I might be completely wrong, or thinking of something else.)

If something changed in your neighborhood, then this could explain it.

Also, when you use a DSL connection, that's when it notices that the quality has degraded.

You should be able arrange with the phone company or whoever to have them run a test to your router and evaluate the signal quality.

The last time I had intenet issues, it turned out the wiring from my house to the street had actually gone bad. (I have cable internet, but it suffers from quality degradation as well.)

[lucimay]

uh huh. we had that same problem (as wayfriend) only it turned out to be a cable actually in our apartment.

[I'm Murrin]

Update: I decided to give something a try, and unplugged the phone cables, took off the cover on the main socket, and plugged the phone and router directly into the test socket. I've disconnected once in 3 hours.

Conclusion: It's the internal wiring.

[SoulBiter]

I would think about the possibility that it isn't the power at all.

DSL, IIRC, is tricky in that quality degrades with the distance from the station. If the station is too far away from your house, you simply can't use DSL. (I might be completely wrong, or thinking of something else.)

If something changed in your neighborhood, then this could explain it.

Also, when you use a DSL connection, that's when it notices that the quality has degraded.

You should be able arrange with the phone company or whoever to have them run a test to your router and evaluate the signal quality.

The last time I had intenet issues, it turned out the wiring from my house to the street had actually gone bad. (I have cable internet, but it suffers from quality degradation as well.)

Thats correct. DSL has a limitation from your house to the switch. If you are more than 5 miles (I think) from a main switch then you cant get DSL. But, if you are close enough for DSL you shouldnt get good performance, bad performance, good performance.. etc. You have dedicated bandwith with DSL so you arent sharing your connection with others. Thus you should either have a good connection or not unless there is a hardware or software problem either at your home or the lines between your house and the switch. Sometimes they screw around with the switch and mess things up as well. There is a certain way they wire those so that you dont get crosstalk and crossconnectivity which for DSL causes poor signal.

But you figured it out. Just thought I would throw that in. Before I became a logistics distribution analyst I spent alot of years learning networking. (A+ cert, MCSE NT4.0 and a CCNA) All that education that I dont use anymore.

[wayfriend]

BTW to SB, that was what I was trying to say: DSL line quality diminishes over distance, until at 5 mi or whatever it drops below the threshold of functional. If Murrin was just above the threshold before, then if someone mucked with a line or the switch or something, even a little bit, it could have dropped below the threshold. In such a situation, the line might appear to function when idle, and then crap out as soon as activity started up and larger packets started flowing. Which is what I thought Murrin was describing. At least, that was what I was guessing.