daisy chaining problem

[dharris70]

hi,

several days ago we had a power surge and lost a fax machine that was on the same surge protector as our network gear. since then we have experienced high ping times, a lot of timeouts, etc. our set up is as follows:

ISP connects to wan port of linksys wrt54gs, it is configured with a static IP. on the lan side, two computers connect directly to the router. on the other two ports in the router, i have two dumb switches connected to add network drop capacity. one is a retail plus ez820-s, the other is a linksys ezxs88w.

since the power surge i have moved all network gear to a UPS. if i unplug the switches from the router, everything works fine. if i plug either switch into the router, timeouts immediately start to rate of about 80% lost packets. all ping tests are being done from a computer on our lan to the lan IP of our router. the router is also the dhcp server for all computers. there are a total of 5 computers in this configuration and one network printer. OSs are XP and win 7.

this configuration with two switches has worked fine for 2 years. i can't understand why now plugging them into the router causes this problem. i also tried replacing the linksys wrt54gs router with a netgear n150 router and the same behavior persists.

any thoughts on what i'm missing?

tia,
dave

[jdc2000]

Possibilities:

1. The switches memory of your network paths was cleared by the power surge, and they are having to relearn things.

2. One or more of the ethernet ports were zapped by the surge. Try a new (spare/backup/different) switch to see what happens. Test after hours, and plug in ONE device at at time. to see when the trouble begin.

[westom]

several days ago we had a power surge and lost a fax machine that was on the same surge protector as our network gear. since then we have experienced high ping times, a lot of timeouts, etc.

The reasons why a router would lock the rest of a network are too numerous to list. Long before blaming anything, the most basic facts must first be provided. For example, what do one or two LEDs for each end of each Ethernet cable report?

Meanwhile, you had damage because a protector was too close to electronics and too far from earth ground. That power strip does not even claim to provide effective protection. Read its numeric specs. Where are the numbers that list protection from each type of surge? No protection claimed. That power strip is recommended by a majority because the word 'protector' and 'protection' sound alike.

View spec numbers for that UPS. Its only protection number (joules) is even smaller. No problem. Near zero protection means advertising can claim 100% protection. Subjective claims make obscenely profitable profits at the expense of the most naive.

For over 100 years, protection has always been about earthing a surge before it even entered the building. If your router was at risk, then other more important appliances are at same risk. What protects your furnace, digital clocks, and smoke detectors? Informed homeowners earth one 'whole house' protector. So that even the power strip and UPS have effective protection.

To have damage means a surge had both an incoming and outgoing path. One wire may be AC electric. What was the other wire? That other appliance may have suffered overstress. Or protection inside that other appliance may have protected it. But this you know. A surge was all but invited to go hunting for earth destructively via appliances. At least two are cited as destructive connections to earth - routers.

Apparently an adjacent protector did what we engineers have seen often. Unearthed protector too close to appliances and too far from earth ground can even give a surge even more destructive paths. Make surge damage easier. And only do protection that its manufacturer specification numbers say it will do. Near zero. Step one. Learn protection numbers for each UPS or power strip. Then learn why facilities that can never have damage, instead, use one ‘whole house’ protector – that also costs tens or 100 times less money.

[CincyJim]

For over 100 years, protection has always been about earthing a surge before it even entered the building.

100 years ago was 1912, was electricity available to at least 50% of the houses/buildings back then? I assert not.

Eliminating the quoted phrase would eliminate an ostensible error. I love facts, and assert your posted error demeans the value of your fact-based post.

[westom]

100 years ago was 1912, was electricity available to at least 50% of the houses/buildings back then?

Irrelevant how many homes had electricity before 1900. Protection from direct lightning strikes was proven by experience and patented in the late 1800s.

Early 20th Century Ham operators would suffer damage from direct lightning strikes. Disconnected and put the antenna lead inside a mason jar. Still suffered damage. Damage stopped when the antenna lead was earthed.

Telephone operators were routinely killed when lightning struck wires and their headsets? Of course not. Operators that long ago worked through all thunderstorms without fear because phone wires connected, as short as possible, to earth before entering the building and via a protector.

Well proven solutions are that old. That well proven. Is the only solution found in any facility that cannot suffer damage - including munitions dumps.

Any homeowner can implement the same proven solutions for about $1 per protected appliance. More responsible companies (ie ABB, Square D, Leviton, General Electric, Siemens, Intermatic) sell well proven solutions. A Cutler-Hammer product sells in Home Depot and Lowes for less than $50.

Meanwhile, what or why those routers failed requires more facts, as discussed previously.

[dharris70]

hi All,

the issue was resolved with a firmware upgrade. interesting discussion though...

dave

[westom]

the issue was resolved with a firmware upgrade.

Firmware worked just fine did not suddenly go bad on its own. Something external changed that firmware. Facts in those previous posts provide possible reasons for an unacceptable firmware change.

You have apparently identified and fixed the problem. Now fix what caused that problem.