Is there a "trick" to plugging in a Switch or did I kill it?

[EdH]

Static addresses will work fine in a network as long as there is only one piece of equipment for each unique address. You would need to know the first three parts of the router's base address to connect with it.** I don't know your provider's setup. If you try to use a static address that your provider is already using elsewhere, there will be a conflict.

However, you can test your equipment using static addresses. Without the switch, you can only connect machine to machine, but with the switch, you should be able to connect all 4 to each other.

The first test would be to set up two machines with static addresses and try to ping each from the other. Modern machines should be able to do this with a standard cable, but older machines might need a crossover cable. I would try a standard one. If it works, then that's one step proven.

After that test, you can connect the two machines to two standard ports on the switch and see if you can still ping (or, can now ping) each machine. If the machines can ping each other without the switch, but they can't with the switch, I would consider the switch (or, supply) as being bad.

Power supplies can be bad and still turn on lights. The circuitry in the switch wants a pretty noise free and load capable voltage and some switches draw some serious amperage. If the ripple on the input voltage is too high, it will interfere with the communication packets.

**I think one of your posts mentioned being able to connect the individual machines to the incoming ethernet cable. If you do this you should be able to see what address is assigned and use that address as your static "testing" address. Be sure to choose different ones for your other machines.

[error]

Originally Posted by petrw1

I don't think it is the switch power supply because the Power light comes on....just NOT the Link/Act lights.

There is power but it is not stable enough - DC filtering has failed because of a bad capacitor.

At least that was the case with my device and the symptoms were just like yours.

[LaurV]

Yes, static IPs work well, I use static for home, 4 computers (including 2 wireless laptops, the router has 4 wired ports, plus wireless and adsl/wlan), 3 telephones (wireless), 1 tablet (wireless) and printer (wired/wireless). You can try the classic 192.168.0.x or 192.168.1.x (my router has the default 192.168.1.1). It is good to use the default subdomain because in case you reset the router you do not need to change the IP address of one machine to connect to it. For mine, the machine which connect for admin purposes has to be wired and in the same IP range. Therefore I assign IP addresses starting with 192.168.1.10 (reserving the lower 10 for whatever purposes may appear) for wired equipment, and with 20 or 50 for wireless. I mean, I did this manually for all devices in the house, and I disabled DHCP, which is also safer, as most unwanted guests will be refused connection as their default DHCP is on, and they can not get and IP address from my access point. This way I also know at any time which machine is connected and what it does.

[Madpoo]

Originally Posted by petrw1

...I don't think it is the switch power supply because the Power light comes on....just NOT the Link/Act lights....

Do the computers show an active ethernet link?

If the switch is that old, I wonder how old the router (or whatever CPE) is? Same vintage? When was it last rebooted? Maybe the DHCP services on it are stalled and everything just happened to be working fine until it all got moved around.

Anyway, if the computers are showing an active link but the link LED on the switch isn't lit, it may just be that it'll only light up when there's activity, not just a link? No idea... cheap switches, who knows.

What IP addresses are showing up on the PCs? Are they what they should be? Or are they the annoying 169.254.x addresses? (which means no DHCP server could be reached).

Note that after you reboot the router, if that was the problem you'll still need to go to each machine and make it request an IP address ("ipconfig /renew" on Windows), or just reboot again.

But there are a # of things that could be wrong, all of which become progressively harder to troubleshoot in a forum setting.

[petrw1]

Considered the value of my time and just got a new switch....only $20 .... worked.

Thanks for all the tips

[snme2pm1]

Originally Posted by petrw1

Considered the value of my time and just got a new switch....only $20 .... worked.
Thanks for all the tips

Evidently there are all kinds of switches with a variety of qualities.
The small and cheap switch that I thought would be lovely perhaps more than ten years ago failed whilst in use at another site.
I repaired it. I don't immediately remember the particular electronic fault after the passage of some time without consulting records.
Whilst I could deploy it, I know that it is a watt waster due to a particularly poor design which I became aware of during such repair.
I'm not sure, but the fault might have been related to overheating of the lame series power regulator due to that particularly power inefficient design arrangement.
I recall augmenting the heatsink for the significant watt waster component, but it remains a poor design, so will only use it for occasional purposes.
PS. I wouldn't be surprised if my records showed that some electrolytic cans near the main watt waster device were cooked real good.

[EdH]

Originally Posted by petrw1

Considered the value of my time and just got a new switch....only $20 .... worked.

Thanks for all the tips

Glad to see you're running...

[Madpoo]

Originally Posted by snme2pm1

PS. I wouldn't be surprised if my records showed that some electrolytic cans near the main watt waster device were cooked real good.

I recently did a repair on something that needed some caps installed (some home automation stuff). The caps in a switching power supply are apparently vulnerable to ESR (equivalent series resistance) degrading.

The fix, as researched by some other folks, was to swap out the caps with roughly equivalent ones (higher voltage rating in a few key places, for an extra margin), but with low ESR ratings. Also having some rated for higher temps is a plus.

About 4 USD in parts (plus $8 for shipping...) got me 4 spare sets of capacitors, to replace all 5 of them. Not bad considering it put a $60-$70 part back in business with only 30-60 minutes of soldering.

The problem is you don't know if the ESR went wacky or not without removing it from the circuit and testing it with a meter capable of that measurement. Normal cap measuring might show it's within spec of the microfarads. The few people who replaced their and actually had an ESR meter were able to test and show just how bad one or two of them became.

Anyway, for an el-cheapo switch, it's definitely easier (and probably better) to just get a new one. Now, if I had an expensive stackable L3 switch with a high density of 10Gb ports, I just might take the time (if it didn't already have a swappable PSU).

[snme2pm1]

I reckon that Aaron has a reasonable grasp, but to flesh it out a tiny bit, and for others...
Those not interested in electronics might like read something elsewhere.

Originally Posted by Madpoo

...swap out the caps with roughly equivalent ones (higher voltage rating in a few key places, for an extra margin), but with low ESR ratings. Also having some rated for higher temps is a plus.

Caps have a wide range of actual value compared to their marking. Depending on the particular circuit purpose, it might be reasonable to adopt a differently marked component.
Usually we endeavour to retain the same nominal parameters for a replacement, which avoids confounding the next person to work on the board, and very often it is physically impossible to adopt a larger nominal capacitance or voltage anyhow.
Physical dimensions are often a significant selection parameter.
I would suggest only stocking 105 degree, not 85.
The higher price for low ESR must be paid for high frequency or high ripple current situations, or if stiff low impedance is otherwise deemed important.
Sometimes a suitable MLCC (multi layer ceramic cap) might be gratuitously paralleled copper side if that might round off some of the sharp edges on stress current.
It might be possible to use a replacement solid dielectric cap, if you can source some with legs, which I have achieved on rare occasions.

Originally Posted by Madpoo

About 4 USD in parts (plus $8 for shipping...)

I tend to gamble a tiny bit and buy out of China with negligible shipping cost included.

Originally Posted by Madpoo

The problem is you don't know if the ESR went wacky or not without removing it from the circuit..

For purposes here we're not chasing caps that already have visible symptoms of bulging or venting like Chernobyl.
An ESR meter can be used in circuit to some extent, but caution is needed.
For casual get-lucky purposes, it can be used to probe around a board in an effort to quickly identify a cap that has dried out showing high impedance, however if the reading isn't high, then yes the component needs to be lifted off for any determinism.
For comprehensive whole board rework, I would mark, lift off and record measurement of all liquid cans, which can be a time consuming struggle for a multi-layer board.
If the manufacturer fitted dubious caps then it may be desirable to replace them all.

Originally Posted by Madpoo

Anyway, for an el-cheapo switch, it's definitely easier (and probably better) to just get a new one.

Yes, a newer module might (or not) incorporate a superior design or specifications.
There are also situations where it would be a plain nuisance to attempt tracking down a new module that will operate decently in some not so new system.

There seems to be a diminishing realm of folk that would contemplate repairs of any kind.
I'll happy to see that Aaron is not shy of using a soldering iron.

[kladner]

I replaced some caps which seen to blow very often in my partner's particular Samsung monitor.. Getting the replacements involved a fifty mile round trip to a suburb of Chicago. The monitor is still running years later. I did get 105 C capacitors.

There is something very satisfying about making a good, mirror-surfaced solder joint.

[snme2pm1]

Blow very often? Still running years later? Two different tunes there!
What's the truth?
Re Chicago tripping: Well these days, you can have components delivered to your door from Shenzhen or similar at no noticeable fee; I don't know how that is possible.

A good solder joint would be one making connection with real metal and without the rubbish, from oxides etc.

PS: I am aware that monitor internal modules are frequently poorly ventilated, thus affording a hightened risk of thermal failure.
PPS: Depending on the situation, you might want higher grade low ESR caps, or perhaps MLCC parallels, and maybe provide some vents.