Sfp issues / 1gig trendnet

[givemesam]

Hi there

Spent most the day working on a no sync issue on a fiber pair. Netonix does detect the sfp and displays model number.

I've checked and have good light at both ends.

New netonix, new sfp ( TEG-MGBSX )

Use same sfp in both ends on other netonix switches.

We exclusivity use the 12port ac. We have rev D working fine with autonegotiation off. Recently upgraded to 4.7rc7, no problem.

New switch, I think it's rev F, port is down no matter what we do. I saw in the logs after moving the sfp from port 13 to 14, this show in the log: FAULT on SFP 1, resetting...

Looking on this forum there were 3 things when sfp not present (mine is present)

1- insects, switch is new, no bugs
2- bad sfp
3- bad switch

Any thoughts?

I'm going to swap sfp, cables, and then the switch to flush it out.

I use this sfp exclusivly due to its reliability in my experience. Anyone else seen this?

Any fixes or other things I should look at if my swaps fail?

Wish there was ddm. Spent a half day rolling fibers trying to get it to sync. It eventually did on the edgeswitch end. But link still down.

[sirhc]

Well by your own omission you have used WS-12-250-AC switches before with the brand SFP and it works.

So you're right in that it is one of following:
1) Bad fiber and or fiber jumper
2) Bad SFP
3) Bad SFP port in one of the switches

Unless you have some expensive equipment swapping is the best and fastest way to determine your issue.

Instead of spending the day spinning your wheels I would have done swap testing first?

[givemesam]

Agreed. I couldn't swap test as the spares were across town.

After retrieving then, I determined the switch, the sfp, and the jumper cables were all fine.

We actually have low/no light in one direction which may be the cause for that log error? Not very descriptive.

I could really use some of your expertise here,

The fiber run is campus fiber, it is 62.5 / 125 fiber. On this particular run the campus fiber goes to a junction Midway. From this point it's running the rest of the way on a 3 pair. All jumpers and campus fiber are all the same spec.

What we discovered is from the netonix back to our headend the light is fine. But going the other way we have no light at the netonix. Oddly on that same 3 pair cable we have been successfully using 2 pairs without issue. Same setup. Same netonix. Same stfps.

Even more odd is inside the junction at the halfway point there is light in both directions. I can black out the room and see red light coming from the netonix and I see you red light coming from the head end on the other fiber.

But at its destination that single strand has no light visible to me when the room is blacked out. Of the six five have worked fine. Total distance is about 800 feet. And we've had no problems on the two pairs that we've been using from the same three pair cable. I also tried using a different color pair on the first segment of fiber from the head end to the junction room.

So now that I've explained the setup and that we have low or no light on one strand what would be the best way to diagnose this. I don't have a fiber termination kit but I could get a low-end one and I don't have the meter to give me light readouts. We're generally ethernet guys unless we come across some campus fiber. Or we use pre termed fiber. It may be time to get dirty on that Strand and try and repair it or we can get a SFP that's compatible to do two different frequencies on the same strand so we can send and receive on the same cable but I'm not sure they make multi-mode SFP for that. If anyone knows of one that could Band-Aid this issue on multimode fiber that would be one way to go about it.

I was quick to add to think that maybe there's damage to the cable but there's light in both directions at the junction and I can't tell how bright it is or if there's any light loss that's significant and I also changed the SC jumper to a brand new one that connects the two fibers together. That literally did nothing. I also changed the coupling that they click into in case there was some lost there but those are pretty much Hollow straight through design.

Can anyone give me some coaching on how best to attack this and get it solved in a reasonably straightforward way?

[jowang]

The same 3 pair cable (Same setup. Same netonix. Same sfps) you have been successfully using 2 pairs without issue. And you don't have light on only one strand. Am I right about your current situation?

I also agree that it might be one of the following:
1) Bad fiber and or fiber jumper
2) Bad SFP
3) Bad SFP port in one of the switches

I wonder if you have tried to use the good cables/SFPs that already working in the switch to the port (not working). if this is not helping, maybe you should buy a fiber optic tester to save all your troubles.

[givemesam]

Agreed.

The weird part is there is light on both ends in the center J-box. I really don't think its the SFP or Switch. I swapped tested both. (we did not try it on the working pairs as they are busy and im afraid of introducing any compatibility issues.

So something is happening in the junction for at the base of the riser. Its either to weak and the strand is bad glass introducing too much loss or its a bad head. Pretty certain on that.

We will get a light tester to get some real numbers. Anyone have a good small recommendation on a light tester? (for 850 + 1310)

We decided to bring in some new fiber into the mix. After following the path, we are not happy with the campus fiber we adopted. We are thinking to go with a pre-term single mode reel of 400 ft. Something heavily armored and bring it in. (the single mode is way cheaper than the OM3, both are armored pretty well so no need for EMT. (Its a garage + Riser only)

Thinking of going with this:
http://www.discount-low-voltage.com/PT6 ... ated-Fiber
or
http://www.discount-low-voltage.com/Cab ... ated-Fiber

We really need to move away from OM1 anyway to be 10gig ready for when Netonixs gives us a 16 port 48VH switch with SFP+

Any reason to go OM3 or OM4 over Single Mode? (nowadays optics are very cheap)

[jowang]

You can buy a handheld optical light source (1310nm/1550nm) to test your network. I have used this one from fs, which is cost-effective and functioning. Maybe you can have a look at it.

Why not use the single-mode fibers (since cost is not the issue here)?

[michwave]

Is DDM support on the agenda?

[Eric Stern]

Is DDM support on the agenda?

Not at this time.