At what distance does SFP GLC-LH-SMD work?

[Tagir Temirgaliyev]

At what distance does SFP GLC-LH-SMD work?

Recently I had a conversation with a colleague. Between the two buildings installed SM optic cable. Length is about 15 km, attenuation is about 8 dB. I say, we will install SFP LX and it will work. The colleague says that according to the datasheet will not. I knew perfectly well that according to the datasheet 10 km.

https://www.cisco.com/c/en/us/products/collateral/interfaces-modules/gigabit-ethernet-gbic-sfp-modules/product_data_sheet0900aecd8033f885.html

We open the datasheet, it shows 10 km on the SM cable and the maximum allowable loss on the cable 6 dB.
With a transmission power of -3 to -9.5 dB.
Receiver sensitivity is -3 to -20 dB.

I think this information in the datasheet is a bit outdated.

Firstly, a few years ago and maybe more than 10 years ago SM cables could have losses of 0.4-0.5 dB per km, which results in a total loss of 6 dB per 10 km,
but now the cables are 0.35 dB / km and the welding places are not 1 dB and already 0-0.1 dB.
Now even for 15 km attenuation can be about 6 dB.

Secondly, the maximum allowable loss of 6 dB also raises the question, if we take a new SFP module, then the transmitter will give output power -3, -4 or -5 dB and the sensitivity of the receiver is -20 ie. even in the worst case 15 dB, and even if we take a margin 5 that remains 10 dB. Previously, it was always recommended to take a margin of 3 to 8 dB. This was due to the fact that in the event of breakage and new welding, a loss up to 2 dB could appear at the breakage site, because it is necessary to weld in two places and 1 dB each welding place but now the welding quality is different 0-0.1.

As a result, SFP LX installed and work fine. Up to alarm levels far.

I am asking here my colleagues to say their thoughts and comments.

[Bill Ferguson]

In my experience, as long as the RX power level exceeds the minimum value (-20 dBm), it works.  Distance is secondary to span loss,  Distance is (almost*) irrelevant.

Cisco (and the industry) publishes the extremely conservative 10km because they know that many people don't have the proper fiber inspection and cleaning tools, because fiber quality varies greatly, and they can't predict the number of patches you will have.  Also, many people don't even understand optical power measurement and span loss, but almost everyone understands what 10km is.

I have many spans that exceed 10km where I'm using LX optics.  I like to keep the RX above -18 dBm.

The same thing happens with 10GE optics.  I frequently exceed 10km with LR, and I often use ER optics on one side and LR optics on the other side to stretch the loss budget even further.

It's not so much that it's out of date, because we've always been able to exceed 10km if the span loss is low.  That's what the LX standard is, and standards tend to never change.

*at some distance, chromatic dispersion comes into play and will compromise the clean signal, but at 1310nm, power loss is so much more significant that nobody even bothers with thinking about the dispersion affects.

[andrecsthompson]

@Bill Ferguson

What you have said is very interesting.

I also would like to know what distances you achieved so far.

- Andre