"worst case" says to take the lowest trasmit power vs. the poorest
sensitivity, but also checking to make sure that the maxiumum input
power or other limiting factors aren't exceeded. Likewise you take
the worst fiberloss, allowance for aging and splicing/conectors, etc.
This gives you some assurance that the stuff will work when you plug
it in and continue to work thereafter, if you have something that
isn't mission critical you can work with more favorable assumptions.
You may notice that the numbers aren't exactly what you'd exepect:
transmit -15 dBm this is the lowest power that the device can emit,
it may or may not including aging
-8 dBm you shouldn't see any higher output than this, either
by design or testing for stuff that's "too hot" and won't
last or might threaten the budget or regulatory limits.
received -31 dBm this is the weakest signal level that should allow
the unit to deliver the specified error-rate performance.
-8 dBM this is the strongest signal that the receiver is
guaranted to handle without errors due to overloading or
loss of sensitivity due to damage.
For the case you cite -15dBm - -31dBm gives you a 16dB loss budget
to start with, and you could safely jumper input to output without
fear of damage or overload. There might be other factors (typically
distance related) that could still prevent it from working over some
fiber link that has < 16dBM loss.
George
> From: Joe McGuckin <joe@monk.via.net>
> Date: Thu, 28 Jun 2001 15:31:39 -0700 (PDT)
> To: cisco-nsp@puck.nether.net
> Subject: [nsp] Computing optical link power budget
>
> Most manufacturers quote power output and sensitivity of
> optical interfaces as ranges. e.g.:
>
>
> Transmit power Receive Power
> -15 to -8 dBm -31 to -8 dBm
>
> Which numbers should IU use to figure out how much budget to work with ?
>
> Thanks,
>
> Joe
>
>
This archive was generated by hypermail 2b29 : Sun Aug 04 2002 - 04:12:43 EDT