It is technically RF from a physics point of view.  You can swamp out any 
receiver with too much signal and it will lose its ability to properly 
demodulate.  All optical receivers have a range of powers where they will work. 
 Too much or too little and you have problems.  Exactly like RF.  

From: Josh Reynolds 
Sent: Thursday, August 3, 2017 7:36 AM
To: [email protected] 
Subject: Re: [AFMUG] UBNT OLT Transition from Active Ethernet to GPON

We had to pad much of our gear in the Datacenter as well as the NOC in the 
office on short single mode runs, including our Juniper MX960s.

On Aug 3, 2017 8:33 AM, "Adam Moffett" <[email protected]> wrote:

  Don't forget that Sterling is doing ethernet, not PON.  I've plugged in a 
pair of 40km ethernet SM SFP's at 3' away and they work fine because they'll 
turn down the tx power.

  With PON there's a single Tx source for the entire network, it can't turn 
down because you may have subs at 1km and other subs at 20km.


  ------ Original Message ------
  From: "Josh Reynolds" <[email protected]>
  To: [email protected]
  Sent: 8/2/2017 1:29:28 PM
  Subject: Re: [AFMUG] UBNT OLT Transition from Active Ethernet to GPON

    To clarify, there is no real intelligent attenuation on any optical product 
I have ever seen. Some can do a 2-3db depending on the product, but it's never 
really a truly intelligent system with bidirectional communication between the 
optics to negotiate power levels. 

    I may be wrong but this is just my experience.

    On Aug 2, 2017 12:26 PM, "Josh Reynolds" <[email protected]> wrote:

      How could they? TX and RX are different optic sources. You might have a 
TX power level much higher on one end than the other due to manufacturing 
differences or different equipment.

      On Aug 2, 2017 12:10 PM, "Sterling Jacobson" <[email protected]> wrote:

        Hmmm, that's not good if it can't auto-attenuate down.

        Sounds like they need to fix that.

        Most of my SMF lasers and links are short and 'hot', but doesn't seem 
to bother anything I'm currently using.



        -----Original Message-----
        From: Af [mailto:[email protected]] On Behalf Of Brett A Mansfield
        Sent: Wednesday, August 2, 2017 10:16 AM
        To: [email protected]
        Subject: Re: [AFMUG] UBNT OLT Transition from Active Ethernet to GPON

        I need to print a retraction here. I have been talking to Martin at 
UBNT and he shows me the error of my ways. I do not have a 50% failure rate. In 
fact, it's a 0% failure rate. My signal was just too hot. I'm new to the SM 
Fiber game, so I'm learning as I go. I didn't realize the signal could be too 
hot at only -3 dB. All of my multi mode Fiber sits at -2 dB and works really 
well.

        You learn something new every day.

        Thank you,
        Brett A Mansfield

        > On Jul 31, 2017, at 2:46 PM, Brett A Mansfield 
<[email protected]> wrote:
        >
        > I have heard a lot of complaints from DirectCom customers about their 
Fiber never being close to what they pay for, but that may be more related to 
the way it's throttled then the GPON.
        >
        > I've been playing with several of these ONT/OLT over the past week. I 
really like them. Though I have a 50% failure rate on the nanoG's. The fiber 
port breaks very easily.
        >
        > Thank you,
        > Brett A Mansfield
        >
        >> On Jul 31, 2017, at 2:31 PM, <[email protected]> <[email protected]> 
wrote:
        >>
        >> We have good luck with 32 customers per 2.4 Gbps down on GPON.  
Lotsa overhead.  No problems, not even close, so far.  And we are selling more 
Gig circuits than ever before.
        >>
        >> -----Original Message----- From: Sterling Jacobson
        >> Sent: Monday, July 31, 2017 2:19 PM
        >> To: '[email protected]'
        >> Subject: [AFMUG] UBNT OLT Transition from Active Ethernet to GPON
        >>
        >> Anyone tried their PON OLT CPE and OLT 8 port (128 clients per port) 
1U unit?
        >>
        >> I see pricing around $70 retail for OLT, but haven't seen pricing 
yet for the OLT 1U unit.
        >>
        >> Also, I'm active fiber right now, so I have full 1 to 1 panels in 
the rack already.
        >>
        >> If I wanted to 'migrate' to OLT from active I would need some sort 
of transition panel/setup right?
        >>
        >> Right now my density is 48 ports per 1U 1 to 1 single family home 
connections.
        >>
        >> The UBNT Fiber OLT has 8 ports handling up to 128 clients each, with 
20Gbps uplink capability (not quite sure on those split details yet).
        >>
        >> I currently only take a max of 576 per cabinet on active, so I could 
easily use just one of these UBNT fiber OLT units.
        >> If I don't care about the share ratio I guess, I would just get 
another 576 panel count that spliced 72 count to each port and I'm done.
        >>
        >> I'm unclear what that panel/splice would look like though since I've 
never actually done GPON.
        >>
        >> And I would probably want to not load up that many per port, and 
instead maybe get four of the UBNT Fiber OLT units.
        >> That would take up 4U of rack space, the fanout would probably still 
take up 4U of rack space, for a total of 8 U.
        >> And I would have instead an 18 customer to 1 port on the GPON 
instead of 72 which I like better for future use.
        >>
        >> Do these UFiber OLT 1U rackmount units share just 1Gbps per each of 
the 8 ports? That would only be 8Gbps needed total.
        >> So I assume the GPON spec they are using can transmit more than that 
per each of the 8 GPON ports, right?

Reply via email to