Splice a Fibre - Hacker News

Posted by matt-p 12/27/2025

Splice a Fibre(react-networks-lib.rackout.net)

It's interactive, try and splice one!

Written in react, it's moderately heavy and not entirely mobile optimised.

101 points | 43 comments

sathackr 12/27/2025|

The hardest thing about splicing fiber is not splicing fiber(at least not anymore)

It's cable management and routing to keep things from kinking and breaking while accounting for cable flexing, thermal expansion, and unforseen circumstances like another company lashing their cables to yours for vertical support.

All while maintaining future serviceability

gerdesj 12/27/2025||

"like another company"

What kind of org shares cabs? Monster!

Your cab's environment should be pretty thermally stable. Your switches are probably venting to the hot isle through their front ports. Thermal expansion of glass and the sleeve is going to be negligible over a metre or so.

Kinking is a possible issue. The minimum radius these days is quite tight. However, if you don't leave enough space inside the doors to allow for the terminations it will go horribly wrong. If you don't allow a gap for cable management between all switches and top and bottom, it might go wrong.

john01dav 12/28/2025||

I think that the person to whom you replied is speaking of outdoor installations, while you are speaking of controlled (maybe datacenter) installations. I have outdoor fiber running aerially between buildings on my property, in a region with massive seasonal temperature changes. Multiple local FTTH and Coaxial ISPs also run fiber on shared utility poles (the same ones that the electrical grid maintains) and when I look at the poles I see communications lines all in the same general area, often mere centimeters apart, if that.

summa_tech 12/27/2025|||

Can always go splice some PCF or PMF if you like to feel appreciated for your splicing. I swear I'd rather splice 100 SM fibers than 1 PMF.

vasco 12/27/2025||

And avoiding the NSA submarine taps!

candiddevmike 12/27/2025||

I was a fiber installer once upon a time in the 00s. A guy I worked with who was "the splicer" for our team and has years of experience using the little easy bake oven thing swore by going off the smell to know when it's "ready". Probably not the greatest thing for your health considering he did at least 10-12 of these a day.

bcrl 12/27/2025||

The old manual tools were extremely slow. Modern fibre splicers mean that a dozen fibres can be spliced in maybe a bit more half an hour, although cable prep cam take a significant amount of time depending on the cable type, number of cables and splice closure. Even more if you're using a ribbon splicer that fuses 12 fibres per burn.

Sesse__ 12/27/2025||

Modern fiber splicers are fully automatic, so you don't need to smell :-) The only thing that's still mostly by hand is the cutting (mostly stripping away the various layers of insulation).

duskwuff 12/27/2025|||

Modern fusion splicers are also shockingly cheap. You can get one for under $600.

Sesse__ 12/27/2025|||

Yup, I have one from AliExpress :-) Only done ~50 splices or so, though, but works really well.

dboreham 12/27/2025|||

Plus tariff.

duskwuff 12/28/2025||

Under $600 delivered from Amazon, e.g. https://www.amazon.com/dp/B0D6R91Y8F

Hikikomori 12/27/2025|||

Even 20 years ago when i was learning they were automatic.

ndom91 12/27/2025||

This would be a perfect plugin for Netbox! (https://github.com/netbox-community/netbox)

Already__Taken 12/27/2025|

I've always had stuff like this turned down by Netbox, they argue they want to model the logical topology as a source to trust, not the physicality, but then they model rack U placement. I'm always puzzled by their stance.

Like you can't model 1 cat5 split into two 100mb terminations, patch panels are kinda of hack, I think you can now but forever you couldn't just swap a termination direction because logically why would you (but their UI gets messy when 44 are done A-B and the 45th B-A)

Anyway that's thoughts as of maybe v2 or 3? Before the new UI when it was all jquery.

dfc 12/27/2025|||

Netbox project used to go on and on about the philosophical justifications for not including n-type connections or different types of LMR. But the most recent release notes that I read had a blurb about all the new coax cable types they are supporting. I understand having limited time but instead of saying "no" they always had to make lofty philosophical arguments. It's weird.

Already__Taken 12/27/2025||

Honestly that's fine I'm just glad I'm not crazy.

wkat4242 12/27/2025||||

> Like you can't model 1 cat5 split into two 100mb terminations

Ugh I don't really blame them there, that's really a dirty hack. Sure I've done in a pinch but not for permanent stuff.

I wouldn't call that professional network management. If you really wanna do it, just split the pairs over two patch ports IMO.

bc569a80a344f9c 12/27/2025|||

One of the achievements in my career I’m lowkey proudest of is sneaking in the rewire of about 45,000 ports on a campus that were split pair after an explicit project to do so was shot down.

matt-p 12/27/2025||||

Of course, but a splitter in a PON network or a WDM device are perhaps better examples of things that are hacky to model. Multi-fibre cables and splices are another. Netbox is great for some simple applications, and it's fantastic OSS, but in practice falls short for many use cases.

wkat4242 12/27/2025||

I understand, my cabling OCD got a bit triggered, sorry :)

Already__Taken 12/27/2025||||

Yeh it's awful, but all of our CCTV was wired like this through patch panels with 24v/48v power injectors. 2 cameras a cable. So that's what I needed to document, because in reality I can't book scaffolding and change rooftop cameras for a documentation tool.

toast0 12/27/2025|||

> Ugh I don't really blame them there, that's really a dirty hack.

I certainly wouldn't do it today, but using two pair for a connection designed for two pair isn't a dirty hack, it's as designed.

Today, using 4 pair for 1G or more and a small switch on the host side to get more ports is probably a better plan.

wkat4242 12/27/2025||

Oh I wasn't aware of this actually being an intended usecase. And yes like the other poster said, pairing it with a phone infrastructure was more common (in the days before these went all IP of course).

It was a bit of my OCD being triggered as well. I love neat cabling at work (at home it is chaos funnily enough).

pbh101 12/27/2025|||

Any links to PRs or discussions?

erinnh 12/27/2025||

Theres many.

Here is one Discussion/issue that is currently annoying me again.

https://github.com/netbox-community/netbox/discussions/9515 https://github.com/netbox-community/netbox/issues/20005

Netbox is full of these kinds of things. Where people ask for stuff or even create PRs for it and the Maintainer of Netbox shoots it down because reason.

hju22_-3 12/29/2025||

Wow, that second one was just straight up mean spirited. Doesn't feel like the responder even read it, given the comment about needing a discussion at the end despite it being linked. Not even mentioning the frankly unrealistic expectations of quality is annoying when the contribution guide didn't have those kinds of insane requirements then and still doesn't now.

zwnow 12/27/2025||

We were splicing some fiberglass in job training a few years back and it was honestly pretty cool! The website is also really nice, I remember seeing the color codes on the splicing machine. Mesmerizing piece of technology.

matt-p 12/27/2025|

Definitely mesmerising the first time! We have ribbon fibre these days as well which is very cool too.

Thank you :)

Prcmaker 12/27/2025||

In an old job I spliced a lot of fibre for our research, got the hang of it and got pretty proficient.

My research shifted, and I start making sensors using fibre and similarly sized capillary tube. It all got cleaved, spliced to assemble, cleaved some more, spliced again, then polished to spec. Getting the he process down was a pain, but the results were super satisfying.

kotaKat 12/27/2025||

Are there a set of controls I'm missing? I'm confused on how to work with this.

matt-p 12/27/2025|

Click a unused fibre and drag it to an empty 'splice holder' then do the same on the other side. You can also use double tap on mobile I think.

Sorry! In practice manual usage is normally very rare, these are typically auto generated!

crote 12/28/2025||

In case you are interested in improving the drag-and-drop usage:

There are a lot of connections which look like they should work, but don't - such as patch panel to patch panel. On the other hand, it is possible to connect a single LC connector to two splices at once. There doesn't seem to be any way to unmake a connection. Overriding existing connections sometimes works, but sometimes doesn't. The tube nodes don't have a clear drag-and-drop point towards a cable. By default the tube nodes have a colored edge matching the incoming tube connection, but this edge color doesn't change when you drag a different tube from the cable onto it. The cursor sometimes wrongly shows a "+" to indicate a possible connection, such as trying to drag from one tube exit to another. It is somehow possible to connect two cable tube exits to a single tube node. It is somehow possible to connect a cable tube exit directly to a splice node.

Those probably don't matter all that much if it is almost always auto-generated, but considering the amount of weird behavior it might perhaps be better to disable the edit functionality for now?

richvos 12/27/2025||

This looks really nifty, matt. I might be in the minority here, but would you consider making a similar model for PBX‘s? I’m having the hardest time visualizing all of the phone lines I have to deal with. This could be a really nice tool.

dtgriscom 12/27/2025||

Are fiber splices really only a 0.02dB drop? That only a 0.23% reduction in signal (if I have my math right). Impressive.

fsh 12/27/2025||

Fiber splicers are marvels of technology. They align the fiber cores with sub-micrometer accuracy and produce just the right amount of heat and pressure to melt the ends together. They are also usually very rugged, fully automated, and surprisingly cheap (a few thousand euros). It is remarkable what is possible when the entire internet relies on a technology.

matt-p 12/27/2025||

That is probably the very best case scenario, but possible yes. Typically you'd accept anything less than 0.1dB.

cycomanic 12/27/2025||

I did my PhD on fibre lasers, 0.1 DB would have been considered a ver bad splice and I would have recut and respliced (if you have 1-10W in your cavity that 0.1 dB loss would risk burning and the fuse propagating through your cavity destroying everything in its path (as a side not look up Videos of fibre fuse, looks fascinating). In my experience 0.01-0.02 is much more typical than 0.1 dB loss.

matt-p 12/27/2025||

I’m speaking mainly within the context of telecom field splicing - the numbers I mentioned are typical for that application in my experience. You’re only sending on the order of 5 mW down a fiber, so none of those high-power concerns apply. Obviously, different networks have different thresholds: if you’re building a greenfield, low-latency long-haul route, you want to minimize loss and it’s reasonable to spend the extra time and use higher-end equipment. For FTTH, with something like a 30 dB overall budget, nobody really cares whether a splice is 0.03 dB or 0.1 dB.

Hikikomori 12/27/2025||

You might accept a bad splice but you'd almost have to fuck it up on purpose with a decent automatic splicer.

maartenh 12/27/2025||

Nice! You might want to fix your GitHub link in the footer though, it 404's for me right now :)

matt-p 12/27/2025|

Thanks! Sorry, looks like I made the repo private at some point I'll take a look later but for now I've fixed the link.

eru 12/27/2025|

What am I supposed to do?