NBN fixed wireless

Edit: New readers to the topic can see the latest experiences as of May 2022 by going directly to post 185.

I rely on NBN fixed wireless, so this hits close to home.

NBN Co considers a cell to be officially congested when its “30 day … average busy hour download throughput [falls] below … 6Mbps" per user.

In view of the impact of RSP provisioning on performance for the end-user, I wonder how they measure that? Note that it’s an average, so an individual service could be consistently slower than 6 Mb/s.

But it can be really fast:

According to NBN corporate public affairs manager Tony Brown, only 20 percent of NBN’s fixed-wireless towers are connected by fibre, meaning the 1Gbps speeds would only be available for a fraction of fixed-wireless users – those with fibre-connected towers, and with 11 carriers available in both the 3.4GHz and 2.3GHz spectrum bands.

With the world moving toward 10 Gb/s services, how long will it be before Australia hits a wall?


Should be a great topic and resource @n3m0 for those on or about to be connected to FW.

The NBN Co holds it’s performance data close to it’s chest. There are so many ways to present stats and skew results or avoid difficult questions. It does not share the source data, or offer even summary stats by tower or location. It’s about more than just congestion, it is also about the maximum reliable connection speed at each premise for each tower site. It’s also as mentioned about the limited backhaul due to the cost cutting microwave links used for most towers, and so much else.

I read one report that suggested personal data consumption in Australia has gone up ten fold over the past 5 years. For FW, and also Satellite it would be more than essential for the community to be told just how close both services are to capacity, The NBN may talk about future technology, however after 6 years they are still another two from finishing even the current plan. It seems an unlikely outcome that FW will be any further improved in 5 years time if the substantial financial commitments needed to deliver this work are not already in place.


A little more grist for the mill:

Under the design rules:

  • the 2017 iteration allowed for “up to” 2640 premises to share “up to” 900 Mb/s;
  • the 2018 iteration changed that to the same number of premises (or maybe less, depending on interpretation), sharing “up to” 4Gb/s. but;
  1. the design rules document is inconsistent and open to interpretation;
  2. the fixed wireless network is largely complete and
  3. the new rules are not retrospective.

Links to current and archived NBN design rules documents can be found here:


You have to wonder what the real facts from the NBN Co are when reading the following from the linked IT News item, July 2018.

NBN Co said this week that 99.92 percent of premises outside major urban areas are either in design, under construction or complete. That means the vast majority of the fixed wireless network has been deployed. - end of quote.

How the term “In Design” could ever mean in an Engineering Sense “ Network Deployed” does Hansard hold the answer?

‘Nearly Finished’ or ‘yes, we are up to finishing that one’ are similar but more openly evasive routine project responses?

I could suggest that 0.08% is only one tower remaining if there are 1200 tower sites planned for FW. I have not yet found a source that adds them all up?

To allay fears about any ambiguity in how to read the NBN press release from last year.

Per the NBN Co 2019-2022 Business Plan there will only ever be 600,000 premises serviced by NBN FW. Might be rounded to the nearest?

Per the NBN Co weekly progress report (03 Jan 2019) fixed wireless reports 633,395 premises Ready to Connect. That’s premises with an active service available if I read the data correct? Just needs a receiver and modem install at the premises to complete connection when the owner elects?

By fact the NBN Co reported at the end of June 2018 last year FW Ready to Connect covered 609,000 premises! Per the July press release it would leave (0.08%) or only 500 premises nation wide without the promised FW available?

However another 24,000 have been added since then? So Ready to Connect must have a different meaning to the literal one? :upside_down_face:

Which data is reliable?
Which stats are correct?
Who is being honest here?

It would aid transparency to know just how many FW sites are currently not active or providing full service, and exactly how many premises are genuinely unable to connect? :rage:
IE How much left to complete the FW solution?

That’s a minor aside perhaps to having better information on the actual backhaul capacity vs demands for each site? :wink:

P.S. There have been various predictions going back over 24 months FW would be stretched to more than 700,000+ premises? Alternately are those in the never never land heading for an alternate solution?


Most FW NBN customers live outside the cities i.e. in the country. These folks lived in the dark ages in relation to internet and mobile services for so long that most would have welcomed the change to FW. Country people are normally easy to please but they do not like to be fooled with statements that all’s well or as good as can be.
We feel the congestion every weekend especially on long weekends when the download speed often goes down near nothing but who are we to complain as we know that it will be better again in a day or so (thanks NBN Co!). Look at other countries, often 3rd world countries, that have better internet and mobile services than Australia (thanks Australian Government!).


Speaking as one of them, country people are none too pleased that Telstra neglects copper 'phone lines. We were told that they’re too expensive to maintain. Telstra gets (IIRC) $300 million per annum for the Universal Service Obligation. That was one of the carrots John Howard dangled to get agreement to privatise the network. We really should take it back (and I do mean “take”).

1 Like

FW is being provided a lot in non Rural locations as the FTTN service can be “too difficult” to provide when on certain copper circuits they are too far from the node to maintain 25/5. Mr Morrow acknowledged this back in 2017. FTTC was the next fix to be trotted out to help address this and also to plug the hemorrhage that is HFC services.

Even Satellite has been acknowledged as a service that is too commonly used to fill gaps so that they don’t need tp address FTTN, HFC and FW failures/lack of service/poor service.


Not surprising. What government often dresses up as support, the recipient sees and accepts as a donation. Since government never reminds them of the support/donation purpose, it remains a donation. Does that ring true with your observation?


I gave up in disgust, years ago, but that’s about right. Telstra accepts our funding and avoids doing what they’re being funded to do. For me, the last straw was when Telstra’s excuse for not maintaining a certain line was that it wasn’t “commercially viable”. Why the :face_with_symbols_over_mouth: do they think they’re being paid?

Telecommunications is a vital service. If the private sector (Telstra) will only bother with the profitable bits, then we (the people) should take it all back. I gather that we can’t just do that. The Constitution says something about due compensation. I’m sure that, with a little creative accounting (offsetting the damage they’ve done against the value of the network), we could end up presenting them with a bill in place of compensation.


13 posts were split to a new topic: Privatisation or government ownership of NBN and telecommunications - what’s best?

Communications Alliance argues that fixed wireless performs so differently to fixed line services that the two can’t be measured in the same way.

NBN Co has been under pressure over the past year to fix congestion issues on its fixed wireless network, issues that users say has seen them achieve a fraction of anticipated speeds.

NBN Co will ultimately stop selling speed-tiered fixed wireless services and instead offer a “best effort” service.

So; “Crap, it might be, but it’s the best we can (be bothered to) do. Suck it up!:face_with_symbols_over_mouth:

If this is anything to go by, then 2 Mb/s is acceptable on any speed tier.


Another joyless online doc from Telstra with their assessment of FW. I was able to post in Dec on the “Never Never Broadband”.

It may be worth keeping the Telstra lack of commitment in the one place. Same horrific 2Mbps minimum on all speed tiers, but an unclear promise on the 50Mbps speed tier at peak times of achieving 20Mbps. Published Nov 18!

I’ve just come back from a chat with our Federal Govt MP (LNP) and am none the wiser on where the NBN is going after the next election assuming the current govt is returned. For now they either they don’t know themselves or avoiding it in the hope it will go away?


The following interview with the NBN per ZDNET middle of last year provides two alternate view points from the NBN Co.

The following statement from the NBN Co suggests that it has never been capable of delivering even 2Mbps average per connected customer at many sites due to limitations on backhaul over a 900Mbps link.

“The maximum bandwidth planned for the microwave hub site back to a FAN [fibre access node] site has been 900Mbps, but is now moving to 4Gbps to support capacity growth, allowing for the aggregation of up to eight eNodeBs [base stations], with a maximum of 2,640 end users.”

You need to do some maths and read the full article. The 2,640 end users on a hub site is a target half of the original design spec, hence there may be sites with more users. To achieve the NBN minimum service standard of 6Mbps at peak time that the NBN Co was committing to at that time with a 900Mbps backhaul there is only sufficient capacity for approx 150 concurrent users at that level of demand. Even with the upgrade to 4Gbps it only provides for approx 666 concurrent users?

The 900 Mbps capacity provides for approx 450 concurrent users out of 2,640 at 2Mbps average. Accepted this is all an over simplification as there are network overheads and other limitations. Data is also accessed in bursts when browsing, however streaming, downloading updates, Utub (Sorry - Youtube) etc require a steady data feed.

Riordan also described two ways in which NBN could improve visibility over how it is tracking and addressing its fixed-wireless network, including publishing which cells are congested.

“Commence in their dashboard reporting or similar how they’re tracking in addressing the cells which don’t currently qualify or meet their own design standard for the 6 megabit per second per end user in the busy hour – that would be one simple step which they could take, and that would at least give people an informed information base to make some policy and other decisions,” he suggested.

The 6 Mbps standard is far from being acceptable and is not consistent with the original NBN promise of 25Mbps minimum speed of service. It is however 3 times faster than the minimum even Telstra is prepared to offer on their fastest NBN FW offer at present.

FW even with improvements looks to have minimum standards that barely match many customers previous ADSL based services. In particular for those who now have an ADSL exchange connected to fibre eliminating physical backhaul limitations, being pushed over to a less than adequate and potentialy more expensive NBN seems very very wrong!


Meanwhile, back in the world that some of us think might be real:

Technical documentation shows that the rebate applies to 25-100/5-40Mbps services, 100/40Mbps, 250/100Mbps, 500/200Mbps and 1000/400Mbps services.


As a brief distraction, how many may have seen this Telstra promotional on the side of a near extinct public phone box?

Do they have a sense of humour or was this another inspired moment that others might interpret in unexpected ways?

On the drive for more NBN customers it fits nicely into the not too distant vision of the NBN Co readying the business for sale. The fundamentals still look poor, however without the FW and Skymuster it may only a function of getting the income stream to improve a little more. Best saved for the topics on selling the NBN and public vs private ownership, it may explain a current low level of NBN and govt interest in FW customers?


I have not seen that one but the Telstra public phone near our local burger and fish shop has a classic.

“A network faster than seagulls on chips”.



This might be of some interest. I’ve been playing around with speedtest-cli. Setting up a cron job to record performance hourly proved easy enough. Sadly, each test consumes about 80 megabytes. During peak periods, that exceeds my data allowance. Anyway, I’ve graphed a day’s figures.

The drop-off in performance during the evening peak is pretty dramatic. Of course, whether that’s due to limitations of the NBN or of my RSP is impossible for me to tell.

The service is through NBN’s Mount View Road fixed wireless tower (ACMA site ID 9027703) in the Hunter Valley. The target is server #2629 (Telstra, Sydney).


“National Broadband Network” was always an oxymoron. I suggest that the term only exists for political reasons.

Perhaps. Can anybody deny that Australia needs telecommunications infrastructure that delivers equitable access to services?

The comment was about what “is”, not what “ought to be”.

Cost means that a certain amount of non-uniformity is unavoidable. However as low as “6 Mbps” may be unacceptable.