New install, disappointing speeds - could/should I expect better?

[#1]

Just had my old FTTC line upgraded to full fibre. I went for the top package which estimated my connection at 900Mb and a minimum speed guarantee of 500Mb. We're in a newly provisioned village and I am one of a handful of customers on the FTTP product in the village (B4RN are currently rolling out and most people are holding off - I'm having both!) which I would hope means I should expect close to the advertised speed. 

I'm currently connecting the ONT to my UDM-SE and on ethernet my speeds are fluctuating between 380-470Mb. I will run some tests using the hub2 later but a WiFi test I did was dire (less than 80Mb when stood right next to the hub). Not so much bothered about the wireless speeds as I have my own UniFi network, but the wired speeds seem unacceptably low to me.

Not sure if it matters / is a contributing factor, but we're served via a telegraph pole installation.

Anything I can do? Am I right in saying the old "training period" doesn't apply with FTTP?

[#2]

Did the engineer run a test before he left? That should have shown the full expected speed (you are right that there is no training period for FTTP). 

[#3]

He did some testing but didn't show/tell me anything other than "it's working".

[#4]

@djbenson 

It is the wired speed that matters as WiFi can be influenced by your own equipment.

You should be getting near full speed so it'll be interesting to see what the Hub2 speed is on a wired connection.

Funny that OR have installed FTTP in an area that B4RN are building in. They tend only to go where no one else is building as they need signicant properties to sign up. As a community owned company they don't borrow to invest.

Can I ask the villiage name just out of interest and is the B4RN build due to goverment money being available?

Brian

[#5]

I assumed the WiFi would be poor because my own kit would be competing with it.

I've sent you a DM regarding the B4RN project.

[#6]

@djbenson 

Does your current router have QoS enabled, I would disabling it if so.

For my Full Fibre connection I see on a Hub two Download: 923.86 Mbps / Upload: 106.66 Mbps

A wireless speed of around 80 Mbps would suggest a 2.4GHz connection.

A wired speed test using the Hub two would be best as plusnet would prefer their router for speed investigations.

[#7]

@djbenson 

Thanks for the PM. If I'm not mistaken the village went "live" on the B4RN network very recently.

Given the time it takes B4RN and the local volunteers to set up a project I would guess it's OR who are overbuilding the area. The difference is that B4RN will cover every property in the project area no matter how difficult or distance involved. OR will cherry rick.

Look forward to the Hub2 results.

Brian

[#8]

Yes it went live within the last week, I get mine next Wednesday. I keep two connections as I WFH and need a second connection which is on a different platform, so historically I was FTTC+5G and now will be BT-FTTP+B4RN.

Just need to work out the speed issues!

[#9]

This was the first thing which sprung to mind so I've disabled everything I can think of;

• QoS (probably don't need this any more)
• IDP
• IDS
• Smart Queues (this should be disabled anyway for faster connections)

[#10]

Soup to nuts consideration for FTTP speed concerns...

FTTP data test speeds not as expected

There have been numerous reports of measured data speeds not fulfilling users' expectations of their shiny new FTTP service. An end-to-end data speed test is a somewhat blunt instrument with which to test the connection speed. In most instances the FTTP service is running at the product's connection speed, but the user's environment or the data speed test service cannot make full use of (saturate) the FTTP connection's bandwidth.

Users can no longer have confidence that the cause of slower-than-expected data transfer speed is due to the broadband connection's performance. If there is any doubt that the user's technology, infrastructure or chosen test service cannot make full use of the FTTP link's bandwidth, then a data speed test is likely to deliver misleading results.

Here are a few things to consider / investigate...

1. Is the ethernet cable between the router and the ONT (Optical Network Terminator) CAT5e or better?
   CAT5 or below will not transfer data faster than 100mbps thus a simple cable could be the cause of slow data speeds. The ethernet link rate between hub and ONT can be observed in the Hub Two GUI under 'Status'. This is a good indicator of router <> ONT cable function.
    
2. Is the device on which the data speed test is being run connected over (CAT5e) ethernet cable (not WiFi)?
   A device connected over 2.4GHz WiFi absolutely will not get anywhere close to 300mbps. It will typically be less than 100mbps. Typical 'good' 5GHz performance using the Hub Two is likely to be below 500mbps and will get lower the further away the device is from the router. More optimal Wi-Fi performance is possible but involves sourcing your own third-party equipment that is capable of the newer standards. Also Wi-Fi performance is equally dependent on the type of device and the age of its technology standards. Some older devices do not support 5GHz, even the most recent smartphones only have a 2x2 antenna array, which limits the number of concurrent Wi-Fi streams. The Plusnet Hub Two can use more streams (go faster), but most types of device cannot.
    
3. Is the device on which the data speed test is being run connected over (CAT5e) ethernet cable with no intermediary devices (like Powerline adapters, switches and physical ports) on the LAN side?
   Irrespective of cabling, if there is a 10/100 ethernet switch in the chain then it is going to bottleneck things. Similarly Powerline adapters are highly unlikely to be able to pull anywhere close to 300mbps through electrical wiring. Many Powerline adapters that are rated at e.g. 500mbps, actually have only 10/100 ethernet ports which means the advertised speeds are near fiction.
    
4. There are no other devices active on the network at the time of the test.
   Other concurrent use of the network will reduce the bandwidth available for the data speed test, thus delivering a misleading result.
    
5. If on a static IP or using the Plusnet firewall (configured in the user portal) has the Plusnet IP profile updated?
   Data speed tests performed soon after the service has been connected (but before Plusnet's systems have updated) might be constrained by the account's line speed profile. Disconnect the router from the ONT, wait a short while, reconnect then check that the Current Line Speed (Download) reported in PN Broadband Status reflects the FTTP product's profile.
    
6. If using a third party router ensure that QoS (Quality of Service) and other traffic shaping tools are disabled.
   QoS and other tools have been seen to cause havoc with attempting to saturate FTTP links. Prioritisation and QoS features are very processor intensive and can (probably will) significantly bottleneck speeds. Notably there have been numerous reports of Netgear Orbi maxing out at 550mbps when connected direct to the ONT, rather than being connected to a Plusnet Hub Two.
    
7. Is the test device configured to use any VPN or corporate intranet connection?
   VPNs and corporate intranets can also be potential bottlenecks.
    
8. Is the data speed test device capable of moving the data at FTTP speeds?
   Some kit simply does not have the power to make full use of FTTP connection speeds. Whilst undertaking data speed tests, if running on MS Windows, use Task Manager's performance tab to check throughput limitations do not arise from...
   • CPU performance
   • Disk read/write speeds - FTTP will easily outpace a 5400 rpm disk
   • Memory utilisation - the data has to go somewhere!
   • Network performance - unless the machine is reasonably new, the ethernet port will be limited to 100mbps!!
      
9. Speed test service
   Not all speed test services are dependable nor consistent, so run some comparable tests on different speed test services, at different times of day and night (both on and off-peak) where possible. The following are good points of reference.
   • Think Broadband who also offer a number of download files to soak test connections.
   • Ookla aka SpeedTest.net who also have a client-side speed test application which is preferable.

Please check out the above factors, to ensure that the perceived data speed is not compromised by some other factor limiting the data throughput-speed of your full-fibre FTTP.

~~~~~~~~~~~~

A user has observed that different devices ran at different speeds and their problem was discovered to be the LAN card settings on their Laptop...

My PlayStation was connecting at around 535mbs all the time, but the laptop was only running about 130mbps max. On checking the internet settings I found that speed and duplex were set to auto. I changed this to 1 Gigabit and now the test speed is 838mbps.

To find the settings for your ethernet adapter, go to advanced after selecting properties.

[#11]

If it's not considered rude or big-headed to say so, those are great tips for a beginner but I think I've covered them off before even asking.

Just to cover all bases;

1. The ONT is connected to my UDM-SE using CAT6
2. I'm discounting WiFi - these are pure wired tests
3. As per 1 it's a direct connection between the UDM-SE and the ONT
4. Network usage is minimal (<1Mbps currently)
5. My portal hasn't updated yet as it seems the contract hasn't updated to reflect switching from FTTC to FTTP but my static IP has switched over and I'm already seeing speeds in excess of FTTC so not sure this is relevant?
6. As above all traffic inspection/shaping is disabled - it made zero difference
7. I'm testing directly within the CLI of the UDM-SE
8. According to the specs it should be able to handle 1Gb WAN with ease (it's rated to 3.5Gb without IDS/IPS and around 850Mb with those enabled)
9. I'm using speedtest-cli tool on the UDM-SE which is distributed by Ookla.

[#12]

Hi, neither rude, nor big headed ... but experience has shown that there is wisdom in affirming that all of the usual culprits have been examined and eliminated.  On too many occasions when investigating IT system issues, I have assumed that the coal-face tech has checked the usual suspects ... only to find after days of wasted effort, they had not and the fault was one of the known issues.

This potentially brings us in the the space of muted issues with ONTs.  I recall reading around here of a user reporting that a BTOR engineer alleged widespread issues with ONTs ... but there's no substantive evidence to verify that.

[#13]

Thanks for taking my comment graciously, and I agree, many an issue leads back to something we've either assumed has been tried or we're too pig-headed to try in the first place 😂

Which leads me to this...

When I connect my Windows laptop directly to the PN router which is connected to the ONT, I get higher speeds. I did three tests and got;

• 550
• 850
• 900

So it's a variant of PEBKAC - something like Problem Exists Between ONT and User?!

Need to figure out where the bottleneck is because from that one test I can safely assume the connection to the ONT is fine. 

Thanks for your help people - I'll go away and try and find what I've missed.

[#14]

As a side question, when did BTOR start rolling out such tiny ONTs? The ones I've seen on YouTube etc. were massive - the one that's been installed is the size of a single gang UK socket and sits really nicely alongside my (to be activated) B4RN connection.

[#15]

@djbenson 

Can't comment on the smaller size ONT's but when I got mine in late 2017 I got a big 4 port ONT with a separate battery back up unit. A few weeks late a neighbour got the following installed. Can't remember if it's 4 port but still has the BBU.

Brian