[Betatester548]

Hi,

Thanks very much for your help. I downloaded the latest firmware on the DSL-N16. It did help with the packet discards.

But, I think we decided to just buy a new router in the end, as some of these offer some very useful features like SQM.

I think we will use the DSL-N16 in bridge mode. This allows the device to act as a modem only, is that correct?

I think all I need to do is set the device to ‘Bridge’ mode, like this:

After, that I think you just connect the 1st LAN port on the DSL-N16 to the WAN port on the new router?

In effect, the modem handles the DSL part of the connection and the router handles the actual internet connection to an ISP (PPPoE in this case)?

Presumably, I should also disable the DHCP server on the DSL-N16? And change the IP address of the device?

The router we’ve brought is a used Asus RT-AX56U (should arrive in a few days), for £65. It has a quad core 1.5ghz CPU and supports SQM with Asuswrt-Merlin.

You need to login in order to vote

[Betatester548]

The difficult part to work out is, if it my line that is causing the ‘outDiscards’ when uploading files, or, is it the Asus DSL-N16 router?

I tried enabling bitswap, but my router still reports packets being discarded.

I wonder if it has anything to do with firmware?

You need to login in order to vote

[Betatester548]

Hi, is it normal for a router to discard packets?

I’ve been checking this with this telnet command on my router:

tcapi show Info_Adsl

The number of ‘outDiscards’ on my router has gone up a lot since I started uploading a large 10GB file to Google Drive.

Its going up by hundreds every minute, but generally only when files are being uploaded.

Note – I also noticed that there are just over 2 billion FEC downstream errors after about 2 days connection uptime.

 

You need to login in order to vote

[Betatester548]

I have a question about packet loss, do CRC errors lead to packet loss?

Or are these errored packets discarded by the router, before this happens (resulting in a slightly reduced bit rate).

You need to login in order to vote

[Betatester548]

0.001% packet loss on Stadia now after 20 mins, it’s looking good. That miniscule amount of packet loss was probably caused by local network congestion.

Liked by:

You need to login in order to vote

[Betatester548]

Indeed I will, I know it’s important not to fiddle with stuff that already appears to be working quite well.

There’s also the possibility that that upstream SNR strength has been boosted (I’ve no way to to check as my upstream SNR isn’t plotted on my Asus router. It might explain why I’m not getting packet loss on Stadia now, at least at 1080p.

 

Liked by:

You need to login in order to vote

[Betatester548]

156 upstream CRC errors in 3 hours 30 mins. I suppose the line needs to settle again, but hasn’t effected the performance.

You need to login in order to vote

[Betatester548]

Yeah, nice of them to arrange a visit (should be with no charge I think).

I also think my line may be packet loss free now :)

I tried it at 1080p in stadia with no packet loss for 11 minutes. Will try it for longer periods, but suspect its fine now.

I noticed that the power output has  increased on both the downstream and upstream, according to my router. By about 1dbm for both. I’m unsure if this indicates they made any changes or, if this is related to the line DLM being reset.

Lots of FEC errors reported when I telnet to the router, but maybe these don’t matter?

You need to login in order to vote

[Betatester548]

So, a BT engineer showed up today, due to a query sent to my ISP. They decided to randomly send an engineer to our house, who had a look at the master socket and replaced the front plate. He did a line test after this, his device said ‘line fault’ a couple of times.

There was another engineer outside who did some work on the cabinet, which was apparently not where I thought it would be, at the bottom of my street, but maybe a bit closer.

I asked the engineer who did the work on our master socket if they’d improved the line (I asked if the SNR had been improved), he said yes. They looked like they were going so didn’t ask them anymore questions.

Interestingly, even though the line was reset, there is still interleaving on the line.

Then I checked the line, but same line stats (attenuation and SNR graph the same), and Mlab’s test is still reporting 0.5% packet loss or more on the line.

It looks like the BT broadband engineer just reset the line at the cabinet (line stats reset and SNR margin now at 6DB), wonder if I can find out what was actually done.

 

You need to login in order to vote

[Betatester548]

My dad isn’t willing to have a BT engineer around to check our line. I think he is worried the engineer will blame us for changing the master socket to an NTE5C, from the much older 2/1a BT type master socket, and charge us as a result.

The funny thing about this is, there’s really only 2 wires to connect (excluding the telephone wires on the front part, which we know aren’t too blame) and little to go wrong when connecting these wires with an NTE5C.

Is it actually against BT’s rules to rewire /upgrade the master socket? Even if it was done correctly?

You need to login in order to vote

[Betatester548]

Thanks for the advice.

Do loose connections happen a lot at FTTC street cabinets then? Or, would this generally only happen occasionally?

Realistically, can a BT Openreach engineer do any other work, other than check the street cabinet?

Do they ever resolve street wiring issues, or would that be too time consuming?

You need to login in order to vote

[Betatester548]

Do you think if I contacted my ISP, and said that I think there is a fault with the line, they would send a BT engineer, who would then charge us for any repairs / work at the cabinet or elsewhere?

You need to login in order to vote

[Betatester548]

I checked my line at BT’s wholesale line checker here:

https://www.broadbandchecker.btwholesale.com/#/ADSL

Here are the results:

It says the downstream handback threshold for my line is either 44.8mbps or 50.7mbps.

My line is currently only syncing at 32.3mbps, with the SNR set to 9dB. When it was at 6db, it synced around 40mbps.

We were worried that we might’ve left the white and orange wires (network) a bit loose on the back of the master socket, so we re-inserted those into the cam locks. But the result was the same, so it looks like there is a fault with the line.

You need to login in order to vote

[Betatester548]

No luck I’m afraid. Tried the test socket on a couple of routers, still getting some packet loss on both. It’s likely to be the upstream that’s the problem, as limiting the downstream with QOS didn’t reduce the amount of packet loss.

The upstream line attenuation was a little better on the TP-Link VR200.

Overall, I think it’s just an unreliable line, but at least there are no disconnections.

We are in a ‘white area’ (eligible for subsidy for internet access) for Project Gigabit, so hopefully by 2024 we will have FTTP, which should avoid any line issues.

You need to login in order to vote

[Betatester548]

I think we might take another look at the network wiring on the master socket, there’s a pretty helpful video here that shows this being done:

Liked by:

You need to login in order to vote

[Author]

Posts