Moving on to a better ISP, suggestions? (South of England)

[Zennon]

8.8.8.8 is a bad DNS to test it has known issues.

 

Test from bbc.co.uk at an interval of 2.5.

 

The only hop that matters is the round trip.

[TryHardUndies]

Twisted pair RJ45 to RJ11 turned up today. They were right about RJ45 being a better, snug fit in the socket!

 

I haven't seen any positive difference in performance though. Base latency to 8.8.8.8 is actually about 20% higher and line is jittery, more so than yesterday. Even first hop latency is bouncing all over the place, rather than the usual 1-1.4ms. Time to smash head off wall again.

 

Still getting connection issues even after transferring to BT, lightning does strike twice?!

 

Funnily enough I've also been getting packet loss on the Netduma diagnostic test since switching to BT, never had that problem with VM. Usually between 3 & 5 packets but it doesn't happen on every test. When it does happen I usually just reset the HH & Netduma & complete the test again a few times & usually find that it doesn't happen again.

 

With BT I'm now getting Exceptional, Exceptional & Exceptional ratings, whereas with VM I usually had Good, Exceptional & Exceptional. Ping on the test with VM was 16ms, BT was initially 16ms but seems to have reduced over time & I now get 13ms.

 

Connection feels fine when I play BO3, I've even been playing Core rather than Hardcore without any issues.

 

I'm interested in buying a twisted pair RJ45 to RJ11 that you & a few guys in this post have mentioned to see if it improves further.

[RL317]

Lol yeah pretty much mate. I'm running a test to BBC right now and while the jitter isn't as high as VM's, it's still up and down more than I've seen in other people's Infinity plots. Nothing else is using the line, and I've got that new DSL cable between the socket and Hub which is both twisted pair AND shorter than the flat one they provide (apparently shorter is better).

 

Tests to 8.8.8.8 were fine in the first week: 7ms and stable, no bouncing around. I was playing BO3 on a TV with something like 70ms input lag and still dominating. Even though aiming was a struggle because of that delay, I could hear the hit detection was crisp and registering instantly. Since getting my new monitor and trying again it's been playing EXACTLY like it did on Virgin, clearly behind the action, delayed or simply no hitmarkers at all in some cases, and just leaving games fuming again. IW was totally unplayable for me, and BO3 hasn't been much better.

 

I'd love to know what I'm missing or if I'm simply cursed because this is absurd. I watch my American pal stream on cable (which he has serious complaints about as he used to be on Verizon FiOS fibre), but his hit reg is like nothing I've seen on my screen before. It plays exactly like Ghosts did for two years for me, even with my VM line which gave me issues especially in the evening.

 

I went back to Ghosts to test the new connection a few days back and connections were great. Even running into Saudis/Kuwaitis on 150ms pings I had no issues. Now back on BO3 I'm just dying over and over to lies again.

 

I'd love to know what else I'm supposed to do to get those connections I've seen in maybe 2% of games on BO3, where bullets landed instantly and I didn't see WTF moments every 30 seconds...

[ColonicBoom]

Have you changed something since you got it installed? It looked great at the beginning.

 

Are you setup like this:

BT hub > Netduma > Devices ?

 

Is the computer that you are testing from wired to the Duma (assuming that you have the above setup) ?

 

That packet loss at the first hop is not good, have you tried a different ethernet cable between the laptop and the Duma AND between the Duma and the BT hub?

[RL317]

I'm not even using the Netduma at the moment mate; I still need to order a new one (think I'll do that either today or tomorrow). It was great at first and now it's not the same

[ColonicBoom]

Was it better before you switched the cables?

 

Just check the ends are seated nicely - pull them out and click them back in.

If it still doesn't seem right try the previous cables again obviously.

 

I find that without my congestion control set (and yours obviously isn't without your Netduma in there) that phones and things would still communicate with the router periodically, like email fetch etc, which would cause spikes when they communicated.

 

I haven't played BO3 for a little while so can't say whether that's playing ok.

But I can confirm that I didn't get a single smooth game of IW. Most days I didn't get a single game of IW at all though.

[Zennon]

Are the test being ran with only the console online?, if anything else is on and using the line it will create jitter without a Netduma using congestion control or a QOS of some sort.

[RL317]

Was it better before you switched the cables?

 

Just check the ends are seated nicely - pull them out and click them back in.

If it still doesn't seem right try the previous cables again obviously.

 

I find that without my congestion control set (and yours obviously isn't without your Netduma in there) that phones and things would still communicate with the router periodically, like email fetch etc, which would cause spikes when they communicated.

 

I haven't played BO3 for a little while so can't say whether that's playing ok.

But I can confirm that I didn't get a single smooth game of IW. Most days I didn't get a single game of IW at all though.

It was about the same before switching cables. On the original cable it was fine for a week or so until my monitor turned up. On the same cable it went back to the same old crap, and it's been about the same since on the new cable.

 

Funny enough in the first week I was testing WiFi, powerline adapters AND wiring directly to the router and it was all flawless. My first nuclear back online was on wifi (with the router downstairs), on my 16" laggy TV, without fast hands or quickdraw, just decimating. Even WiFi was incredibly stable, and the only difference on a Pingplotter test was a little extra latency on the first hop as to be expected (around 5ms).

 

I checked and rechecked all cables (and tried new ones), checking they were seated snugly. The twisted pair one is also shorter than the flat RJ11 I used from the box which is only going to improve things.

 

Plot twist though: I ran two tests last night, one to Google DNS and another to the BBC website, and both were stable. Latency barely moved at all, latency to the first hop wasn't fluctuating, using the same new cable too. However on both tests, on the 2.5 sec interval, I still saw a few random red blocks of 100% packet loss. I repeated both tests, same cables, on 2.5 sec again and there was no packet loss.

[RL317]

Are the test being ran with only the console online?, if anything else is on and using the line it will create jitter without a Netduma using congestion control or a QOS of some sort.

All tests are done with nothing else using the line. Devices disconnected or wifi disabled, wired directly to the Hub before testing on computer.

 

Sometimes it's stable, sometimes there are spikes despite literally nothing else using the line :s

[Zennon]

Can you post the new plots so we can judge the spikes.

[RL317]

They're pretty much identical to the last one I post: fairly stable along the bottom of the plot for the most part, and then random spikes up to 80ms or so before dropping straight back down. They're not frequent (ie not showing often in that 10 minute or so window that's visible at once) but they pop up randomly.

 

Then I could restart the test and it won't stray from 8-11ms at all over the course of half an hour or so.

[RL317]

Here's one for you Zennon. I haven't switched DSL cables since my bipolar tests yesterday, and I haven't changed anything in terms of settings. I unplugged my old Ethernet (which as far as I could tell was working just fine, I've tested with it) and plugged in a 10m cat5e foil shielded cable instead.

 

BO3 is playing like a dream past midnight. No WTF moments, no raging, no leaving games. Bullets are registering crisply and I'm having fun. Granted it's not accurate anyway because it's inflated by some kind of additional delay, but the in-game ping graph isn't doing anything different.

 

Is FTP considerably better than UTP? Or could that cable simply have been giving me inconsistent performance?

[Zennon]

Hi you have not posted a picture but it sounds like switching cables has helped?

 

Cables can be bad I see it all the time at work one day they work next day they start acting up, the copper in them can snap with movement it seems.

[RL317]

That's why I'm confused mate. Using my shorter, newer Ethernet on Pingplotter tests, there seems to be no difference? The cause of variation in test results is hard to pinpoint because I can get bad results using one set of cables and then stable results using the same cables again.

 

This one is a huge length of FTP Ethernet; too long for what I need actually, yet if anything performance is better, not worse. Then again Ethernet doesn't really cause performance loss as quickly as DSL cable does it? Hence short DSL cable > modem/router close to socket > Ethernet everywhere else as long as you need it.

 

The irony of it all is this 10m cable is damaged with the plastic coating showing through in some parts haha. Maybe I should order another of the exact same cables, it only cost me £3 or so. Then I think I'll start tidying it up with cable clips up the stairs as powerline adapters are bipolar in this house and WiFi is obviously a big no no for serious gaming.

 

Could the foil shielding really be making a difference? Or is my newer, shorter Belkin cable simply not performing? Maybe I just got a bad one, but it's odd because it gives me good results on Pingplotter tests too.

 

Will run some more tests tonight when the line is quiet and try to remember to bring the screenshots here with me!

[RL317]

Oh and just another quick noob question: my new twisted pair cable from socket to router is cat6, but my ethernet is cat5e. Would getting cat6 ethernet reduce interference or generally improve performance?

[Zennon]

Time of day matters if the ISP is congested , you would notice more congestion if it is that at peak times so 6pm to 11 pm say.

 

You should be fine with either 5e or 6 in a house environment.

[RL317]

Gonna test it again now (6pm onwards). But before BO3 would play like crap 98% of the time regardless of time of day, so we'll see. Hopefully it's on the up from here on! Thanks for all your help brother.