Superhub : Just got the SuperHub 3

[Kushan]

Originally Posted by Sephiroth

No - I got it this time. Before my very eyes.

I also got the update at about the same time 8pm! Well done, Virgin. Not like I wasn't in the middle of something at peak time or anything.

Did I read somewhere that firmware updates weren't actually meant to cause a reboot? Or am I making that up?

[ianch99]

Originally Posted by Kymmy

http://www.cableforum.co.uk/board/12...-channels.html

You want post #3 onwards, though 12 channel and less will do 200Mbps OK... Not sure though that the new hub with it's NAT/Firewall issues will do what you want any better than the other superhubs

Many thanks. I would be using modem mode so the router part of Hub 3.0 would be not used .. It was the remark about lower latency on the Hub 3.0 that intrigued me.

Silly question: how do I find my CMTS MAC address? It is not in my SH2's current Network Log entries

[Sephiroth]

Originally Posted by Travelstar

Recently have moved across to VM from BT and purchased the Vivid 200 service.

Whilst the overall service seems pretty decent and in fact I have even received over 300 Mbit/sec on some days. Unfortunately last night my speed (wired) seemed to drop to around 13-15 Mbits/sec which is a significant and unexpected decrease). In addition to this my experience with the SuperHub 3 has been pretty poor.

Port 4 which happens to also be the only port which works when the unit is in modem mode seems flaky. Sometimes it decides it just wants to be a 100 Mbit/sec port rather than gigabit. I've tried other cables just to check and i've been able to replicate the issue but it can run a day without dropping down to a slower speed.

In addition to the above, the Virgin Hub UI must be the slowest I have ever used. Pages take literally 30 seconds+ to load which seems more than a little extreme. I really can't see how this got out of UAT with that performance. Wifi range also seems fairly poor and my Home Hub was significantly better so I am taking the plunge and getting an R8000 to take over those duties. That said, I need to get the SuperHub fixed so that port 4 does not decide to drop to the lower speed.

Questions:

1. Have other people found the UI of the SuperHub 3 to be as poor as me where pages literally take 30 seconds (sometimes more) to load? I've read on this forum that the hub is slow, but I was not expecting it to be as bad as this.
[SEPH]: Which firmware are you on? It's shown in the Info subsection of the ADMIN section. Until yesterday I was on the 85J firmware and it wasn't as slow as you describe.

2. Have others noticed that the propagation performance of the SH3 wifi is poor? Two walls seems enough to pretty much kill the signal.
[SEPH]: Two walls can kill any WiFi signal. I've found it to be OK - as good as the SH2ac (which is acknowledged as fine). That said, on principle I am in modem mode with an external router that has external antennae and many more router functions that the hubs have got.

3. Anyone else noticed port 4 (i.e. the master port) so be flaky? Or has I just ended up with a dodgy router?
[SEPH]: Dodgy router perhaps. Go onto the VM forum and tell your story about Port 4. Within two or three days the VM Forum Team will arrange a technician to come and sort that out.
4. Does the performance of VM really vary as much as 13-320 Mbits/sec? The dramatic slow down last night (admittedly during 'peak time') was unexpected - personally I would have expected I would have obtained at least 50 Mbit/sec.
[SEPH]: Depends on what him next door's six children are doing with their streaming and mega-downloading, etc. It could also have been some temporary blip and you may care to post your downstream and upstream router stats for an opinion.

As a side note I find the overall 'advanced options' within the SH3 pretty poor. It might be good enough for basic users, but for anyone who has even slightly more requirements, you really are hamstrung by what the SH3 offers.
[SEPH]: Yep - as with all the broadband suppliers' gateway devices. Hence modem mode.

[Travelstar]

Thanks for coming back to me Seph!

1. I've just remoted into my modem (thanks to a VPN I have setup) and things this afternoon seem a little faster so I have a feeling I have just been upgraded in respect to firmware.

New firmware is 9.1.88J5

2. That is what I am hoping the R8000 will fix. I can't see the use case yet for the 8500 until devices get updated so I feel the R8000 should at least help me solve the problem.

3. VM have already agreed to send out a tech thankfully.

4. My levels are shown below:

Downstream:

1 323000000 -6.7 35.7 256qam 116
2 315000000 -6.5 35.7 256qam 115
3 307000000 -6.4 36.3 256qam 114
4 299000000 -7.0 36.3 256qam 113
5 291000000 -7.0 36.3 256qam 112
6 283000000 -7.0 36.3 256qam 111
7 275000000 -7.0 27.3 256qam 110
8 267000000 -6.7 37.3 256qam 109

Upstream:

5 46200000 ATDMA 51 32qam 6400000 5120
7 32600000 ATDMA 51 16qam 6400000 5120

Cheers!

[horseman]

Originally Posted by ianch99

This is interesting. Is the HUB 3.0 now a standard option on Vivid 200? If it is and you are on the upgraded CMTS kit, can you ring in and buy/haggle for one? I ask because my son is an avid PS4 gamer and if this helps with gaming then it may be worth asking for one.

....

Currently Hub 3.0 still on pilot release to new V200 customers only although this can vary by area/inventory levels and also unpredictably variable as VM's processes can be quite dynamically changeable - (diplomatic phrase for totally inconsistent).
If you're naively placing your optimism in a half developed Hub with 16 (D/S) channels and just ICMP results from TBB hosted server (without checking UDP/TCP latency to your specific game servers) then be prepared for potentially disappointment/disillusionment or conversely staggering delightenment!
(I think that's wat I "ment"? )

[Synthetic]

Originally Posted by Travelstar

Thanks for coming back to me Seph!

1. I've just remoted into my modem (thanks to a VPN I have setup) and things this afternoon seem a little faster so I have a feeling I have just been upgraded in respect to firmware.

New firmware is 9.1.88J5


--snip--

Isn't that the current, non-trialist firmware? I'm on that firmware and have been since I got my SH3.

Thought there might have been a non trialist SH3 update by now

[Sephiroth]

Originally Posted by Travelstar

Thanks for coming back to me Seph!

1. I've just remoted into my modem (thanks to a VPN I have setup) and things this afternoon seem a little faster so I have a feeling I have just been upgraded in respect to firmware.

New firmware is 9.1.88J5
[SEPH]: That is the non-triallist firmware. A few things don't work but as ever, modem mode comes to the rescue.

2. That is what I am hoping the R8000 will fix. I can't see the use case yet for the 8500 until devices get updated so I feel the R8000 should at least help me solve the problem.

3. VM have already agreed to send out a tech thankfully.

4. My levels are shown below:

Downstream:

1 323000000 -6.7 35.7 256qam 116
2 315000000 -6.5 35.7 256qam 115
3 307000000 -6.4 36.3 256qam 114
4 299000000 -7.0 36.3 256qam 113
5 291000000 -7.0 36.3 256qam 112
6 283000000 -7.0 36.3 256qam 111
7 275000000 -7.0 27.3 256qam 110
8 267000000 -6.7 37.3 256qam 109
[SEPH]: Noise on this channel is unacceptable high (SNR=27.3). This is because your power levels are outside of the acceptable threshold (-6dBmv to +10dBmv). The tech will fix that.

Upstream:

5 46200000 ATDMA 51 32qam 6400000 5120
7 32600000 ATDMA 51 16qam 6400000 5120
[SEPH]: A noisy upstream too. All this wrecks speed. If it were working perfectly you'd be o 64QAM modulation. I suspect that the same impairment that is dropping your downstream power is causing your modem to push power harder and on inspection you might see it maxing out at 54dBmv.

Cheers!

[Travelstar]

On downstream I have a typo. Should be:

7 275000000 -7.0 37.3 256qam 110


Upstream however I have typed correctly. Sounds like when the tech comes I should point out the noisy upstream.

[ianch99]

Originally Posted by horseman

Currently Hub 3.0 still on pilot release to new V200 customers only although this can vary by area/inventory levels and also unpredictably variable as VM's processes can be quite dynamically changeable - (diplomatic phrase for totally inconsistent).
If you're naively placing your optimism in a half developed Hub with 16 (D/S) channels and just ICMP results from TBB hosted server (without checking UDP/TCP latency to your specific game servers) then be prepared for potentially disappointment/disillusionment or conversely staggering delightenment!
(I think that's wat I "ment"? )

When the Hub 3.0 is "mainstream", it would be interesting to see if there are any descernable gains in switching to it (in modem mode) esp. for gaming ..

[Sephiroth]

Always put all hubs into modem mode.

[MUD_Wizard]

Originally Posted by Horseman

Question for you: If you're on 16channel D/S have you noticed about 15-20% decrease in ICMP max latency?

Originally Posted by ianch99

This is interesting.

It was wishful thinking, nothing more.

SH2ac, 8 downstreams:



SH3, 16 downstreams:

[horseman]

Originally Posted by MUD_Wizard

It was wishful thinking, nothing more.

......

It was indeed, but Mud ol chap, if you're going to waste your time quoting my useless wildly inaccurate posts then at least quote them in context? pretty please...

Originally Posted by horseman

.....

Question for you: If you're on 16channel D/S have you noticed about 15-20% decrease in ICMP max latency?
Nope - forget that insufficient samples - looks the same for me now...

[Kushan]

As far as I can tell, TBB graph performance on the SH3 is either identical or worse depending on what you're looking at (jitter, lowest, highest, etc.).

However, real life performance seems just dandy. No issues gaming or streaming. I can only presume there's some kind of rudimentary QoS or something going on that has ICMP at lower priority? Or the spikes are so marginal that it makes almost no difference.

[Sephiroth]

... or they'e got diagnostic code running.

[horseman]

Originally Posted by Sephiroth

... or they'e got diagnostic code running.

That would conveniently explain both the max latency on ICMP and the tardy GUI response?