300M : Bad Broadband Quality Monitor results after Homeworker+ 300Mbit upgrade

[beardsley]

Hi. I upgraded to Homeworker+ 300Mbit and yesterday received a shiny new Hub 3 (surprising black rather than the white I expected). I connected last night and after a few reboots if seemed to be working. I did a speedtest and all looked great. Then I switched it to modem mode and the fun started. Over the next couple of hours, if would keep dropping the connection and I had to give it a series of reboots. Eventually it seemed to settle down and is now maintaining the connection.

However the graph from thinkbroadband Broadband Quality Meter look pretty dire.



You can see the red bar where I swapped from the existing Shub 2ac to the Hub3.

This is what I get from the hub status pages

------------

Cable Modem Status
Item Status Comments
Acquired Downstream Channel(Hz) 299000000 Locked
Ranged Upstream Channel(Hz) 46200000 Locked
Provisioning State Online


Downstream bonded channels
Channel Frequency(Hz) Power (dBmV) SNR(dB) Modulation Channel ID
1 299000000 -0.2 38.9 256 qam 13
2 323000000 -0.5 38.9 256 qam 16
3 315000000 -0.7 38.6 256 qam 15
4 307000000 -0.4 38.9 256 qam 14
5 291000000 -0.5 38.9 256 qam 12
6 283000000 -0.5 38.9 256 qam 11
7 275000000 -0.5 38.9 256 qam 10
8 267000000 -0.2 38.9 256 qam 9
9 259000000 -0.2 38.9 256 qam 8
10 251000000 0.4 38.9 256 qam 7
11 243000000 0.4 38.9 256 qam 6
12 235000000 0.7 38.9 256 qam 5
13 227000000 0.7 38.9 256 qam 4
14 219000000 1 40.3 256 qam 3
15 211000000 1.2 38.9 256 qam 2
16 203000000 1.4 40.3 256 qam 1


Upstream bonded channels
Channel ID Frequency(Hz) Modulation Power (dBmV) Mode Channel Bandwidth(Hz) Symbol Rate (ksps)
13 46200000 ATDMA 47.5 16 qam 6400000 5120
15 32600000 ATDMA 45.8 16 qam 6400000 5120

Operational Configuration
General Configuration Value
Network Access Disabled
Max number of CPEs 1
Baseline Privacy Enabled
Docsis mode Docsis30
Configuration file Vec1d7393ac210ed6.cm

Primary Downstream Service Flow
SFID 241747
Max Traffic Rate 330000000

Primary Upstream Service Flow
SFID 241746
Max Traffic Rate 21000000

-----------

It struck me as strange that 'Network Access' was listed as disabled, but I assume that if this was a problem, it would not be working at all.

Also Upstream power levels are higher than I expected, but I thing are still within spec?

I am hoping to get some collective wisdom from here before I log a call with VM. It is easy enough to explain a fault when something is not working, but trickier when it is sort of working.

Anybody got any ideas?

[General Maximus]

your power levels are fine, including the upstream. Unfortunately you are royally screwed now. There is a known fault with shub3 in modem mode where it will disconnect every ~30 mins and you will lose your internet connection. Vm are aware of it but don't seem to be doing anything about it; maybe it is their secret way of passively forcing everyone to use router mode (which works fine). Unfortunately for you there isn't anything you can do about it and ringing them up and moaning will be pointless because you will just be wasting your time. If it causes you too much agro the only thing you can do is drop back down to 200mbits and ask for a new shub2 (they won't reactivate your old one) which you might have to go through retentions for.

I wish the government would pass a law where they can get fined for this kind of thing because it is so frustrating. They are knowingly selling a faulty product.

[jb66]

They should be able to put the sh2 back on if you downgrade as the MAC should be still on your account as a second deactivated hub

[General Maximus]

sweet, I would do that asap dude.

[beardsley]

It was working fine from 08:00 to 17:00 today while I was working from home. No reboots apart from one I triggered manually. However, It has gone down 4 times in the last 20 minutes. It does not seem to match the disconnect every 30 minutes.

[MUD_Wizard]

Originally Posted by beardsley

Hi. I upgraded to Homeworker+ 300Mbit and yesterday received a shiny new Hub 3 (surprising black rather than the white I expected). I connected last night and after a few reboots if seemed to be working.

Those will be installing the initial firmware updates.

Originally Posted by beardsley

I did a speedtest and all looked great.

Yes the Hub 3 works great in router mode, even with the slightly spikey BQM.

Here's mine for comparison (router mode):

Originally Posted by beardsley

Then I switched it to modem mode and the fun started. Over the next couple of hours, if would keep dropping the connection and I had to give it a series of reboots. Eventually it seemed to settle down and is now maintaining the connection.

You should post your network log, so we can see what's what.

Originally Posted by beardsley

However the graph from thinkbroadband Broadband Quality Meter look pretty dire.



You can see the red bar where I swapped from the existing Shub 2ac to the Hub3.

Your BQM graph looks fairly average for a Hub 3, excluding the red packet loss events (which may or may not be explained by the combination of firmware updates in the first few hours, events in the network log and your reboots).

The slight increase in minimum latency (green) and increased thickness of the blue average latency will be due to the additional 8 channels of downstream you have locked.

Each 256 QAM downstream channel incurs 0.32 ms of additional latency per channel due to interleave. See the table at the bottom of this link:
http://www.kitz.co.uk/adsl/interleav...g_introduction

16 downstream channels locked (as you have) means a total of 5.12ms due to interleave instead of 2.56ms, when you only have 8 channels locked.

Since the BQM only plots every 100 seconds a single spike within that time is enough to produce the yellow spikes you see. I think those are maintenance tasks and data being transferred. The SH2ac also experiences similar yellow spikes in router mode (I have one running on the same line as my Hub 3), though not as often and not as high.

Where hub 3 users only have 8 or 12 channels locked their BQM graph has less yellow spikes and less thickness of blue.

Hope that helps in some way to understand what you're seeing.

The blue spikes on the graph look like times when data is being transferred?

Originally Posted by beardsley

It struck me as strange that 'Network Access' was listed as disabled, but I assume that if this was a problem, it would not be working at all.

Correct. It's a known issue in the GUI which is fixed in firmware 9.1.88T, so you must have been on 9.1.88J5 at that point. Your firmware may have updated again since then?

Originally Posted by beardsley

Also Upstream power levels are higher than I expected, but I thing are still within spec?

Anything in the 40's is perfect. Though keep an eye on them as that could be the reason for your disconnects (assuming it's not you rebooting your hub)?

---------- Post added at 21:24 ---------- Previous post was at 21:06 ----------

Originally Posted by General Maximus

Unfortunately you are royally screwed now. There is a known fault with shub3 in modem mode where it will disconnect every ~30 mins and you will lose your internet connection. Vm are aware of it but don't seem to be doing anything about it;

That's only IF he is on an Arris CMTS (which he can check by looking up the CMTS MAC in his network log).

If he's on a Cisco CMTS then the Hub 3 will work fine, even in modem mode.

Even if this user does experience port flapping while in modem mode (not everyone on an Arris CMTS does and not all the time) then he can apply in this thread to get a push of the latest firmware 9.1.116V, which VM believes fixes the issue:

http://community.virginmedia.com/t5/...187183#M113286

http://community.virginmedia.com/t5/...188961#M113510

In the mean-time, if necessary, he can run his hub in router mode.

Though we haven't established what his issues are at the point.

[Sephiroth]

I'm not sure that the upstream is fine. Muddie was right to say keep your eye on the levels. Here's why:

Upstream bonded channels
Channel ID Frequency(Hz) Modulation Power (dBmV) Mode Channel Bandwidth(Hz) Symbol Rate (ksps)
13 46200000 ATDMA 47.5 16 qam 6400000 5120
15 32600000 ATDMA 45.8 16 qam 6400000 5120

The power level 45.8 suggests to me that there are power adjustments taking place. The Hub 3 is a bit buggy in the GUI, but if you dive in and out of the GUI and bring up the upstream, any variations heading towards the 50s should show.

[vm_tech]

Seph, are you concerned about the difference between the 2 upstream transmit levels?

[jb66]

Originally Posted by vm_tech

Seph, are you concerned about the difference between the 2 upstream transmit levels?

I think he is concerned that it ends with a .8

If the superhub 2 worked fine it's safe to say it's not network related

[vm_tech]

Is there some sort of bug then with them ending in .8? I've never actually used a hub 3

[jb66]

Originally Posted by vm_tech

Is there some sort of bug then with them ending in .8? I've never actually used a hub 3

It usually indicates there are power adjustments on the upstream on the sh2. Not sure about 3s either. Prefer 2ac

[Sephiroth]

Originally Posted by jb66

It usually indicates there are power adjustments on the upstream on the sh2. Not sure about 3s either. Prefer 2ac

As per what JB says about the .8 reading.

[adduxi]

Originally Posted by beardsley

Hi. I upgraded to Homeworker+ 300Mbit and yesterday received a shiny new Hub 3 (surprising black rather than the white I expected).

That's interesting? I thought all Hub 3's were white?

[beardsley]

Hi Mud_Wizard

Thanks for the extensive reply

Originally Posted by MUD_Wizard

Those will be installing the initial firmware updates.

Thanks. That makes sense. In the quick start guide with the router they tell you to reboot after it has been online for 3 minutes. Looks like that figure might be a bit optimistic.

Originally Posted by MUD_Wizard

[COLOR="blue"]Yes the Hub 3 works great in router mode, even with the slightly spikey BQM.

This is what my BQM looked like when I was using the Superhub 2ac. You can see why I thought things had got worse

Originally Posted by MUD_Wizard

You should post your network log, so we can see what's what.

Here it is. I have blanked out the MAC address of my modem

Network Log
Date And Time Error Number Event Description
2016-08-25 21:20:42.00 69010200 SW Download INIT - Via Config file
1970-01-01 00:01:26.00 84020200 Lost MDD Timeout;CM-MAC=xx:xx:xx:xx:xx:xx;CMTS-MAC=00:00:00:00:00:00;CM-QOS=1.1;CM-VER=3.0;
1970-01-01 00:01:26.00 84020300 MDD message timeout;CM-MAC=xx:xx:xx:xx:xx:xx;CMTS-MAC=00:00:00:00:00:00;CM-QOS=1.1;CM-VER=3.0;
2016-08-25 21:30:58.00 69010900 Disruption during SW download - Power Failure
2016-08-25 21:31:27.00 82000200 No Ranging Response received - T3 time-out;CM-MAC=xx:xx:xx:xx:xx:xx;CMTS-MAC=00:01:5c:81:c6:50;CM-QOS=1.1;CM-VER=3.0;
2016-08-25 21:31:39.00 69010900 Disruption during SW download - Power Failure
2016-08-25 21:31:41.00 69010200 SW Download INIT - Via Config file
2016-08-25 21:41:25.00 69011200 SW download Successful - Via Config file
2016-08-25 22:06:58.00 66060300 Map Request Retry Timeout;CM-MAC=xx:xx:xx:xx:xx:xx;CMTS-MAC=00:01:5c:81:c6:50;CM-QOS=1.1;CM-VER=3.0;
2016-08-25 22:10:25.00 82000400 Received Response to Broadcast Maintenance Request, But no Unicast Maintenance opportunities received - T4 time out;CM-MAC=xx:xx:xx:xx:xx:xx;CMTS-MAC=00:01:5c:81:c6:50;CM-QOS=1.1;CM-VER=3.0;
2016-08-26 12:10:18.00 66060300 Map Request Retry Timeout;CM-MAC=xx:xx:xx:xx:xx:xx;CMTS-MAC=00:01:5c:81:c6:50;CM-QOS=1.1;CM-VER=3.0;
2016-08-26 16:24:27.00 82000400 Received Response to Broadcast Maintenance Request, But no Unicast Maintenance opportunities received - T4 time out;CM-MAC=xx:xx:xx:xx:xx:xx;CMTS-MAC=00:01:5c:81:c6:50;CM-QOS=1.1;CM-VER=3.0;
2016-08-26 16:55:41.00 66060300 Map Request Retry Timeout;CM-MAC=xx:xx:xx:xx:xx:xx;CMTS-MAC=00:01:5c:81:c6:50;CM-QOS=1.1;CM-VER=3.0;
2016-08-26 16:59:51.00 82000400 Received Response to Broadcast Maintenance Request, But no Unicast Maintenance opportunities received - T4 time out;CM-MAC=xx:xx:xx:xx:xx:xx;CMTS-MAC=00:01:5c:81:c6:50;CM-QOS=1.1;CM-VER=3.0;
2016-08-26 18:53:42.00 66060300 Map Request Retry Timeout;CM-MAC=xx:xx:xx:xx:xx:xx;CMTS-MAC=00:01:5c:81:c6:50;CM-QOS=1.1;CM-VER=3.0;
2016-08-26 18:58:19.00 82000400 Received Response to Broadcast Maintenance Request, But no Unicast Maintenance opportunities received - T4 time out;CM-MAC=xx:xx:xx:xx:xx:xx;CMTS-MAC=00:01:5c:81:c6:50;CM-QOS=1.1;CM-VER=3.0;
2016-08-26 19:01:56.00 66060300 Map Request Retry Timeout;CM-MAC=xx:xx:xx:xx:xx:xx;CMTS-MAC=00:01:5c:81:c6:50;CM-QOS=1.1;CM-VER=3.0;
2016-08-26 19:06:32.00 82000400 Received Response to Broadcast Maintenance Request, But no Unicast Maintenance opportunities received - T4 time out;CM-MAC=xx:xx:xx:xx:xx:xx;CMTS-MAC=00:01:5c:81:c6:50;CM-QOS=1.1;CM-VER=3.0;
2016-08-26 19:10:24.00 66060300 Map Request Retry Timeout;CM-MAC=xx:xx:xx:xx:xx:xx;CMTS-MAC=00:01:5c:81:c6:50;CM-QOS=1.1;CM-VER=3.0;
2016-08-27 09:58:33.00 82000200 No Ranging Response received - T3 time-out;CM-MAC=xx:xx:xx:xx:xx:xx;CMTS-MAC=00:01:5c:81:c6:50;CM-QOS=1.1;CM-VER=3.0;

Originally Posted by MUD_Wizard

Your BQM graph looks fairly average for a Hub 3, excluding the red packet loss events (which may or may not be explained by the combination of firmware updates in the first few hours, events in the network log and your reboots).

The slight increase in minimum latency (green) and increased thickness of the blue average latency will be due to the additional 8 channels of downstream you have locked.

Each 256 QAM downstream channel incurs 0.32 ms of additional latency per channel due to interleave. See the table at the bottom of this link:
http://www.kitz.co.uk/adsl/interleav...g_introduction

16 downstream channels locked (as you have) means a total of 5.12ms due to interleave instead of 2.56ms, when you only have 8 channels locked.

Since the BQM only plots every 100 seconds a single spike within that time is enough to produce the yellow spikes you see. I think those are maintenance tasks and data being transferred. The SH2ac also experiences similar yellow spikes in router mode (I have one running on the same line as my Hub 3), though not as often and not as high.

Where hub 3 users only have 8 or 12 channels locked their BQM graph has less yellow spikes and less thickness of blue.

Hope that helps in some way to understand what you're seeing.

The blue spikes on the graph look like times when data is being transferred?

Thanks for the explanation. I was not aware of the extra latency due to channel interleve. I look forward to studying the link in more detail when I can devote the time.

Originally Posted by MUD_Wizard

Correct. It's a known issue in the GUI which is fixed in firmware 9.1.88T, so you must have been on 9.1.88J5 at that point. Your firmware may have updated again since then?

Correct I am on 9.1.88T

Originally Posted by MUD_Wizard

Anything in the 40's is perfect. Though keep an eye on them as that could be the reason for your disconnects (assuming it's not you rebooting your hub)?[COLOR="Silver"]

I was asking about power leveles because in the past mine have fluctuated quite a bit. In the past the engineer had left a selection of forward path attenuators for me to swap around when the levels changed too much. This had not been such a problem since another enginner changed the tap in the cabinet about 6/8 months ago. I have not had to use an attenuator since then. He also added an inline filter between the wallplate and my modem.

Originally Posted by MUD_Wizard

That's only IF he is on an Arris CMTS (which he can check by looking up the CMTS MAC in his network log).

If he's on a Cisco CMTS then the Hub 3 will work fine, even in modem mode.

CMTS MAC is 00:01:5c:81:c6:50 which I think decodes to Arris

Originally Posted by MUD_Wizard

Even if this user does experience port flapping while in modem mode (not everyone on an Arris CMTS does and not all the time) then he can apply in this thread to get a push of the latest firmware 9.1.116V, which VM believes fixes the issue:

http://community.virginmedia.com/t5/...187183#M113286

http://community.virginmedia.com/t5/...188961#M113510

Not 100% sure that the problem I have is the port flipping. Connection seems to be stable most of the time - It has been up since about 20:30 last night. However when the does drop, it does it repeatedly at short intervals for a while. This is my snapshot QBM graph for last 24 hours. (If I have got the formatting right)


Do you think it is worth requesting 9.1.116V firmware? Are there any known issues with it that I need to be aware of?

Originally Posted by MUD_Wizard

In the mean-time, if necessary, he can run his hub in router mode.

When I received my first Superhub 2ac I did try running it in router mode. After a short time it went into a flappy state and went into a constant reboot cycle every few minutes. I tried to backup my configuration prior to full reset but it refused to recognise my credentials. Apparently this was a known problem. Speaking to support there was noting they could do, so we did a factory reset and I entered my setings from paper notes. The issue re-occured and a replacement Superhub 2ac was sent. This one immediately had the same problem as soon as I loaded my configuration settings. Putting the hub into modem mode and using an external router cause all the issues to do away.

I was prepared to give the hub 3 a try in router mode but the GUI was so slow to be almost unusable and the real show stopper was that would not accept the IP address I needed in the DMZ setting. If seemed happy to accept other addresses, but not the .5 I needed to set.

The 2ac also would not let me set the subnet mask that I wanted. I have a virtualisation lab set-up at home and can get through a large number of IP addresses. I use a netmask of 255.255.248.0 which the 2ac would not accomodate, so I had to work around that. My TP-Link router is happy to let me set whatever subnet want. I do not know what the hub 3 will accept.

If I do have to sun the hub 3 in router mode, I am thinking of setting my router to an address that the hub 3 will accept as a DMZ and taking the hit on double NAT. I have aleady disabled UPNP, so apart from a small latency increase I think I should be OK.

Also, I understand that the hub 3 is rather picky regarding what port it will forward in port forwarding. Does this also apply with DMZ? I am running a few external facing sirvices which I need to route to the correct internal addresses.

Originally Posted by MUD_Wizard

Though we haven't established what his issues are at the point.

Now that you have explained the yellow spikes on the BQM graph, my main issue is the connection staying up.

Thanks agian for you very detailed reply.

---------- Post added at 15:21 ---------- Previous post was at 15:20 ----------

Originally Posted by adduxi

That's interesting? I thought all Hub 3's were white?

So did I. The power brick and cable are white, but the modem is shiny black.

---------- Post added at 15:22 ---------- Previous post was at 15:21 ----------

Originally Posted by Sephiroth

As per what JB says about the .8 reading.

The .8 has gone now.

Upstream bonded channels
Channel ID Frequency(Hz) Modulation Power (dBmV) Mode Channel Bandwidth(Hz) Symbol Rate (ksps)
13 46200000 ATDMA 47.5 16 qam 6400000 5120
14 39400000 ATDMA 47 16 qam 6400000 5120

[Sephiroth]

Originally Posted by MUD_Wizard

<SNIP>.

Each 256 QAM downstream channel incurs 0.32 ms of additional latency per channel due to interleave. See the table at the bottom of this link:
http://www.kitz.co.uk/adsl/interleav...g_introduction

16 downstream channels locked (as you have) means a total of 5.12ms due to interleave instead of 2.56ms, when you only have 8 channels locked.

Since the BQM only plots every 100 seconds a single spike within that time is enough to produce the yellow spikes you see. I think those are maintenance tasks and data being transferred. The SH2ac also experiences similar yellow spikes in router mode (I have one running on the same line as my Hub 3), though not as often and not as high.

Where hub 3 users only have 8 or 12 channels locked their BQM graph has less yellow spikes and less thickness of blue.

Hope that helps in some way to understand what you're seeing.

Muddie,

With regard to the bit I've highlighted in red, what are you saying? It seems that you have made the latency additive (as in 16 * .32ms = 5.12ms).

How can that be? Bonding is a parallel function using Frequency Division Multiplexing. It is not additive.

So perhaps you should adjust the explanation you have made for the change in "yellow" on the BQM as between SH2ac and Hub 3. Incidentally, my Hub 3 has seen 8, 16 and now 20 DS channels and the "yellow" hasn't changed at all.