Common problems and bebox stats interpretation
Contents |
Where to find your statistics
Your connection statistics can be found on this page of your BeBox's web interface (the default username is Administrator with no password). If you are using Internet Explorer 7 and cannot log into the BeBox, see IE7 and the BeBox on the Wiki for more information and the workaround fix.
If you want to view an automated analysis of your BeBox's connection statistics, and also see some graphs comparing some of your stats to those of other Be* users, use one of the online analysis tools mentioned in the How to analyse your connection statistics wiki article. NB: you should still get an expert to analyse your statistics if you are unsure of anything, as the way your connection is configured (Annex M or fastpath enabled, etc) may cause the BeBox to show false or inaccurate statistics.
Explanation of terms used in statistics
Attenuation
Your downstream attenuation should be approximately double your upstream. If it isn't then you probably have a broken microfilter or a serious line fault.
Modulation
- G.992.5 Annex A is ADSL2+
- G.992.5 Annex M US 56 is ADSL2+ with UploadPLUS
- G.992.3 is ADSL2
- G.992.1 is ADSL
Ideally you will be on ADSL2+ Annex A (if you are on a Pro product you can request Annex M via ticket). It seems there is a bug in the 585v7 that causes it to sometimes revert to lower modulations after a resync, typically rebooting the BeBox should resolve this.
Note: ADSL2 can result in better sync speeds for attenuations between 45 and 60*.
Note2: ADSL can result in better sync speeds for attenuations over about 55*.
(*): These are very approximate attenuation values. ADSL and ADSL2 could also be more stable on these longer phonelines.
Vendor ID
Some firmwares report the remote endpoint is unknown (μ, mew) and some report Broadcom (BDCM). This makes no difference whatsoever to your connection.
Sync Rate
You can find your expected sync rate compared to your line length in this thread here.
Extension wiring is the #1 cause of issues on these boards. If you have a new style 'NTE5' faceplate, then removing the bottom half will allow you access to the test socket. Connecting your modem to this socket disconnects your internal wiring. If you have significantly different stats in the test socket and where you usually connect the router you have internal wiring issues you need to resolve. There is a good guide by Reggie here.
Noise margin
Your noise margin directly affects your sync speed, the lower the margin the higher your sync speed will be. If your margin is too low it will lead to connectivity issues. By default this is set at 6db, and if you have no other faults this SHOULD give the best balance of stability and sync speed. If your margin is very high, it means the line management process has increased it to compensate for problems on your line. You can request noise margins of 3-15 (in 3db steps) from be* by raising a ticket.
Output power
If your downstream output power is unusually low (it should be ~16-19) then you may be using a buggy firmware version (notably 6.2 variants).
Loss of Framing
Framing errors indicate interference, either internally or externally. Sky boxes are often known to cause issues and double filtering them may help. Regarding external interference, you may wish to perform a quiet line test using an analogue telephone (dial 17070 option 2). If there is any noise you can report a voice fault to BT. You can also use the DMT tool to see bitloading per tone if you suspect interference, this will show you a graph which should start high and result in a smooth falloff (high frequencies attenuate more quickly). Any noticeable gaps or troughs indicate the frequency the noise is on (and this may assist you in locating the source).
Loss of Link
Link loss errors are either the result of total signal loss, or the framing loss threshold being met, these errors indicate that a resync was required. The same troubleshooting applies as above, as the framing and signal losses usually go hand in hand.
Error Seconds
These will be reasonably frequent if you are on fastpath (I believe I get about 1 per 2 minutes or so). If you have a lot of errored seconds on an interleaved profile then this may also indicate interference as the redundancy in interleaving should take care of these.
FEC, HEC, CRC
Upstream error counts for these can never be cleared by you, as they are stored at the DSLAM. The BeBox miscounts FEC downstream errors so you can usually disregard these. Your FEC downstream error count will be zero if you are on a fast profile.
Other things to consider when troubleshooting
LAN connection - wired or wireless?
Wireless is somewhat inferior for sustained, high-speed throughput, and is not at all reliable for testing your bandwidth (especially if you sync at 10Mbps or higher).
You can help your wireless throughput if it is extremely slow by using the TCP Optimiser, or by selecting a wireless channel that is not used by neighbouring networks (you can use Network Stumbler to search for these), ideally with a gap of 3 channels between yours and the neighbouring ones.
If you are having throughput issues, please test with a well seeded torrent or by downloading a reasonably large file from a mirror close to you (Linux distributions are good for this) using an Ethernet connection.
Fast connection but slow ping responses?
The second and third hops from your traceroute are unlikely to respond, this is not an issue, icmp packets have low priority on these routers. There appears to be a bug with the 585v7 which causes the BeBox to return with 50 or 100ms responses, this doesn't seem to affect the actual trace or latency calculations for subsequent hops.
