Instability with old ISP (BT) and new one (BEThere)

Hi All,

I was hoping someone in the know could cast an eye over the following router stats to see if there is an obvious problem.

We have a BT engineer coming next week to check noise on the line (after a quiet line check). We are about 4km away from the exchange.

DSL Type: G.992.1 annex A

Bandwidth (Up/Down) [kbps/kbps]: 1,072 / 4,534

Data Transferred (Sent/Received) [kB/kB]: 0.00 / 0.00

Output Power (Up/Down) [dBm]: 12.0 / 18.5

Line Attenuation (Up/Down) [dB]: 28.5 / 48.0

SN Margin (Up/Down) [dB]: 7.0 / 6.0

Vendor ID (Local/Remote): TMMB / BDCM

Loss of Framing (Local/Remote): 13 / 0

Loss of Signal (Local/Remote): 2 / 0

Loss of Power (Local/Remote): 0 / 0

Loss of Link (Remote): 0

Error Seconds (Local/Remote): 11 / 0

FEC Errors (Up/Down): 77 / 140,161

CRC Errors (Up/Down): 77 / 4

HEC Errors (Up/Down): 48 / 3

Many Thanks,

Raas

Stats don't look too bad.
Impossible to say if the errors are high or low as you haven't given an "up" time.

Sync is a little low for a 6dB margin @ 48dB.

Have you tried the usual steps (removed the face-plate at your master socket and connected the router to the 'test' socket inside - if there is one)? This will eliminate your internal wiring as the cause of the slightly low sync or the instability.

Also; does your phone line come in via overhead wires (telephone poles)? This can cause drop-outs due to noise, especially during the long dark winter nights.

Thanks for replying Adebov,

It's been up for just over an hour when I took those readings.

Yes we've got it connected via the test socket now and it's been connected the longest it has all day. I installed an iplate last week and would have thought this would do a similar thing.

Not sure about the overhead wires, but as we're in a rural location I would assume so. I can see one outside with a wire to the house.

If that is the case, do you think a BT engineer would be able to improve the noise problem?

Also, if I instructed my ISP to implement Interleaving, would it help?

Thanks,

Raas

Just thought Id post this after being connected for 5.5 hrs. Getting lots of FEC errors, but seems stable using the test socket.

Uptime: 0 days, 5:33:12

DSL Type: G.992.1 annex A

Bandwidth (Up/Down) [kbps/kbps]: 1,072 / 4,534

Data Transferred (Sent/Received) [kB/kB]: 0.00 / 0.00

Output Power (Up/Down) [dBm]: 12.0 / 18.5

Line Attenuation (Up/Down) [dB]: 28.5 / 48.0

SN Margin (Up/Down) [dB]: 7.0 / 5.5

Vendor ID (Local/Remote): TMMB / BDCM

Loss of Framing (Local/Remote): 13 / 0

Loss of Signal (Local/Remote): 2 / 0

Loss of Power (Local/Remote): 0 / 0

Loss of Link (Remote): 0

Error Seconds (Local/Remote): 11 / 0

FEC Errors (Up/Down): 183 / 6,284,166

CRC Errors (Up/Down): 183 / 152

HEC Errors (Up/Down): 93 / 109

Your line is interleaved (FEC Forward Error Correction is only used on interleaved lines).

The iPlate filters the bell/ring wire. It may be better to actually detach the bell/ring wire, if present, and dump the iPlate.

You still need filters when using the iPlate, if there is a difference between test socket and iPlate stats perhaps you have a faulty filter.

As John says; pay little attention to FEC errors. They're ones which have been corrected by the router without having to request the data again.
CRC errors are the bad ones. These are ones which didn't get through and 152 in 5.5 hours is pretty good for a long(ish) line.

If it's stable in the test socket and no so much when you've got the extensions wired in, you need to look at the possibility of fitting a filtered ADSL face-plate (which installs in place of your current regular face-plate) and locating your router next to the master socket (using CAT5 ethernet to distribute the local LAN to your PC, rather than cheap telephone extension cable to extend the ADSL signal).

If you're very rural it's always possible the lines from the exchange to your village are overhead all the way. You quite often see that round here (mile upon mile of very large overhead wires) and this will make a pretty good AM radio antenna (which would result in a larger than normal drop in overnight noise margins).

I think you need to leave it in the test socket (along with a regular ADSL micro-filter; so you can have at least one telephone connected) for maybe 2-3 times longer than the typical drop-out interval (if it used to drop out every 3-4 hours; leave it connected into the test socket for 9-12 hours - better still; a couple of days), but it does sound like it's more down to internal phone wiring issues (ring wire connected, cheap extension cable, etc.) and if your house is anything like mine the "electricians" employed by the builders used untwisted burglar alarm cable for the phone extensions and lost me around 2Mbps in the process.

Why are you on ADSL1 not ADSL2+?

In reply to a post by BatBoy:

Why are you on ADSL1 not ADSL2+?

Don't you find it odd that G.992.1 implies he's on ADSL1 yet the sync (4534kbps) implies he's on ADSL2/2+ (mainly because for ADSL1 the sync divided by 32 must be an integer and 4534/32 = 141.6875).
Something weird is going on.

The up sync being more than 832 also suggests ADSL2+. Although I don't know if you can have higher than 832 syncs on ADSL1, just don't remember seeing any.

In reply to a post by john2007:

Although I don't know if you can have higher than 832 syncs on ADSL1, just don't remember seeing any.

In theory it should go to about the same speed as annex A of ADSL2/2+ but it seems to be capped by the wholesale providers at around 800kbps (some lower; some higher).