Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[William Grimsley]

Hello,

Really don't want to go insane like before, but I was going to ask, are SES on G.INP a bad sign, I was thinking if bursts of CRC's are happening with G.INP enabled, then "We've got a problem, Houston!"

[broadstairs]

I suspect any errors CAN happen even on G.INP, it all depends on how many and how often. I would think zero CRCs all the time is probably unlikely on any line except one which is only a few feet away from the cabinet.

Stuart

[William Grimsley]

Yes, I understand. The main thing I'm slightly concerned about is I had around 30 - 40 Downstream CRC's in one go causing a Downstream SES on G.INP, that would probably been hundreds on interleaving...

[burakkucat]

I would advise that you make a note of the event and all the relevant details in your circuit's "log book". Then just carry on using the service as normal.

The next time you experience another event, log it and review all past events. If you see a pattern developing, then we can discuss it and see, with the collective knowledge, if we are able to deduce the cause.

[WWWombat]

I think the best answer is that SES are definitely more serious than ES, in that they depict a pretty horrible concentration of errors.

That seriousness stays, whatever error protection methods have been applied to the line - G.INP, interleaving, FEC, banding, high SNRM targets, or nothing at all.

As @b*cat says, a one-off isn't a worry. It is only when it becomes part of a pattern that you need to worry - and you need to take into account all the other things going on too.

[Chrysalis]

I agree with the above posts, a SES isnt good, but its the frequency thats important, if you just getting e.g. one a day, then not a big deal.

On my line with 25 days uptime I have had 11 SES, line performs fine so is no concern.

[NewtronStar]

Only seem to get SES errors when the modem has done a quick resync by the DLM or the stats software gets confused during a resync.

Here is a graph showing SES errors all that happened at that time was a resync.

[William Grimsley]

Ok, thanks for your help guys. I will keep monitoring.

[NewtronStar]

With G.INP I tend to keep an eye on RTX_UC (uncorrected errors) then tie it up with RTX_TX (RE-Transmission) spikes and also keep a close eye on the LEFTRS.

G.INP is odd to say the least but if its working on your line then most of your error corrections are  being fixed on the fly 

[William Grimsley]

I see. Well, I've had another Downstream SES in the night due to a burst of Downstream CRC's. I will keep monitoring over the next few days to see if I can find a pattern occuring...

[William Grimsley]

There have been 301 Downstream CRC's on the line in the past 19 hours, slightly concerned about that to say G.INP is activated. The Downstream CRC's seem to have stopped for the moment but will check again at some point in the not too distant future.

xDSL
Mode   VDSL2
Traffic Type   PTM
Status   Up
Link Power State   L0
Downstream   Upstream
Line Coding (Trellis)   On   On
SNR Margin (dB)   6.3   5.7
Attenuation (dB)   25.9   0.0
Output Power (dBm)   12.0   6.1
Attainable Rate (Kbps)   43091   7912
Rate (Kbps)   42756   7912
B (# of bytes in Mux Data Frame)   243   239
M (# of Mux Data Frames in an RS codeword)   1   1
T (# of Mux Data Frames in an OH sub-frame)   0   42
R (# of redundancy bytes in the RS codeword)   10   0
S (# of data symbols over which the RS code word spans)   0.1817   0.9639
L (# of bits transmitted in each data symbol)   11182   1992
D (interleaver depth)   8   1
I (interleaver block size in bytes)   254   120
N (RS codeword size)   254   240
Delay (msec)   0   0
INP (DMT symbol)   46.00   0.00
OH Frames   0   0
OH Frame Errors   443   6101
RS Words   1761537656   4179611
RS Correctable Errors   31471439   0
RS Uncorrectable Errors   0   0
HEC Errors   0   0
OCD Errors   0   0
LCD Errors   0   0
Total Cells   3382804391   0
Data Cells   2438724482   0
Bit Errors   0   0
Total ES   52   5342
Total SES   10   4
Total UAS   31   31

[WWWombat]

There have been 301 Downstream CRC's on the line in the past 19 hours, slightly concerned about that to say G.INP is activated.

G.INP can't fix everything. If a retransmission is also lost, it can't keep on retransmitting forever. After a short period, it gives up ... and the missing block of data becomes a CRC error.

When you see CRC counts, you will also be seeing rtx_uc counts too.

[William Grimsley]

Ah, that's interesting, thanks for that, WWWombat!