Print

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

[burakkucat]

Thank you, I shall await your results

Considering the QLN and SNR graphs (attached below), the data that produced them was tabulated in one file. Close inspection showed that there were four areas to investigate.

Code: [Select]

Tone QLN SNR

200 -112.8750 42.6250
201 -110.8125 40.4375
202 -108.0000 37.7500
203 -103.5000 33.1875
204 -94.3750 24.1250 ####
205 -93.6250 23.3750 ####
206 -103.3125 32.6875
207 -107.9375 37.3125
208 -110.8125 39.8125
209 -112.5625 41.2500

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

248 -119.7500 45.0000
249 -118.2500 43.6250
250 -116.1250 41.7500
251 -111.8125 38.0000
252 -102.8750 28.7500 ####
253 -102.1875 27.8750 ####
254 -111.6875 37.5000
255 -115.4375 41.3125
256 -117.9375 43.3125
257 -120.4375 44.5000

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

276 -121.1875 42.6875
277 -119.5000 41.8125
278 -118.1875 40.0000
279 -115.9375 37.6875
280 -112.1250 33.8750
281 -104.6875 26.1250 ####
282 -95.8125 30.3125 ####
283 -109.3750 30.3125 ####
284 -114.2500 35.5625
285 -117.1875 38.1875
286 -119.3750 39.6875
287 -120.8750 41.1250
288 -122.3125 41.8125

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

3186 -123.5000 30.3750
3187 -121.2500 29.8125
3188 -119.5000 29.9375
3189 -115.6250 29.2500
3190 -108.1875 29.2500 ####
3191 -100.2500 29.2500 ####
3192 -113.2500 29.2500
3193 -117.6875 29.2500
3194 -121.0000 30.1875
3195 -122.3125 30.4375

The frequencies (+/- 2 kHz) corresponding to those areas of interest are --

Code: [Select]

882 kHz

1089 kHz

1216 kHz

13759 kHz

Unfortunately there was only correlation with the third frequency, 1216 kHz (+/- 2 kHz) --

Code: [Select]

1215 Virgin AM Norfolk         Postwick 1200 TG303086
1215 Virgin AM Northern Ireland Lisnagarvey 16000 IJ258619
1215 Virgin AM Humberside         Hull         320 TA169258
1215 Virgin AM Tyne and Wear         Wrekenton 2200 NZ274598
1215 Virgin AM Worcestershire         Droitwich 105000 SO929663
1215 Virgin AM Yorkshire         Moorside Edge 200000 SE070154
1215 Virgin AM Somerset         Washford 100000 ST058410
1215 Virgin AM Cornwall         Redruth 2000 SW709403
1215 Virgin AM Dartford Tunnel Dartford Tunnel 4 TQ571769
1215 Virgin AM London                 Brookmans Park 125000 TL259050
1215 Virgin AM Lothian         Westerglen 100000 NS868773
1215 Virgin AM Grampian         Redmoss 2300 NJ942024
1215 Virgin AM Hampshire         Fareham 1000 SU546058
1215 Virgin AM Devon                 Plymouth 1100 SX490585

So I shall assume that Jamie is located in an area served by one of the above transmitters. They are shown on the map (attached below) by the blue markers.

The origin of the other three major spikes remain a mystery. 

[NewtronStar]

Hi Everyone,

Relatively new to VDSL2 stats, got hold of a HG612 and unlocked it today, have managed to extract some stats, but i'm interested if someone could tell me what my Interleaving level is like?

I'm also guessing from my attainable rate, I've a pretty good chance with faster speeds on an 80/20 package?

heres my pennys worth the interleaving depth seems very high but that should mean less errored seconds but with a high latency, and you should do well on the 80/20 profile when you switch over to it.

[Chrysalis]

Colin the spare snr is extremely important.

If someone has a X amount of visible noise at a higher snr margin, then that automatically means that noise is of a larger amount as it is overcoming the higher snr margin.

If there is enough noise on a line with 17db of snrm that DLM has seen fit to apply interleaving, then when the line is down to 6db of snrm, there is going to be a ton of visible noise, no doubt about it.  The interleaving depth will be increased if he upgrades and I think he wont get anywhere close to 80mbit sync speed.

You are comparing your FEC rate to his but you have different snr margins.

I can give you some examples.

I observed error rates between 6db and 8db snrm doubling, so an extra 2db of snrm halved the error rate. On my upstream a extra 5db snrm reduced the error rate to 10% just 1/10th. 

So in a 33 hour period he had 31985 FEC with a 17db snrm.

My guestimate if he had a 6db snrm his error rate for 33 hours would be approx 900k to 1.5million.  He would probably be getting CRC errors as well as that error rate would be too high to fix all CRC.  That in turn would likely increase the interleaving depth from his current setting to manage the new line conditions.

[ColinS]

The interleaving depth will be increased if he upgrades and I think he wont get anywhere close to 80mbit sync speed.

Nor, as I have already said several times, do I think he will achieve the full 80/20.  But I can't predict exactly what he will get, any more than you have with the 'precise' estimate you have made

... not anywhere close

So, you imagine what, 41/2.1? No competent engineer, seeing that kind of performance on an 80/20 product, would let it go uninvestigated.

Something is causing these recovered errors.  But, frankly I think no one would investigate it while his line's capability so far exceeds it's current product banding, that he is achieving 100% of everything he's currently paying for.

[Edit] PS: I guess you would say I was just lucky then to survive bursts of 3.5Million bit errors/min over a 2-3 minute period on ~half his current interleaving depth?

[Chrysalis]

That I do agree with, I agree he has a problem.

I also agree on his current product it wont get investigated as he has full sync speed with no CRC errors.

Just remember tho the reported attainable rate only gets hit on fast path as FEC overheads make the actual sync lower on interleaving. So his line in its current state wont hit the attainable rate as its interleaved, and if he goes to 80/20 his snrm goes down, the error rate goes up, then there is a chance the intelreaving depth increases and this further brings down the sync speed again.  I think he would be lucky to get 65mbit.

[jid]

Thank you for all the replies

As for that map of transmitters B*cat, the only possibility is the one in Somerset, I'm only across the water in South Wales from that transmitter.

It seems from all of your feedback, that it will only appear apparent there is a problem when I'm moved to a faster package? As you've said, there's a distinct noise problem somewhere which it seems my ISP will only consider investigating when there's more serious errors occurring on the line?

Once again, thank you to all who've replied, its been a learning curve and something I'll keep reading up on as part of my degree

[kitz]

882 kHz

Code: [Select]

200 -112.8750 42.6250
201 -110.8125 40.4375
202 -108.0000 37.7500
203 -103.5000 33.1875
204 -94.3750 24.1250 ####
205 -93.6250 23.3750 ####
206 -103.3125 32.6875
207 -107.9375 37.3125
208 -110.8125 39.8125
209 -112.5625 41.2500
iirc jamie had problems with those same tones on adsl1

[Chrysalis]

in my adsl days I had that same dip (along with 2 other dips).  I always assumed was local radio transmission.

[jid]

iirc jamie had problems with those same tones on adsl1

That would be right Kitz, it seems the problem still haunts me

ALthough I did find this, which may answer the questions about the option TalkTalk have to apply a "DLM Profile" for TV customers : http://community.talktalk.co.uk/t5/Superpowered-Fibre-Broadband/Not-So-Superpowered-Fibre-Broadband/m-p/951118#U951118

[kitz]

37mb down to 16mb is a lot.  Streaming services shouldn't take up that much bandwidth!

[jid]

37mb down to 16mb is a lot.  Streaming services shouldn't take up that much bandwidth!

Seems his line is being throttled by DLM.

IPTV varies on usage, I've seen anything from 2 - 5mbps being used.

[Chrysalis]

that talk talk post is interesting, I wonder if talk talk move their customers with their tv service onto the stable DLM profile?

[kitz]

Its a damn shame that I bet that person doesnt use an unlocked HG612.   Would be most interesting to have seen his before and after stats. :/

[jid]

Its a damn shame that I bet that person doesnt use an unlocked HG612.   Would be most interesting to have seen his before and after stats. :/

I've posted asking about it, but will have to see as it could play hell with my IPTV quality if errors increase.

[jid]

TalkTalk confirmed that there is a different DLM profile for TV customers.

After leaving mine stable for a while my interleave depth has decreased significantly - seems DLM resynced the line to do this:

Code: [Select]

xdslcmd info --stats

xdslcmd: ADSL driver and PHY status

Status: Showtime

Retrain Reason: 2

Max: Upstream rate = 32818 Kbps, Downstream rate = 95932 Kbps

Path: 0, Upstream rate = 1999 Kbps, Downstream rate = 39996 Kbps



Link Power State: L0

Mode: VDSL2 Annex B

VDSL2 Profile: Profile 17a

TPS-TC: PTM Mode

Trellis: U:ON /D:ON

Line Status: No Defect

Training Status: Showtime

Down Up

SNR (dB): 20.2 31.0

Attn(dB): 0.0 0.0

Pwr(dBm): 12.7 5.1

VDSL2 framing

Path 0

B: 57 31

M: 1 2

T: 64 36

R: 16 16

S: 0.0461 0.9922

L: 12835 645

D: 701 1

I: 74 80

N: 74 80

Counters

Path 0

OHF: 57974584 852339

OHFErr: 0 0

RS: 2541185293 4027317

RSCorr: 7936 0

RSUnCorr: 0 0



Path 0

HEC: 0 0

OCD: 0 0

LCD: 0 0

Total Cells: 1319058268 0

Data Cells: 86869855 0

Drop Cells: 0

Bit Errors: 0 0



ES: 7 0

SES: 5 0

UAS: 34 34

AS: 128858



Path 0

INP: 3.00 0.00

PER: 2.21 13.39

delay: 8.00 0.00

OR: 86.72 57.33



Bitswap: 358 2183



Total time = 1 days 34 min 1 sec

FEC: 154519 0

CRC: 2605 0

ES: 7 0

SES: 5 0

UAS: 34 34

LOS: 5 0

LOF: 5 0

Latest 15 minutes time = 4 min 1 sec

FEC: 4 0

CRC: 0 0

ES: 0 0

SES: 0 0

UAS: 0 0

LOS: 0 0

LOF: 0 0

Previous 15 minutes time = 15 min 0 sec

FEC: 6 0

CRC: 0 0

ES: 0 0

SES: 0 0

UAS: 0 0

LOS: 0 0

LOF: 0 0

Latest 1 day time = 34 min 1 sec

FEC: 43 0

CRC: 0 0

ES: 0 0

SES: 0 0

UAS: 0 0

LOS: 0 0

LOF: 0 0

Previous 1 day time = 24 hours 0 sec

FEC: 5260 0

CRC: 0 0

ES: 0 0

SES: 0 0

UAS: 0 0

LOS: 0 0

LOF: 0 0

Since Link time = 1 days 11 hours 47 min 37 sec

FEC: 7936 0

CRC: 0 0

ES: 0 0

SES: 0 0

UAS: 0 0

LOS: 0 0

LOF: 0 0
Interestingly I see there's another thread on the retrain reason codes, which I've found this link:

Code: [Select]

define kRetrainReasonNegativeMargin                2
So it seems the line hasn't coped at some point and dropped?