VDSL Performance & Tuning

Forums › New Zealand Broadband › VDSL Performance & Tuning

1 | ... | 204 | 205 | 206 | 207 | 208 | 209 | 210 | 211 | 212 | 213 | 214 | ... | 285

'That VDSL Cat'
13036 posts

Uber Geek
+1 received by user: 3896

ID Verified

Trusted

Lizard Networks

Subscriber

#985127 12-Feb-2014 14:40

BMarquis:
hio77: impressive.. your on one of the odd dslams that has a Max of 71680 rather than 70000, we have seen odd things happen with them..

Hmm, all of our profiles have a max of 71680, not 70000. They are pushed to dslams from a single source based on templates and should all be identical.
Perhaps the fritz misreports in some cases? Have seen many, many cases int he past of modems mis-reporting figures.
the 54mbps upstream value here shows that too.

always assumed it to be a difference between the configs..

we have seen the two different values, on the same firmwares, from different linecards. maybe the fritz is missreading the packets from the linecard? not sure.. not that it really matters anyway! clearly i was wrong here!

BMarquis:
hio77: 11db attenuation, purely from DLM is a first. even if your fritz is miss-reporting (seen the newer firmwares show a 1~3db higher attenuation.

as you can see in your spectrum, distance is clearly a factor for you (note the DL 3 block, tapering off.)

i base my full figures from a chorus employee, along with what we personally have seen in the VDSL section of snap..

by all respects, im hella surprised your on 17a, but congratulations!

My post should be read carefully:
the DLM Choice of 7dB is the ELECTRICAL LENGTH determined by the 5530, not the DSL attenuation. They are different figures (calculated on different frequencies I believe)

e.g, there is a line with:
Electrical Length 6 dB
Attenuation [dB] 10 dB

fully agree, your figures are possibly not the same as the fritz, however i draw conclusions from the likes of sams line, in that the figures generally line up.

as mentioned in that post, the fritz firmware he is on is absolutely known for misreporting of buffering up attenuation values.

this line you speak of however, very interesting example, another line that we would turn around and go WTF about.

BMarquis:

hio77:as i pointed out in the thread, it looks like hes on the funny linecards, the broadcom ones, which seem to have completely different profiles to the usal Ikanos..

thats the only way i can possibly explain it..

Nah, the profiles are the same across those line cards (i.e the profiles are stored on the DSLAM and used by both cards), there is only 1 set of profiles.
Being different chipsets, they have different performance, but the profiles are identical.

on the topic of profiles, have there been more added other than in our list? noticed a few people around with odd profiles (such as 8b 8ms/1ms) or is this simply CPE misreporting?

once again, thank you for clearing our assumptions up!

may have even learnt something as to why DLM throws a big finger at us every so often!

#include <std_disclaimer>

Any comments made are personal opinion and do not reflect directly on the position my current or past employers may have.

hio77

'That VDSL Cat'
13036 posts

Uber Geek
+1 received by user: 3896

ID Verified

Trusted

Lizard Networks

Subscriber

#985133 12-Feb-2014 14:44

gabesmith:
gabesmith: So I now have a good line after some initial glitches I ironed out with Snap and Chorus a while back. I am 150-200m from the exchange and have 7db attenuation and very low CRCPMs being 0.01 and 0.03. I am on 43.4 Mbit/s with 8b profile and what understand to be DLM-1. I feel that DLM has kinda given up on my line after my initial glitches.

I am pretty certain my line should be capable of 50-55 Mbit/s sync in theory.

Do y'all think getting port reset would be a worth while experiment given my stats?

So update I requested a port reset from Snap via email. They came back with "This reset has been requested and you should note the DLM profile reset in the next 24 hours." I have not seen any evidence of a single re-sync during this time on my Fritz. I have replied to them on the ticket asking them to check this with no response yet. Sounds like a phone call to Snap is now required.... I have found in the past a port reset is a noticeable event that will result in a dsl re-sync at the very least. I am also assuming that a port reset will trigger off DLM again as well?

still hopeful of 17a one day.....

your line is another where by all our logic, you should get it, but DLM seems to have other ideas in mind for you.

i can only assume your electrical length value might just be over, even given your distance.

stevehodge:
BMarquis:
Hmm, all of our profiles have a max of 71680, not 70000. They are pushed to dslams from a single source based on templates and should all be identical.
Perhaps the fritz misreports in some cases? Have seen many, many cases int he past of modems mis-reporting figures.
the 54mbps upstream value here shows that too.

I'm hitting that limit though, so the Fritz would have to be misreporting the actual speed as well as the DSLAM limit. That seems unlikely. Is it possible there is some overhead counted in the 71680 that the Fritz is removing to get 70000?

with steves line, is where the max limit does get interesting.. considering it should be 71680

#include <std_disclaimer>

Any comments made are personal opinion and do not reflect directly on the position my current or past employers may have.

stevehodge

393 posts

Ultimate Geek
+1 received by user: 52

#985208 12-Feb-2014 15:51

BMarquis:
stevehodge:
BMarquis:
Hmm, all of our profiles have a max of 71680, not 70000. They are pushed to dslams from a single source based on templates and should all be identical.
Perhaps the fritz misreports in some cases? Have seen many, many cases int he past of modems mis-reporting figures.
the 54mbps upstream value here shows that too.

I'm hitting that limit though, so the Fritz would have to be misreporting the actual speed as well as the DSLAM limit. That seems unlikely. Is it possible there is some overhead counted in the 71680 that the Fritz is removing to get 70000?

You'll never hit the DSL line rate (being a layer 1 rate), so the 1.68mbps is probably not noticeable in any practical sense.

If you are sitting on an actual sync rate ("Current throughput on the FB) of 70000, pm me your address I'll look into it.

I've PM'd my address. I think I've figured it out though. 71680 = 70 Mbps if 1 Mbps = 1024 kbps. So there is some confusion over binary kilo/mega vs metric kilo/mega happening somewhere. I believe for line speeds 1M = 1000k = 1000000 is the usual definition, right?

Here are my current stats:

The high total error seconds is due to the connection having been up nearly 150 days.

BMarquis

466 posts

Ultimate Geek
+1 received by user: 581

Trusted

Chorus

Lifetime subscriber

#985237 12-Feb-2014 16:05

stevehodge:

I've PM'd my address. I think I've figured it out though. 71680 = 70 Mbps if 1 Mbps = 1024 kbps. So there is some confusion over binary kilo/mega vs metric kilo/mega happening somewhere. I believe for line speeds 1M = 1000k = 1000000 is the usual definition, right?

Possibly, but that doesn't explain why other people see 71680 though - unless different fritzbox software versions interpret the speed differently.
It would also make more sense if it said 70Mbps instead of 70000 though!

will look into it at our end based on your address.

BMarquis

466 posts

Ultimate Geek
+1 received by user: 581

Trusted

Chorus

Lifetime subscriber

#985240 12-Feb-2014 16:07

Anyone have the Max shown as 71680?
if so, PM me your address, so I can do some comparisons :)

stevehodge

393 posts

Ultimate Geek
+1 received by user: 52

#985286 12-Feb-2014 17:05

BMarquis:
stevehodge:

I've PM'd my address. I think I've figured it out though. 71680 = 70 Mbps if 1 Mbps = 1024 kbps. So there is some confusion over binary kilo/mega vs metric kilo/mega happening somewhere. I believe for line speeds 1M = 1000k = 1000000 is the usual definition, right?

Possibly, but that doesn't explain why other people see 71680 though - unless different fritzbox software versions interpret the speed differently.
It would also make more sense if it said 70Mbps instead of 70000 though!

will look into it at our end based on your address.

Silverwolf is running the newer firmware, I'm still on 5.22 (mine's a 7340). If the actual value is 71,680,000 bits per second then perhaps the older firmware uses 1024 bits per kb to convert and the newer firmware uses 1000. Might also partly explain the people who say other modems are faster.

Dyson

Shop now for Dyson appliances (affiliate link).

hio77

'That VDSL Cat'
13036 posts

Uber Geek
+1 received by user: 3896

ID Verified

Trusted

Lizard Networks

Subscriber

#985290 12-Feb-2014 17:17

purely out of an interest factor, i would be interested to see how the fritzboxes calcuation for distance (electrical length) compares directly to the DSLAMs calculation too..

#include <std_disclaimer>

Any comments made are personal opinion and do not reflect directly on the position my current or past employers may have.

Item

1739 posts

Uber Geek
+1 received by user: 726

Subscriber

#985380 12-Feb-2014 18:51

So I replaced my $20 VB104W from Slingshot with a shiny new ASUS DSL-N66U as the Wifi performance on the latter was crap...

Annoyingly, I have now gone from 32Mb/s down to about 22Mb/s down - I was anticipating the device at least matching, if not exceeding the old one. Upload speed remains solid at around 9.9Mb/s

Any pointers for troubleshooting this?

Item

1739 posts

Uber Geek
+1 received by user: 726

Subscriber

#985381 12-Feb-2014 18:52

^^^^^

Ok scratch that - found a newer firmware that sorted it!

:)

6297a1679f9

371 posts

Ultimate Geek
+1 received by user: 29
Inactive user

#986931 13-Feb-2014 18:32

stevehodge:

Here are my current stats:

The high total error seconds is due to the connection having been up nearly 150 days.

That current throughput.... What page was your story on?

hio77

'That VDSL Cat'
13036 posts

Uber Geek
+1 received by user: 3896

ID Verified

Trusted

Lizard Networks

Subscriber

#986932 13-Feb-2014 18:35

quite awhile back IIRC...

i dont remember seeing steves connection sitting at exactly max though!

#include <std_disclaimer>

Any comments made are personal opinion and do not reflect directly on the position my current or past employers may have.

Geekzone support

Support Geekzone with one-off or recurring donations Donate via PressPatron.

SilverwolfNZ

62 posts

Master Geek
+1 received by user: 3

#986975 13-Feb-2014 20:29

stevehodge:
BMarquis:
stevehodge:

I've PM'd my address. I think I've figured it out though. 71680 = 70 Mbps if 1 Mbps = 1024 kbps. So there is some confusion over binary kilo/mega vs metric kilo/mega happening somewhere. I believe for line speeds 1M = 1000k = 1000000 is the usual definition, right?

Possibly, but that doesn't explain why other people see 71680 though - unless different fritzbox software versions interpret the speed differently.
It would also make more sense if it said 70Mbps instead of 70000 though!

will look into it at our end based on your address.

Silverwolf is running the newer firmware, I'm still on 5.22 (mine's a 7340). If the actual value is 71,680,000 bits per second then perhaps the older firmware uses 1024 bits per kb to convert and the newer firmware uses 1000. Might also partly explain the people who say other modems are faster.

Hey the firmware I'm running is one customized by snap themselves I believe, maybe their rep on here could get a copy of it uploaded to mega or something of the sort?

hio77

'That VDSL Cat'
13036 posts

Uber Geek
+1 received by user: 3896

ID Verified

Trusted

Lizard Networks

Subscriber

#986978 13-Feb-2014 20:33

SilverwolfNZ:
stevehodge:
BMarquis:
stevehodge:

I've PM'd my address. I think I've figured it out though. 71680 = 70 Mbps if 1 Mbps = 1024 kbps. So there is some confusion over binary kilo/mega vs metric kilo/mega happening somewhere. I believe for line speeds 1M = 1000k = 1000000 is the usual definition, right?

Possibly, but that doesn't explain why other people see 71680 though - unless different fritzbox software versions interpret the speed differently.
It would also make more sense if it said 70Mbps instead of 70000 though!

will look into it at our end based on your address.

Silverwolf is running the newer firmware, I'm still on 5.22 (mine's a 7340). If the actual value is 71,680,000 bits per second then perhaps the older firmware uses 1024 bits per kb to convert and the newer firmware uses 1000. Might also partly explain the people who say other modems are faster.

Hey the firmware I'm running is one customized by snap themselves I believe, maybe their rep on here could get a copy of it uploaded to mega or something of the sort?

beta firmwares can be a tad dodgy...

given the 6.0 rollback issues, ild probably be rather careful about it..

#include <std_disclaimer>

Any comments made are personal opinion and do not reflect directly on the position my current or past employers may have.

stevehodge

393 posts

Ultimate Geek
+1 received by user: 52

#986990 13-Feb-2014 20:54

hio77: quite awhile back IIRC...

i dont remember seeing steves connection sitting at exactly max though!

I'm sure I mentioned somewhere that I'd tweaked it to hit the limit. Don't think I posted a screenshot before though.

hio77

'That VDSL Cat'
13036 posts

Uber Geek
+1 received by user: 3896

ID Verified

Trusted

Lizard Networks

Subscriber

#986991 13-Feb-2014 20:55

stevehodge:
hio77: quite awhile back IIRC...

i dont remember seeing steves connection sitting at exactly max though!

I'm sure I mentioned somewhere that I'd tweaked it to hit the limit. Don't think I posted a screenshot before though.

lack of screenshot would probably be why i didnt remember it..

photos are far more memorable for me..

#include <std_disclaimer>

Any comments made are personal opinion and do not reflect directly on the position my current or past employers may have.

1 | ... | 204 | 205 | 206 | 207 | 208 | 209 | 210 | 211 | 212 | 213 | 214 | ... | 285