VDSL Performance & Tuning

Forums › New Zealand Broadband › VDSL Performance & Tuning

1 | ... | 183 | 184 | 185 | 186 | 187 | 188 | 189 | 190 | 191 | 192 | 193 | ... | 285

729 posts

Ultimate Geek

#945716 4-Dec-2013 22:06

TL;DR Downgraded Fritz firmware to 99.05.22. Back on DLM1. Happy!

Thought I would share another firmware downgrade success story. I must apologise for the lack of screenshots. I wasn't intending to make a post on this but since a few people have been talking about downgrading recently, I thought I should share.

Here's what everything looked like back in July when my Snap VDSL went live, using the latest firmware:

Given all those errors it wasn't surprising that DLM kicked in and I moved to DLM2 and then finally settled on DLM4. From this point I would get a resync maybe once a month where I would move from DLM4 to DLM2 and then back again as the error rates crept back up. My line sync also settled in around the 47000-48000 mark. I was happy, the speed was good, but we all know us geeks like to tinker...

On November 13 I finally got around to downgrading the firmware on my Fritz 7340 to the 99.05.22 version, and then proceeded to let DLM do it's thing naturally. I know I could have called Snap and got them to do a manual reset, but at the same time I was interested to see what would happen if left on it's own. I'm assuming at this point I was on DLM4 - hard to say as I believe the .22 firmware seems to report the INP values as zero even if that isn't true. I was on DLM4 before the downgrade. One noticeable difference was the lack of CRC errors, they were zero, where-as on the latest firmware it was always reporting a value <1 per minute.

On December 1 I had a resync which I believe moved me to DLM2, again hard to tell. Still zero errors.

On December 3 I had another resync and this time we were back at DLM1! The first thing I noticed was the CRC errors heading towards values of about 400 per hour. At this point I promptly changed the Line Settings (since they actually work on this version):

This resulted in a slight drop in download speed as expected, but it's still enough for my needs. Currently this is what I'm seeing (stable since 6am December 3):

So, if you're considering downgrading, my advice is to do it. This version behaves so much better and I'm yet to see any downside.

SamF

1578 posts

Uber Geek

Trusted

#945722 4-Dec-2013 22:16

@Kraven: Thanks for sharing that! Odd that you were dropped from DLM-1 originally; our testing has shown that a CRCPM up to 10 is fine on DLM-1, but I guess there may be other factors also at play.

Assuming your error levels stay low, you could probably go a bit lower (towards performance) on the SNRM as well IMO.

Kraven

729 posts

Ultimate Geek

#945742 4-Dec-2013 22:35

SamF: @Kraven: Thanks for sharing that! Odd that you were dropped from DLM-1 originally; our testing has shown that a CRCPM up to 10 is fine on DLM-1, but I guess there may be other factors also at play.

Assuming your error levels stay low, you could probably go a bit lower (towards performance) on the SNRM as well IMO.

That screenshot was shortly after plugging everything in. Over the next few hours those errors ramped up to values much, much higher.

I may look at lowering the SNRM but for the sake of a couple of Mbps it's not worth going through the whole DLM process again, I like my 12ms ping to work too much ;)

SamF

1578 posts

Uber Geek

Trusted

#945745 4-Dec-2013 22:43

Yeah fair enough. Looks like you have some room to move if you want to tho, which is nice.

sidefx

3711 posts

Uber Geek

Trusted

#945750 4-Dec-2013 22:50

SamF: @Kraven: Thanks for sharing that! Odd that you were dropped from DLM-1 originally; our testing has shown that a CRCPM up to 10 is fine on DLM-1, but I guess there may be other factors also at play.

My last drop from DLM1 to 2 happened with CRCPM well under 10; closer to 5 I think or not much over - which was odd because as you know I've run with much higher for extended periods previously. As far as I could see nothing else had really changed. I subsequently raised the TSNRM a bit to drop errors back down and after a good few weeks went back to DLM1 but haven't bothered to push the TSNRM back down since then. Perhaps they've dropped the threshold? Or yeah, other factors at play.

"I was born not knowing and have had only a little time to change that here and there." | Octopus Energy | Sharesies
- Richard Feynman

SamF

1578 posts

Uber Geek

Trusted

#945772 4-Dec-2013 23:11

Resyncs seem to play a large roll.

hio77

12999 posts

Uber Geek

ID Verified

Trusted

Lizard Networks

#945779 4-Dec-2013 23:43

SamF: Resyncs seem to play a large roll.

this. i was sitting with my slightly high CRC errors perfectly fine on DLM-1, until someone decided they were gonna try to find a cable to connect to the internet with....

seems to be about 2~3 resyncs in a day, and DLM-1 is gone.

#include <std_disclaimer>

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

Equinox IT

Cloud spending continues to surge globally, but most organisations haven’t made the changes necessary to maximise the value and cost-efficiency benefits of their cloud investments. Download the whitepaper From Overspend to Advantage now.

Dairusire

298 posts

Ultimate Geek

#946098 5-Dec-2013 15:31

So I've got something VERY interesting for you guys, one of my customers whos VDSL we've litterally just setup (within the last few hours), didn't even get a look in at 8b. litterally straight onto a 17a profile

Heres a picture to have a gawk at

hio77

12999 posts

Uber Geek

ID Verified

Trusted

Lizard Networks

#946099 5-Dec-2013 15:45

Dairusire: So I've got something VERY interesting for you guys, one of my customers whos VDSL we've litterally just setup (within the last few hours), didn't even get a look in at 8b. litterally straight onto a 17a profile

Heres a picture to have a gawk at

this seems VERY odd.. what attenuation is he at? distance from the cabinet?

am i reading the SNR correctly there? says its measured in 0.1db, so wouldnt that be 1.1db SNR? or is that possibly a firmware bug in the draytek?

#include <std_disclaimer>

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

Eitsop

583 posts

Ultimate Geek

ID Verified

#946589 6-Dec-2013 14:02

Kraven: TL;DR Downgraded Fritz firmware to 99.05.22. Back on DLM1. Happy!

How can I get a copy of older firmware?

Dean

hio77

12999 posts

Uber Geek

ID Verified

Trusted

Lizard Networks

#946626 6-Dec-2013 15:09

attewell:
Kraven: TL;DR Downgraded Fritz firmware to 99.05.22. Back on DLM1. Happy!

How can I get a copy of older firmware?

Dean

check the service ftp server.

ftp://service.avm.de/Downgrade/

#include <std_disclaimer>

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

Dairusire

298 posts

Ultimate Geek

#946636 6-Dec-2013 15:41

hio77:
Dairusire: So I've got something VERY interesting for you guys, one of my customers whos VDSL we've litterally just setup (within the last few hours), didn't even get a look in at 8b. litterally straight onto a 17a profile

Heres a picture to have a gawk at

this seems VERY odd.. what attenuation is he at? distance from the cabinet?

am i reading the SNR correctly there? says its measured in 0.1db, so wouldnt that be 1.1db SNR? or is that possibly a firmware bug in the draytek?

Yeah the SNR reading is probably a glitch. When we ran a pre-qual test on his lines it came back as a 1-1 attenuation.

hio77

12999 posts

Uber Geek

ID Verified

Trusted

Lizard Networks

#946649 6-Dec-2013 15:51

Dairusire:
hio77:
Dairusire: So I've got something VERY interesting for you guys, one of my customers whos VDSL we've litterally just setup (within the last few hours), didn't even get a look in at 8b. litterally straight onto a 17a profile

Heres a picture to have a gawk at

this seems VERY odd.. what attenuation is he at? distance from the cabinet?

am i reading the SNR correctly there? says its measured in 0.1db, so wouldnt that be 1.1db SNR? or is that possibly a firmware bug in the draytek?

Yeah the SNR reading is probably a glitch. When we ran a pre-qual test on his lines it came back as a 1-1 attenuation.

definitely under that 200M mark then.

whats the real world attenuation? 1-1, ild pick about a 2~3?

looks like DLM picked the distance and switched 17a straight away, interesting to see indeed, considering it takes most weeks+ to see 17a pop

#include <std_disclaimer>

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

Dairusire

298 posts

Ultimate Geek

#946651 6-Dec-2013 15:54

hio77:
Dairusire:
hio77:
Dairusire: So I've got something VERY interesting for you guys, one of my customers whos VDSL we've litterally just setup (within the last few hours), didn't even get a look in at 8b. litterally straight onto a 17a profile

Heres a picture to have a gawk at

this seems VERY odd.. what attenuation is he at? distance from the cabinet?

am i reading the SNR correctly there? says its measured in 0.1db, so wouldnt that be 1.1db SNR? or is that possibly a firmware bug in the draytek?

Yeah the SNR reading is probably a glitch. When we ran a pre-qual test on his lines it came back as a 1-1 attenuation.

definitely under that 200M mark then.

whats the real world attenuation? 1-1, ild pick about a 2~3?

looks like DLM picked the distance and switched 17a straight away, interesting to see indeed, considering it takes most weeks+ to see 17a pop

Nah, their attenuation is 1. Its not more than this, because well the cabinet is litterally on their front lawn and their pair runs directly from the cabinet to their comms room lol

SamF

1578 posts

Uber Geek

Trusted

#946652 6-Dec-2013 15:55

We've seen people go straight onto 17a before, and this appears to be a very good line, so it's not unheard of.

1 | ... | 183 | 184 | 185 | 186 | 187 | 188 | 189 | 190 | 191 | 192 | 193 | ... | 285