Print

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

[rpdmallett]

Hi,

I had ADSL switched over to FTTC about a month ago and have been trying to get a very low upstream rate diagnosed.  I'm 1.9km from the cabinet and downstream is very good at 16M (was ~9M on ADSL), but upstream is very poor at ~538k (was double that at ~1037k on ADSL).

So trying to figure out this upstream loss has been very interesting.  Since the copper hasn't changed, and the downstream is very good, and Plusnet have been very good at trying to isolate any faults (including an unbalanced phone line to the exchange), I'm left with "that's all the line can do".  But of course I know it can do more, since ADSL was better on a ~3km line.  VDSL2 uses the same US0 bins as ADSL2+, and even has a higher potential power level, so what could be going wrong?

Well I've noticed that the upstream PSD, as reported by my Draytek 2860, does vary quite a lot - despite the upstream rate never really changing.  The table below hopefully shows this:

Date          PSD   SNR   kbps
18/1/17   9.6   7   528
3/2/17   9.6   7   528
6/2/17   9.0   7   528
11/2/17   8.7   8   538
11/2/17   9.4   8   538
12/2/17   9.6   7   535
14/2/17   8.9   8   542
17/2/17   9.6   7   538
19/2/17   9.1   7   538
19/2/17   9.2   7   538
20/2/17   9.5   7   531
21/2/17   9.8   7   535
22/2/17   9.5   7   528

(The Draytek modem doesn't give a decimal point value for the SNR, so '7' could be '7.9'.)

So, despite the PSD varying from 8.7dB to 9.8dB, the upstream rate only varies from 528k to 542k.

It's as if the ECI cabinet (yes, I know... I'm already damned...) has 'decided' that ~530k is what should be aimed for and the PSD for my modem is adjusted to achieve this (using Upstream Power Back Off?).

Does this sound right?  Anyone come across anything like this.  ECI firmware is 0xd086, with no G.INP.

Thanks,
Richard

[forceware]

Hi, I am the same distance from the cabinet as you I have the same download speed and the same terrible upload. I've had countless engineer and was told it was as good as it get for the distance.

[rpdmallett]

Hi, I am the same distance from the cabinet as you I have the same download speed and the same terrible upload. I've had countless engineer and was told it was as good as it get for the distance.

Do you know the cabinet type you're connected to?

And the firmware version, if it's ECI?

[forceware]

Yes eci version 206

[forceware]

BT wholesale checker upload says .8 impacted and 1.2 clean but I'm .5

[burakkucat]

BT wholesale checker upload says .8 impacted and 1.2 clean but I'm .5

Please remember that those estimates are for the circuit to the DP and not to your NTE5.

[j0hn]

You have the dreaded 0xd086 ECI DSLAM firmware.
I've only seen this once, but upon updating to that firmware it cut the users upload in half. I would concentrate on telling them the DSLAM doesn't have the latest firmware and try that.
Here's the initial post with that strange firmware
Here's the thread made about it.
Every other firmware I've seen on ECI DSLAM's has been either 0xb204 or the newer/latest 0xb206.
Also

I'm left with "that's all the line can do".  But of course I know it can do more, since ADSL was better on a ~3km line.

Your getting that all wrong. VDSL2 signal degrades much quicker over longer distances. A little further from the cab and you would get no service at all.

Hope this helps 

[rpdmallett]

You have the dreaded 0xd086 ECI DSLAM firmware.

Thanks - I've read those 2 posts, and been in contact with user npr.  However, are you saying that 0xd086 is OLD firmware?  I guess there's no published list from ECI with the firmware revisions?  And user forceware has the same issue with '206.

Also Your getting that all wrong. VDSL2 signal degrades much quicker over longer distances. A little further from the cab and you would get no service at all.

From what I've read, this is not the case for the first upstream band US0.  That uses the same bins/frequencies as ADSL2+.  VDSL2 even has a higher potential output (upstream) power.  Of course, VDSL2's US1, US2, etc. bands all quickly die off - but as far as I can see US0 should behave the same as ADSL2+'s US band.

If I'm getting that wrong, please let me know.  Thanks.

[ejs]

While looking for your thread on the Plusnet forums, I also found a different thread in which someone reported getting a much better upload speed using very old Lantiq modem firmware (the classic 5.4.8.0.0.6). It's probably not applicable to you because it was about much faster speeds, but not a lot to lose by trying it.
https://community.plus.net/t5/Plusnet-Feedback/Dramatically-better-upload-speed-with-using-Non-Vectoring-Modem/m-p/1411743

[rpdmallett]

I also found a different thread in which someone reported getting a much better upload speed using very old Lantiq modem firmware (the classic 5.4.8.0.0.6).

Thanks for that - but having just given it a go it actually lowers both my up & down stream speeds.  I've attached the Draytek stats for that firmware, and the 'BT' one (576D17).

[burakkucat]

Surely bit-swapping should be enabled on each device, at both ends of the xDSL link?

[ejs]

@burakkucat
I think the bit-swapping status is often reported strangely on these devices, changing to ON if it's done any bit swaps recently, then going back to OFF.

From what I've read, this is not the case for the first upstream band US0.  That uses the same bins/frequencies as ADSL2+.  VDSL2 even has a higher potential output (upstream) power.  Of course, VDSL2's US1, US2, etc. bands all quickly die off - but as far as I can see US0 should behave the same as ADSL2+'s US band.

I think the higher potential upstream power for VDSL2 applies to the aggregated total upstream power summed over all the upstream bands, and the power limits for the US0 frequencies should be about the same as ADSL2/2+.

Is the vectoring status reported anywhere?

[burakkucat]

@burakkucat
I think the bit-swapping status is often reported strangely on these devices, changing to ON if it's done any bit swaps recently, then going back to OFF.

Ah, thank you. I always have trouble in understanding the statics reported by Draytek devices.