|
|
|
Hi ,
Last week I was connected to a newly installed Huawei Cab, (I'm using a Billion 8800axl) and initially got a attainable rate of >105Mbps, with a correspondingly high SNR, however as more people have been connected to the cabinet my SNR has dropped and my Max attainable is hovering just above 80Mbps. At this rate I'll be down to 60Mbps by next week!.
I'm currently on Fast Path and the modem stats show that G.Inp is not enabled yet. My question is, when the Cabinet is upgraded to support G.Inp will all lines (with compatible modems) utilise it immediately, or will it only be applied to a line when a certain error rate threshold has been reached, or when SNR reaches a certain level?
|
|
|
I had an attainable of ~83, down from an initial attainable of ~95.
With G-INP, it's back up to ~87.
ZeN Unlimited Fibre 2
Fritz!Box 3390
|
|
|
I think most of us where under the impression that all Huawei cabinets where activated with G.INP...
I'm not sure on the compatibility of the 8800AXL, I remember someone having some compatibility issues over on the Plusnet forum's but not sure of the outcome on that one.
I can't see you loosing much more speed on your line, also bare in mind that the DLM might be increasing your SNR to increase stability, I would say that if the speed was to drop below 80 the SNR would start to drop to compensate!
|
|
Register (or login) on our website and you will not see this ad.
|
|
|
G.INP is a replacement for interleaving, so those on Fast path usually won't see it unless the DLM thinks the error rates need some form of error correction.
Most likely cross talk has been dropping your speed as others in the same cable bundle take VDSL2 - waiting on vectoring to resolve that.
|
|
The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
|
|
|
|
Thats interesting, I assume from that that you had fairly low error rates and were on fast path.
Which would point to the fact the G.INP has not been enabled on my cab yet.
|
|
|
G.INP is a replacement for interleaving, so those on Fast path usually won't see it unless the DLM thinks the error rates need some form of error correction.
DLM seems very eager to intervene with G.inp. Far more eager than to interleave.
|
|
|
I think most of us where under the impression that all Huawei cabinets where activated with G.INP...
It would be a fair guess that the problems with ECI modems has halted the G.INP rollout.
No more than a guess, mind, but the A&A reports suggest the issue has gone high in BT.
I'm not sure on the compatibility of the 8800AXL, I remember someone having some compatibility issues over on the Plusnet forum's but not sure of the outcome on that one.
I have an 8800NL with G.INP activated, working with Plusnet.
As @Ignition noted, DLM is far more eager to turn on G.INP than it was the old-style FEC/interleaving.
On an old line, I'd have 600 ES's per day (from 6,000 CRC's) which wouldn't trouble DLM. On my current line, I was getting just 60 ES's per day, but G.INP was still turned on.
that the DLM might be increasing your SNR to increase stability
It's a chicken/egg thing, but ...
On FTTC, DLM doesn't alter the target SNRM. Beyond the use of FEC/interleaving (old-style) or G.INP (combined with minimal FEC/interleaving, new-style), DLM's only other weapon is to use banding - or artificially limiting the speed.
Banding is, however, likely to cause an increase in the actual SNRM.
I think there are some technical advantages to do it this way around - probably in the way in which power settings vary, so reducing the potential for crosstalk.
|
|
|
Thats interesting, I assume from that that you had fairly low error rates and were on fast path.
Which would point to the fact the G.INP has not been enabled on my cab yet. I'm wondering if I'll ever see an effect. My line originally had an attainable of over 90 but gradually settled down at about 69 synch. For a while interleaving kept being put on then taken off then after BT rolled out a firmware update that stopped and I've been on fast path ever since.
If I'm on fast path then presumably G.INP is not needed but equally I wonder what happened to the interference that was affecting my line at first. I seem to recall that the firmware update enabled some bit swapping but don't remember. It's just possible there's a couple more Mb/s that G.INP could give me access to. Not that it would make the slightest difference but it's interesting.
---
Andrue Cope
Brackley, UK
Edited by Andrue (Wed 22-Apr-15 11:34:25)
|
|
|
That may well be the case
Keeping errors down and thus eventually less re-syncs will be the key to ensuring people don't moan about over 2 minute resyncs when vectoring appears.
|
|
The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
|
|
|
If I'm on fast path then presumably G.INP is not needed but equally I wonder what happened to the interference that was affecting my line at first. I have never had interleaving on my home connection and have had long, stable connection times. Attainable rate and sync dropped as crosstalk increased. I now have G.INP and am synced at 80000/20000 from the most recent pre-G.INP rate of about 75000/20000.
|
|
|
I have an 8800NL with G.INP activated, working with Plusnet.
Yes but I was talking about 8800 AXL...
|
|
|
I have an 8800NL with G.INP activated, working with Plusnet.
Yes but I was talking about 8800AXL...
This I was aware of.
Are you aware that they are the same Broadcom 63168 chipset? Highly likely that an incompatibility in one will appear in the other, and vice-versa, at least for firmware containing the same Broadcom driver blobs. This makes comparison between models a useful, but not definitive, thing to do.
|
|
|
I know they're nearly the same thing... I was lead to believe that there was an issue with firmware on the AXL that prevented G.INP from working ...
Whether they have fixed that now or not I don't know...
|
|
|
I know they're nearly the same thing... I was lead to believe that there was an issue with firmware on the AXL that prevented G.INP from working ... Can you supply a link to the source of that misleading information please?
|
|
|
|
According the Billion forums the latest beta dh65 firmware for the AXL does support G.INP
|
|
|
oh ok so theres a compatible beta firmware thats capable then, thats good..
Just to clarify I wouldn't class it working on a "beta" firmware as it compatible due to it requiring advanced installation rather than out of the box support..
Regardless this is good information to have, so thanks for the input!
|
|
|
Can you supply a link to the source of that misleading information please?
Its somewhere in the G.INP thread on Plusnet's forum...
Forgive me but I'm not exactly willing enough to trawl through all them posts just to prove a statement.
|
|
|
|
Well it's a bit of a nonsense as it stands as no-one knows if it's true or not.
|
|
|
|
|
|
|
|
No you didn't you told me it was somewhere.
|
|
|
Yes in the G.INP thread on Plusnet's Forum...
Here's a link: http://community.plus.net/forum/index.php/topic,1362...
|
|
|
|
Looks like you misread the post. Thought so.
|
|
|
|
|
|
|
andrew I think its the new default profile, adslmax on a line with less than 10 ES a day was put on g.inp.
As far as I know the only fast path users left are on ECI cabinets.
|
|
|
If I'm on fast path then presumably G.INP is not needed but equally I wonder what happened to the interference that was affecting my line at first. I have never had interleaving on my home connection and have had long, stable connection times. Attainable rate and sync dropped as crosstalk increased. I now have G.INP and am synced at 80000/20000 from the most recent pre-G.INP rate of about 75000/20000.
The answer was yes! I did get G.inp and actually it's quite a good boost.
Before:
http://www.thinkbroadband.com/speedtest/button/14297...
After:
http://www.thinkbroadband.com/speedtest/button/14298...
That's nearly taken me back to the maximum possible. I'm surprised and impressed.
---
Andrue Cope
Brackley, UK
|