Print

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

[waltergmw]

Gentlefolk,

I thought you might be interested in the following pictures which show the transfer of a Zen capped service on a TI DSLAM to a standard ADSL MAX service (Automatically without a request) to an Infineon MSAN / DSLAM.
This line is one of our horror stories I keep on constant monitoring via the 2700 HGV modem and a LogMeIn laptop.

During the 21CN work @ 01:40:57 on 15 Mar 2011 the TI DSLAM connection tripped and at 01:42:01 the Infineon DSLAM / MSAN started to attempt sync.
It achieved sync at 01:42:43 on a standard max ADSL service (and NOT the capped service we had to have initially as the line then was so unstable).
It was then in Yo-Yo mode up to just after 03:21:09.
At this point I don't know exactly when this session started as the final Date / Time reading is "time now". It only reverts to start of session time when the service trips and re-syncs.
Note also that the re-syncs happen relatively quickly so the ISP's data logs don't usually record a PPP disconnection for these short-term drops.

After that the line stabilised with a download sync speed of 960 Kbps and upload of 384 Kbps. It has remained without tripping now for 36 hours which looks very promising, albeit the throughput speeds are still rather poor.

I'm also attaching a new BT Speed tester picture which provides both down and upload data. Note the downstream IP profile data now shows 749 rather than 750 and a new Upstream IP profile figure has been added which is the same as the upstream sync speed.

Kind regards,
Walter

[attachment deleted by admin]

[waltergmw]

Gentlefolk,

I've been monitoring 7 services in the area and am still considering these results as in the early days.
One rather surprising point, at least to me, is that the 21CN services still on ADSL1, seem faster by about 10% and that they are more stable on our horrible long lines.
If this continues the implication is surely that new firmware, and possibly new hardware, has been improved.

Given the warnings we've had, I suspect most won't contemplate ADSL2+ services here.

Kind regards,
Walter

[roseway]

One rather surprising point, at least to me, is that the 21CN services still on ADSL1, seem faster by about 10% and that they are more stable on our horrible long lines.

That is, as you say, a bit surprising. The lines are the same, so the only difference is the Infineon DSLAM/MSAN. I suppose that it's the case that the 2700HGV and Infineon equipment work well together.

[waltergmw]

@ Eric,

Perhaps even more surprising is that now, over 7 days later, the line hasn't dropped once whereas before it used to go into yoyo mode every couple of days.
(Remember also that the 2Wire 2700 HGV records very short drops which are not seen after the ISPs data filtering.)

Even another AAISP service which hasn't yet been "21CNed" at the same address using the same cable is more stable.
( I observe this as that service only provides the single download BT speed test and it still reports its old DSLAM / MSAN as "Unknown".)

I wonder if these observations suggest a noisy piece of exchange equipment has been decommissioned too ?

Kind reagrds,
Walter

[Azzaka]

What chipset is in the 2700? Is it running an Infineon or a Broadcom?

[roseway]

What chipset is in the 2700? Is it running an Infineon or a Broadcom?

It's a custom 2-Wire chipset.

[Azzaka]

Hmmm...

I know the Broadcom chips have issue and the Infineon work well. I wonder if the 2Wire chip uses a similar DSL code to the Infineon... Anyone wish to hazard a guess?

[roseway]

It's not impossible. I've certainly noticed similarities in performance between the 2700HGV and the DG834Gv3 with the latest (Infineon) firmware, but that may just be coincidence of course.

[waltergmw]

Gentlefolk,

There is only one small hint just after reboot within thousands of log file entries which I've reproduced below.

INF
 INF P0000-00-00T00:00:01 wave:  Atheros driver version: 2.0.0
 INF P0000-00-00T00:00:06 scc:  SCC DMA started
 INF P0000-00-00T00:00:09 initd:  ---------------------------------------------
 INF P0000-00-00T00:00:11 cm:  cm initialized
 INF P0000-00-00T00:00:11 initd:  libcm started

There's a little bit more information here:-

http://www.modem-help.co.uk/2Wire/2700HGV-ADSL2--11g-4P-VoIP-Intelligent-Gateway.html

I have a couple of photos I'll e-mail to Eric & Azz and the word Infineon can clearly be seen on one large IC.

It should also be noted that although 2Wire (Now owned by Pace?) report that the modem is IPv6 capable, I suspect that it will need a new firmware version to enable it.

Kind regards,
Walter

[roseway]

the word Infineon can clearly be seen on one large IC.

It's probably my eyesight, but I can't see 'Infineon' there.

[waltergmw]

Hi Eric,

Have a look at the new photograph now.

K R
Walter

[roseway]

Thanks Walter, I see it now. I was looking at the big 2-Wire chip.

[Azzaka]

Looks like the Atheros is for the Wireless, the Infineon is for the DSL, and the 2Wire Chip is for the main OS of the modem.

[roseway]

Yes, it looks like that.

By the way, Digidave is selling the generic 2-Wire 2700HG again, and seems to have stock. I've ordered one.

[waltergmw]

Gentlefolk,

Herewith even more astonishing results from 2Wire 2700HGV modems.
The first picture shows the migration to 21CN Infineon MSAN now with 15 days of continuous operation at the slightly faster speed and only 8,701 CRC errors.
By contrast the second picture shows a service with 8 drops still on 20CN Unknown DSLAM at the same house fed from the same cable.
The road distance for the cable is just over 7 km.

I'm absolutely delighted with the infineon although I would have been happier if the TI DSLAM had been replaced a long time ago and thus reduce the amount of time we've all devoted to this difficult case.
When on the TI the 2700 HGV was the ONLY modem that would connect properly.

Kind regards,
Walter

[attachment deleted by admin]