Toshiba M200 with Intel wirless card - poor connection with Dlink DI-514

I have a Toshiba Portege M200 with the Centrino wireless adapter
(802.11b). It works fine on most networks but for some reason on my
home network which uses a D-Link DI-514 Wireless router, the
performance is terrible with packet loss, poor response time etc.

I have upgraded the firmware in the router to 1.02 per a note on
D-Link's site which says this fixes problems with Centrino chipset
machines.

I can ping my router and get response times from a few milliseconds to
several seconds to timeouts! I have tried all the usual stuff like
preample short to long and back again. I am running 64 bit WEP but all
other machines on my home network connect fine to it.

I would upgrade the drivers for the Toshiba if I knew which ones they
were.

Thanks

Larry

Larry wrote:

> I have a Toshiba Portege M200 with the Centrino wireless adapter
> (802.11b). It works fine on most networks but for some reason on my
> home network which uses a D-Link DI-514 Wireless router, the
> performance is terrible with packet loss, poor response time etc.
>
> I have upgraded the firmware in the router to 1.02 per a note on
> D-Link's site which says this fixes problems with Centrino chipset
> machines.
>
> I can ping my router and get response times from a few milliseconds to
> several seconds to timeouts! I have tried all the usual stuff like
> preample short to long and back again. I am running 64 bit WEP but all
> other machines on my home network connect fine to it.
>
> I would upgrade the drivers for the Toshiba if I knew which ones they
> were.
>
> Thanks
>
> Larry
Here are some setting in the router the will help.
Go to the tools tab > MISC button > lower your frabmentation to 2100,
Lower your RTS to 2300, and set the preamble to long.

See if that helps.

Koho

"Koho" <(E-Mail Removed)> wrote in message
news:yCJMc.2327$(E-Mail Removed) ...
> Larry wrote:
>
> > I have a Toshiba Portege M200 with the Centrino wireless adapter
> > (802.11b). It works fine on most networks but for some reason on my
> > home network which uses a D-Link DI-514 Wireless router, the
> > performance is terrible with packet loss, poor response time etc.
> >
> > I have upgraded the firmware in the router to 1.02 per a note on
> > D-Link's site which says this fixes problems with Centrino chipset
> > machines.
> >
> > I can ping my router and get response times from a few milliseconds to
> > several seconds to timeouts! I have tried all the usual stuff like
> > preample short to long and back again. I am running 64 bit WEP but all
> > other machines on my home network connect fine to it.
> >
> > I would upgrade the drivers for the Toshiba if I knew which ones they
> > were.
> >
> > Thanks
> >
> > Larry
> Here are some setting in the router the will help.
> Go to the tools tab > MISC button > lower your frabmentation to 2100,
> Lower your RTS to 2300, and set the preamble to long.
>
> See if that helps.
>
> Koho

Koho,

Could you please explain why these settings help. I'm particularly curious
since these values still exceed the 1500 octet maximum payload in Ethernet
frames.

Thanks,
Ron Bandes, CCNP, CTT+, etc.

On Mon, 26 Jul 2004 15:16:57 GMT, "Ron Bandes" <RunderscoreBandes
@yah00.com> wrote:

>> Here are some setting in the router the will help.
>> Go to the tools tab > MISC button > lower your frabmentation to 2100,
>> Lower your RTS to 2300, and set the preamble to long.

>Could you please explain why these settings help. I'm particularly curious
>since these values still exceed the 1500 octet maximum payload in Ethernet
>frames.
>Ron Bandes, CCNP, CTT+, etc.

Ummm... The default values for the DI-514 are at:
http://support.dlink.com/techtool/di...v_perform.html
RTS Threshold = 2432 (256->2432)
Fragmentation = 2436 (256->2436 even numbers only)

Both default values effectively disable flow control and
fragmentation. Basically, the values are set high because under ideal
conditions, neither is really necessary or desireable.

If you have hidden nodes or a large number of clients to deal with, a
lower RTS threshold of about 500 improves reliability at the expense
of performance. If you're getting hit with periodic interference
(implulse noise, microwave ovens), smaller packets are quite helpful.
Multipath is somewhat of a crap shoot. Either way, you need to
monitor retransmissions (not TCP resends) at the MAC level in order to
determine if any such changes are doing any good. You can kinda guess
the effect by looking at the "error" count.

A few articles on the topic:
http://www.wi-fiplanet.com/tutorials...le.php/1468331
http://www.wi-fiplanet.com/tutorials...le.php/1445641

I've also done some tinkering with long and short preamble. The
preamble is the non-data sync pattern sent before each packet to allow
the receivers to synchronize with the access point. It carries no
information and burns quite a bit of airtime. The old 802.11 (1-2Mbit
only) cards, will not associate if short preamble is set. Everything
else works well with short. However, I only gained about 2-3% in
thruput with my testing. Short is fine for everything I do.


--
Jeff Liebermann (E-Mail Removed)
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 AE6KS 831-336-2558