Page 1 of 1

Poor performance of the 3945abg adapter

Posted: Thu Apr 27, 2006 4:29 pm
by redburgundy
I just got my X60, which has the Intel 3945abg WiFi adapter, and it's sitting next to my X40 which has the Atheros Mini PCI adapter.
Both are connected to my DLink home router. Using the same version of Access Connections, the 3945 sees it with 80% signal strength and the Atheros sees it with 100% signal strength.
For the second strongest signal, the 3945 sees it with 22% and the Atheros with 77%. The 3945 sees a total of 4 routers with the two weakest at 17%, and the Atheros sees 7 with the weakest at 30%.
What should I do?
-Call IBM Support and complain?
-Buy an Atheros Mini-PCI Express adapter and install it myself?
-Live with the 3945?
-Something else?

of what?

Posted: Thu Apr 27, 2006 4:58 pm
by ninepoundjammer
100% and 80% of what? Are the scales identical? Normalized values?

Not necessarily meaningful.

You could try bandwidth tests by transfering a large file from a third machine on the network (as long as the third machine's connection isn't the bottleneck).

Posted: Thu Apr 27, 2006 5:29 pm
by trentblase
Plus, the difference could be in the antenna layout between the X60 and X40... you really need to compare two X60s with different wifi cards.

Re: of what?

Posted: Thu Apr 27, 2006 6:41 pm
by redburgundy
ninepoundjammer wrote:100% and 80% of what? Are the scales identical? Normalized values?
It's the scale that Access Connections uses. I don't know exactly what it means, but it implies that the RF link with the Atheros card has more margin than the 3945abg, and the Atheros card will work with a more distant access point than the 3945abg.
You could try bandwidth tests by transfering a large file from a third machine on the network (as long as the third machine's connection isn't the bottleneck).
Bandwidth tests aren't relevant--they're constrained by the cable modem. What I care about is trying to connect to a distant access point. The Atheros card seems better able to do that than the 3945abg.

Posted: Thu Apr 27, 2006 6:42 pm
by redburgundy
trentblase wrote:Plus, the difference could be in the antenna layout between the X60 and X40... you really need to compare two X60s with different wifi cards.
I agree with that. Has anyone done that?

Posted: Fri Apr 28, 2006 3:00 am
by BillMorrow
the differences could, indeed, be the antenna differences..
but more likely it is the wifi card..

just to add some salt to these wounds, my T43p got 100% sitting in the same place as my current T60p..
i have watched the signal strength go from 60% to 80% and back..
all the time just sitting in the same place the T43p got a consistant 100%..

the T60p has the intel card..

you can draw your own conclusions.. :roll:

Re: of what?

Posted: Fri Apr 28, 2006 10:58 am
by ninepoundjammer
redburgundy wrote: It's the scale that Access Connections uses. I don't know exactly what it means, but it implies that the RF link with the Atheros card has more margin than the 3945abg, and the Atheros card will work with a more distant access point than the 3945abg.
This info has to come from different device drivers which query different hardware. Unless somebody that actually knows how this is implemented explains why they are comparable, I would not consider them comparable.
redburgundy wrote: Bandwidth tests aren't relevant--they're constrained by the cable modem. What I care about is trying to connect to a distant access point. The Atheros card seems better able to do that than the 3945abg.
This is why I said on the same network (read same subnet, attached to the same access point). In networking, bandwidth is the primary measure of performance. Signal strength in and of itself doesn't necessarily equate good performance. I haven't seen any reliable evidence that either adaptor is performing poorly in the real world.

Re: of what?

Posted: Fri Apr 28, 2006 11:48 am
by redburgundy
ninepoundjammer wrote:
redburgundy wrote: Bandwidth tests aren't relevant--they're constrained by the cable modem. What I care about is trying to connect to a distant access point. The Atheros card seems better able to do that than the 3945abg.
This is why I said on the same network (read same subnet, attached to the same access point). In networking, bandwidth is the primary measure of performance. Signal strength in and of itself doesn't necessarily equate good performance. I haven't seen any reliable evidence that either adaptor is performing poorly in the real world.
What I care about is the ability to connect to an access point when the signal is weak. I'm willing to accept a lower data rate if that's what it takes to connect. But it seems that the 3945abg can't even see some of the access points that the Atheros card can see with reasonable signal strength.