Which wireless card to use

Message

yozas · #1 Post by **yozas** » Thu Aug 17, 2006 6:36 pm

Hi,

I want some advice, what wireless card model should I use
in my X40 (soon to own) notebook.
I know there are IBM miniPCI cards and INTEL miniPCI cards.
Those made by IBM has atheros chipset, and INTEL cards has an intel chipset (no surprise). If i would use an INTEL card, my laptop would be qualified as a "centrino" platform, but I want to know, would I lose energy saving or anything if I used IBM card instead?
Or maybe one of these cards has better performance and sensitivity, and I should not care about power saving?

Help me to decide, please

Wiz · #2 Post by **Wiz** » Fri Aug 18, 2006 6:23 pm

I guess that someone might disagree, but i prefer the IBM/Atheros. I had a T40 with a Intel 2100b and had a lot of disconnect problems. Then i bought a IBM/Atheros a/b/g wlan card to replace the Intel 2100b and all my problems were gone.
Now i have a T43p with a Intel 2200bg and also had a lot of disconnect problems until i found a known bug and a new driver to fix this disconnect issue. After i got the new driver i have less disconnect problems, but still the problems is not completely gone. It usually happen when i transfer a lot of data over the wireless network. So compared to the IBM/Atheros were i had no disconnect issues at all i'm not really happy with the Intel wlan cards. I also use a Cisco PCMCIA (CB21AG) and a Cisco PCMCIA 350 wlan card and neither does any of them disconnect like the Intel does.
I couldn't see that the Atheros used more power either since the battery did last just as long using the Atheros. I couldn't care less about the Centrino logo since it made my wireless less stable.

yozas · #3 Post by **yozas** » Sat Aug 19, 2006 3:16 am

interesting, thanks. well i did have some discnection problems
with special access points, not all of them, with my T43.
but then i got drivers directly from intel (ibm did not have
updated drivers for this card at that moment) and then those
problems were solved.

what i want to know is, what are other differences except power saving that intel card offers, maybe ibm card does offer same options too.

thanks

duffy · #4 Post by **duffy** » Sun Aug 20, 2006 1:09 pm

Intel cards prior to the latest generation cards seem to have alot of driver issues. I hated them. I did the same thing that one of the other posters mentioned by replaced the Intel card with the Atheros card. As far as benefits of one card over the other (excluding power saving) there aren't many. The Atheros card is more sensitive and picks up more WiFi networks than the Intel. I had a much easier time connecting to Cisco LEAP and PEAP networks with the Atheros card. FYI, the new Intel 3945ABG WiFi card available on new machines is much improved over the older Intel cards. I wouldn't hesitate suggesting that card to people. If you're curious the newer WiFi cards are mini PCI-E so they won't fit into the older machines that have mini-PCI slots.

yozas · #5 Post by **yozas** » Sun Aug 20, 2006 2:22 pm

I have a 2915 a/b/g card in my T43 thinkpad, i guess it is miniPCI express? cause T43 has one pcmcia slot and one expresscard slot, and video is pci express too if I remember right.

Guess I should stick with my IBM card in X40, now that I had read your post

planetf1 · #6 Post by **planetf1** » Mon Aug 21, 2006 11:24 am

I have an intel 3945abg (t60p) and it seems far inferior in pickup to the IBM a/b/g card I had in my T40 (upgraded from the original a/b).

I struggle with low strength /disconnects even 10m from an access point in the house (2 walls), whilst the old laptop could go everywhere.

Same at work...

tried the 10.1 (ibm) and 10.5 (intel) drivers.
Thinking I need to get a new card.

kstuart · #7 Post by **kstuart** » Mon Aug 21, 2006 7:08 pm

I'm sending this using an Intel 2200bg that is over 75 feet from the access point and I'm reading "100% signal strength".

No disconnects whatsoever.

Very latest Intel driver, Intel proset, and IBM AC. Windows ZCFG disabled.

PS I won't be buying a new Thinkpad until they come with Intel Wimax miniPCI cards (not yet released, of course).

DIGITALgimpus · #8 Post by **DIGITALgimpus** » Mon Aug 21, 2006 8:34 pm

Lately I've been having some disconnects with 2915... seems to be attributed to the busy spectrum in my area. I've seen a few reports of similar issues. I might try getting an IBM card... though the thought of spending $60 on this is just bothersome.

holr · #9 Post by **holr** » Wed Sep 06, 2006 7:14 pm

its funny that so many say they have had issues with intels card, and not with atheros' card. I have had the oppertunity to try two laptops, the t43p (which i own) with an atheros, and a sony vaio s5 series with intel wireless. Under linux (suse, fedora, ubuntu whatever) the intel card in the vaio outperforms the atheros in the t43p every time. I get many, many disconnects with the atheros card, while the intel one stays fully connected to wifi points at home, friends houses and at university.

However, this is using linux though. For performance between the two in windows? I'd say they are both comparable. If you dont plan to use linux, then get the atheros. If you do, get the intel.

#10 Post by **jdhurst** » Wed Sep 06, 2006 7:33 pm

@hoir: Your post was most fairly presented. I used Windows, and while I can make a newer Intel card work reliably, my Atheros card has never disconnected on any network. I guess that is why those of us using Windows recommend it. ... JD Hurst

yozas · #11 Post by **yozas** » Thu Sep 07, 2006 5:14 am

holr wrote: However, this is using linux though. For performance between the two in windows? I'd say they are both comparable. If you dont plan to use linux, then get the atheros. If you do, get the intel.

wow this is really strange, because intel drivers for linux was (and I thnk they are still are) very experimental and not stable. And support for atheros chipsets in linux started long time ago and is really reliable (madwifi).

And btw, there are really more options and tuning in atheros drivers for linux, capabilities too (access point, inject, RF monitor, etc)

Looks like some misunderstandment?

holr · #12 Post by **holr** » Thu Sep 07, 2006 5:43 pm

yozas wrote: wow this is really strange, because intel drivers for linux was (and I thnk they are still are) very experimental and not stable. And support for atheros chipsets in linux started long time ago and is really reliable (madwifi).

And btw, there are really more options and tuning in atheros drivers for linux, capabilities too (access point, inject, RF monitor, etc)

Looks like some misunderstandment?

As far as Im aware, the initial sets of intel drivers for wifi were indeed horrible, but i was using the latest / greatest ipw2200 for this particular intel card and it was very nice with fedora core 5.

I must admit however, I have just changed my drivers on the t43p (running suse 10.1) from the "madwifi" ones to using "ndiswrapper" (makes use of native windows drivers) and it seems to be a lot better. I will keep you posted; I have only made the change tonight. The wifi bars are showing a stronger connection anyway!

yozas · #13 Post by **yozas** » Fri Sep 08, 2006 6:51 am

holr wrote: I must admit however, I have just changed my drivers on the t43p (running suse 10.1) from the "madwifi" ones to using "ndiswrapper" (makes use of native windows drivers) and it seems to be a lot better. I will keep you posted; I have only made the change tonight. The wifi bars are showing a stronger connection anyway!

well i think card sensitivity can't change because of drivers, so I guess the drivers now just report strength in a different way.
What I have read all around, is that people don't recommend using ndis wrapper if any other solution exists.
I know that madwifi drivers are loading some code that is not open source, and I think is the same that windows drivers are loading because atheros must have given that little binary piece to linux drivers developers?

alexzabr · #14 Post by **alexzabr** » Sat Sep 09, 2006 1:47 pm

To throw my two cents...my machine has Intel standard 2200BG, at least at home works fine throughout the place (up to about 20-25m away (60-80ft) from my WiFi point. No disconnects at all (actually very rare) even during sustained work (surfing, emails, downloads).
Drivers are probably not the latest, but definitely one-two versions back at most. (I'm not a driver update geek - no reason to touch the things as long as they work fine)

yozas · #15 Post by **yozas** » Sat Sep 09, 2006 4:27 pm

alexzabr wrote:(I'm not a driver update geek - no reason to touch the things as long as they work fine)

and my one cent:
That statement about updating or not updating drivers actually isn't very true nowadays..
Because old drivers also tend to have not just performance/stability improvements, but also several security fixes.

#16 Post by **bill bolton** » Sat Sep 09, 2006 5:34 pm

yozas wrote:well i think card sensitivity can't change because of drivers

While you can choose to believe whatever you wish, that is not a well founded belief.

Cheers,

Bill

DIGITALgimpus · #17 Post by **DIGITALgimpus** » Sat Sep 09, 2006 7:51 pm

Wireless chipsets often have more functionality and fine tuning than programmers ever end up using... as problems arise, and as time/budget permits, or market conditions require, they can enable/modify behavior. So yes drivers can make a HUGE difference. A perfect example is the evolution of linux drivers. Early versions tend to be slow, and use lots more battery than more mature later versions, which support things like low power mode, etc. etc.

Keeping drivers up to date is important. If you don't care about anything else, it's also a security issue.

Aroc · #18 Post by **Aroc** » Sat Sep 09, 2006 11:27 pm

Drivers can make a HUGE amount of difference. I used to think the Intel wireless cards were junk until I tried some of the recent drivers from this year. The 2200BG and 2915ABG are actually [censored] good cards. I'm looking to replace my Atheros with a 2915ABG now after using these at work. Intel had finally delivered.

This is not a suprise. For years their desktop fast and gigabit ethernet adapters have been top notch. A fair bit contributing to that quality was the quality of the supplied drivers.

yozas · #19 Post by **yozas** » Sun Sep 10, 2006 6:34 am

bill bolton wrote:
yozas wrote:well i think card sensitivity can't change because of drivers
While you can choose to believe whatever you wish, that is not a well founded belief.
Bill

well maybe.. but drivers can't change bad radio hardware inside card, can they..

DIGITALgimpus · #20 Post by **DIGITALgimpus** » Sun Sep 10, 2006 9:49 pm

yozas wrote: well maybe.. but drivers can't change bad radio hardware inside card, can they..

Do you have specific references to the bad radio hardware? AFAIK the hardware doesn't seem to have any issues (it's even 802.11 *certified* which you can verify here).

There doesn't seem to be any evidence of bad hardware at all, but drivers aren't quite there.

yozas · #21 Post by **yozas** » Mon Sep 11, 2006 5:22 am

DIGITALgimpus wrote: Do you have specific references to the bad radio hardware? AFAIK the hardware doesn't seem to have any issues (it's even 802.11 *certified* which you can verify here).

There doesn't seem to be any evidence of bad hardware at all, but drivers aren't quite there.

Well i should have noted that I was not talking just about intel or atheros cards..
What I wanted to say that if any cards radio hardware is bad / medium /whatever, the drivers can't change that problem much.
And that link for finding certified devices made me laugh a bit, cause everybody (well maybe every-wardriver to be correct) knows that even if broadcom (which products can be found there as certified) is certified, it has a crappy radio (sensitivity).
So much for certification

alexzabr · #22 Post by **alexzabr** » Mon Sep 11, 2006 7:23 am

yozas, I don't know how deep your're involved into actual hardware world in terms of design and general engineering, but must admit your arguments sound somewhat ignorant.
Contemporary technology is a great equalizer for manufacturers. It can be safely assumed that today's major manufacturers of any technological equipment offer hardware design quality that fails very closely to each other, with ony minor variations, (mostly probably in reliability). Otherwise, no one of them would survive is such competitive market as today's technology. However, the products that combine sophisticated hardware with software (or firmware) often exhibit that quality gaps tendency, but this is mostly due to software/firmware divergencies between the manufacturers. Time-to-market pressure is very tough, so at a times (actually more often then we, customers would like to tolerate) the producst are being pushed to market in non-finished condition such as software drivers aren't thouroughly tested and stabilized, firmware that isn't finished yet in all its features, etc...

yozas · #23 Post by **yozas** » Mon Sep 11, 2006 7:46 am

alexzabr wrote:yozas, I don't know how deep your're involved into actual hardware world in terms of design and general engineering, but must admit your arguments sound somewhat ignorant.

What can I say, I am just reading lots of wardriver forums,
and I myself have some pcmcia cards to compare when wardriving.
I would not mind if different cards were just slightly different in sensitivity, but they sure are not.

alexzabr wrote: Time-to-market pressure is very tough, so at a times (actually more often then we, customers would like to tolerate) the producst are being pushed to market in non-finished condition such as software drivers aren't thouroughly tested and stabilized, firmware that isn't finished yet in all its features, etc...

I surely agree with all things about drivers and firmware, but like I said I don't agree that every modern card differs only slightly in sensitivity. I have a bit more than mediocre knowledge in electronics, and have opened lots of pcmcia cards, and radio part differences are really huge between different cards. Some models are just designed to be satisfactory (and still fall in certification range) and some are many times better.

alexzabr · #24 Post by **alexzabr** » Mon Sep 11, 2006 8:01 am

Well, I don't know how can one judge hardware performance quality by opening the device and looking at its "bones"...
I'm not familar with waforums (actually have no idea what they are), but doesn't sound as some kind of reliable technical info resourse.
You must learn the specifications of the design and the specifications of design parts that are in charge of particular performanc tasks in order to make conclusion how bad or good is the particular hardware alone.
The fact that some operate better then others (such as noticeably better range, for instance) still points probaly to entire conbinaiton of software/firmware/hardware where you're unlikely to be able to point specifically to hardware or software/firmware issues without familiarity with design/testing specifics....

yozas · #25 Post by **yozas** » Mon Sep 11, 2006 8:16 am

alexzabr wrote:Well, I don't know how can one judge hardware performance quality by opening the device and looking at its "bones"...

Not too much magic really

As an example some manufacturers omit few components at the radio stage in a card. That makes the card cheaper but it still works and complies to standards. Smart manufacturers?
And another thing by which one can judge the quality and capabilities of a wireless card radio is to see what chip is used at an antenna stage selector/amplifier..

alexzabr wrote: I'm not familar with waforums (actually have no idea what they are), but doesn't sound as some kind of reliable technical info resourse.

Those forums are called wardriver forums (and there are a lot of them, like http://www.kismetwireless.net/forum.php, which is one of some linux softwares for sniffing wifi), I can say that I am very sorry if you do not know what wardriving. Maybe that's one of the reasons you can't realise how big card sensitivity differences are, and why they are important.

alexzabr wrote: You must learn the specifications of the design and the specifications of design parts that are in charge of particular performanc tasks in order to make conclusion how bad or good is the particular hardware alone.

I can say that I get to read specific datasheets about exact chips that are responsible for radio part in a card. And that is one of the reasons I state that if a card has a poor radio part in it, that can't be improved (much) just by changing firmware, cause there is a direct limit in hardware...

And by the way, there are some cards that use the same firmware but their sensitivity still differs, which proves everything.

#26 Post by **JaneL** » Mon Sep 11, 2006 8:20 am

Keep it civil, now, guys...

alexzabr · #27 Post by **alexzabr** » Mon Sep 11, 2006 8:28 am

Oh, common, how can you now they use the same firmware ? Do you happen to be invloved in their firmware development or you have some other internal info ? Even quite small firmware changes can make big operating difference in general.
I have no interest in Linux in general and in wireless under Linux in particular, so no reason to be sorry for me not getting into these forums. I use my thinkpad as the tool to achieve my study/work objectives, so far have been satisfied by my thinkpad's wireless (it has 2200BG card). And I usually refrain from making an ignorant claims in public such as "bad hardware" in such complex systems unless I gain a deep technical understanding of the matter.

alexzabr · #28 Post by **alexzabr** » Mon Sep 11, 2006 8:30 am

nonny wrote:Keep it civil, now, guys...

Sorry on my part...just couldn't let it pass on...

yozas · #29 Post by **yozas** » Mon Sep 11, 2006 8:44 am

alexzabr wrote:Oh, common, how can you now they use the same firmware ? Do you happen to be invloved in their firmware development or you have some other internal info ?
Even quite small firmware changes can make big operating difference in general.

Not only I do know that the same utilities and firmware files are used for some cards made by different manufacturers (they have the same chipset, but different radio part). I am sure there are some people here to prove this.
I do agree about small firmware changes, but this is not the case, they are the same.

alexzabr wrote: I have no interest in Linux in general and in wireless under Linux in particular, so no reason to be sorry for me not getting into these forums. I use my thinkpad as the tool to achieve my study/work objectives, so far have been satisfied by my thinkpad's wireless (it has 2200BG card). And I usually refrain from making an ignorant claims in public such as "bad hardware" in such complex systems unless I gain a deep technical understanding of the matter.

Well I am geek a bit

So I give some of my free time to linux and I even use it in/at my work. Wardriving with linux is the only choice that I have at my hands for making sure that company wifi network is ok. As a side effect

it helps understand how different stuff works and is interconnected with each other.
By the way, some level of wardriving can be done in windows too.

Now I have tested some wireless cards (intel and atheros) in my thinkpad and can say that both of them sensitivity and driver stability is good. I wish it was like this with other cards.

good luck!

ottadel · #30 Post by **ottadel** » Mon Oct 23, 2006 11:50 pm

I have an R-40 2681, and wich card I can use ??

Which wireless card to use

Which wireless card to use

I have the same question

Who is online