New 802.11n wireless cards

[planetf1]

It seems there's a new Atheros AR5008 based mini-PCI express adapter coming from Lenovo. It's being shipped in some new models, but should be an option on others. It supports 802.11a/b/g/n, but it's most attractive feature is claiming improved coverage for 802.11b/g.

I'm currently experiencing issues with the stability of the 3945ABG and was about to order the current atheros card when I hit on this news.

Anyone know the official part numbers for ordering? (Europe)?

[dfumento]

802.11n is a big improvement over b/g. Just call IBM for the part number. I have it on my new laptop which is on order.

[fbrdphreak]

Configuring a Z61m with the 802.11n card, it is listed as PN 41U4780.

[JohnDrake]

A new intel wireless driver hit 11/17/06, try that first:

http://www-307.ibm.com/pc/support/site. ... 8#Wireless

[planetf1]

Ok, just to confirm I went with the atheros (ibm) a/b/g. Mt T60p only appeared to have 2 antenna cables, not the extra one needed for the n variant.

A world of difference.
a) Range is significantly improved
b) Connection is totally stable
c) Some blue screens I got every few days that looked networking related appear to have stopped (may be due to b/ above)

I just can't believe I took so long swapping the card. Fantastic..

No real feeling on power consumption yet (atheros card set for medium). Looks to be around same order of magnitude.

[kulivontot]

Wait, so let me get this straight, in addition to there being a different physical card for draft-n cards, there also needs to be an extra antenna in the machine? Thus making upgrades from ABG to ABGN impossible? Can anyone confirm this?

I just checked the lenovo site to see if there is indeed an FRU for a different antenna, and yes there is. Unfortunately, my model isn't listed under either the 3rd antenna FRU or the A/B/G/N card. I was planning on waiting until intel released their draft-n card, but I see now that going draft-n might be harder than I originally thought.

http://www-307.ibm.com/pc/support/site. ... 62784.html

Wireless LAN 3rd antenna (14.1in)
XGA

1951-CTO, 2Tx, 2Ux, 2Vx, 4Px, 63x
1952-CTO, 4Px
1953-CTO
2007-CTO, 2Bx, GFx
2008-CTO
2613-CTO
2623-CTO

SXGA+

1951-CTO, 4Qx, 62x
1952-CTO
1953-CTO
2007-CTO, 6Qx, 6Sx, 6Tx, 6Ux, 6Wx, 6Yx, 8Hx, 8Kx, 8Lx, 8Nx, GEx
2008-CTO, 6Qx, 8Hx, 8Kx, 8Lx, GEx
2613-CTO
2623-CTO, K3x

Wireless LAN 3rd antenna (14.1in)
FRU: 42R9909

[planetf1]

The antenna just looked like a piece of wire routed around the screen, so it may not be impossible to upgrade to 802.11n, just require a little creativity.

For me I couldn't wait or suffer the uncertainty, just needed something that worked more reliably than the intel card, and I seem to have found it...

[pinesol]

Ok, just to confirm I went with the atheros (ibm) a/b/g. Mt T60p only appeared to have 2 antenna cables, not the extra one needed for the n variant.
.....

I figured that that would be the case although nobody confirmed my suspicion. That is why I orderd 802.11n card along with my T60. Now I feel vindicated.

[dfumento]

Ok, just to confirm I went with the atheros (ibm) a/b/g. Mt T60p only appeared to have 2 antenna cables, not the extra one needed for the n variant.
.....

I figured that that would be the case although nobody confirmed my suspicion. That is why I orderd 802.11n card along with my T60. Now I feel vindicated.

Actually for an extra $35 and the significantly improved performance in range as well as bandwidth and the fact that by the next year we'll see significant number of 802.11n routers with adoption of the 802.11n standard, it only makes sense.

[jtheun]

I just ordered a T60p (2613CTO) w the following two lines in the FRU's:

42R9190 VBB NO INT.WIFI WL 3RD ANTENNA
41W1685 SBB 11ABGWIFI WL LAN USEULAANZ

What does this mean for 802.11n? Does it mean there is a 3rd antenna already built-in? How do I go about getting it upgraded to 802.11n? Can I call them and ask them to change the order?

Thanks for your help!

[pinesol]

I just ordered a T60p (2613CTO) w the following two lines in the FRU's:

42R9190 VBB NO INT.WIFI WL 3RD ANTENNA
41W1685 SBB 11ABGWIFI WL LAN USEULAANZ

What does this mean for 802.11n? Does it mean there is a 3rd antenna already built-in? How do I go about getting it upgraded to 802.11n? Can I call them and ask them to change the order?

Thanks for your help!

Yes, call them and have them change your order.

[jtheun]

Cool. Does anyone have the part number? I figure it will make the conversation w Lenovo a bit easier. Is it 41U4780 as mentioned by fbrdphreak for his Z61m?

[bfd3832]

I just ordered a T60, and talked to a Lenovo Representative, and I was told that my order will not come with the 3rd antennea, nor will it support 11n. Why is this? Model was 2613CT.

Part Number Description
42R9190 NO INT. Wifi WL 3rd Antenna
41W1685 11ABGWifi WL LAN USEULAANZ


What does this mean exactly? That was on my order print out. Will I have 11n capabilities?

Edit: Just realized this is like the post above. I would call back, but they told me that I didn't have a choice in getting the 11n? It isn't too big of deal for me I guess, but I don't understand why my new order isn't going to get it.

[bfd3832]

I just contacted Lenovo to change my order and get the 11n card, and they said it was not available for the T series currently, just te Z series. I asked if there was anyway to use the same card and have it built into my order, and was also told no, on that idea.

The rep said avaiablity for the n card for the T series could be a few days or a few months.

[rvacha]

According to the latest Tech Spec PDFs n is available for the T60, but I haven't tried to find a specific model

FYI

[kulivontot]

If you check my post above, all the 14" model numbers that the 3rd antenna fits in are listed. You can see the model numbers for the 15" versions by checking the appropriate FRU for the 15" one. What I want to know is, if I buy this part myself, can I add it in my LCD without having a supported model? I'm guessing you can, but there's no documentation on how the third antenna fits inside the lcd case. Is there anyone out there with a T60 with the new n cards who could take pictures of the disassembled lid? That way we can see if it's actually feasible to add 802.11n support to non-supported models.

[mikemi]

I just contacted Lenovo to change my order and get the 11n card, and they said it was not available for the T series currently, just te Z series. I asked if there was anyway to use the same card and have it built into my order, and was also told no, on that idea.

The rep said avaiablity for the n card for the T series could be a few days or a few months.

I have a T60 widescreen (8744-5BU) purchased Nov.30, 2006 that came with the Atheros 802.11abgn card.

[bfd3832]

Wow...then I was lied to possibly?

Its unfortunate that my order shipped today. No chance in changing it now.

[hoya]

I have a T60 widescreen (8744-5BU) purchased Nov.30, 2006 that came with the Atheros 802.11abgn card.

which router are you using that with? did you look to see if the card is attached to 3 antennas? I'm thinking of buying the DIR-655 since it seems compatible with the Atheros n card.

[Patrick B]

Just to add some information about why the third antenna.....there are two big things that 802.11n (draft) has brought to the table to improve performance.

The first of which is a move to a 40 MHz channel instead of the existing 20 MHz channel in 802.11 b/g. Double the channel size, then roughly double the throughput, assuming that no outside interference exists (a challenge, given that the frequency is license-exempt).

Two, the third antenna is for a technology called MIMO (multiple Input Multiple Output).

What MIMO basically does is one of three things: You can split the same signal over different antennas to avoid multi-path interference. IE, send the signal multiple times to ensure that you have a reliable data connection. This happens both on the Tx (Transmit) and Rx (Receive) stations. So if the same signal is transmitted AND received multiple times at each end, you can perform an additive function to get the best possible Signal to Noise Ratio (SNR, or CINR). The better the SNR, the higher the modulation rate, and subsequently higher throughput.

Two, MIMO allows you to split multiple signals across multiple paths, which means that bitstream A for example is sent to Tx 1, bitstream B is sent to Tx 2 etc. At the receiving end, a similar function takes place. This allows you to increase throughput, because you effectively have two transmitters and two receivers sending two different signals, allowing nearly double the throughput. Again, this only happens in ideal circumstances.

Three, a combination of the two above. So what can happen is that Tx 1 is sending bitstream A, Tx 2 is sending bitstream B, and Tx 3 is sending CRC data to maintain A+B. This allows both the increased throughput of Option 2 above, and the increased reliability of Option 1 above, albeit not to the tune of twice the throughput with twice the reliability (in reality, 50-60% greater throughput at 50-60% greater range).

Hence the reason why the 802.11n cards have a third antenna, and may not be fully useable in a machine with only two antennae.


Patrick

[pinesol]

Wow...then I was lied to possibly?

Its unfortunate that my order shipped today. No chance in changing it now.

Yes, you were lied to, but I don't think it was intentional on their part. Some salespeople don't know about it. Maybe it is intentional.

[pinesol]

Just to add some information about why the third antenna.....there are two big things that 802.11n (draft) has brought to the table to improve performance. ....

...
Patrick

Thank you for your comprehensive review on this matter. Although my T60 came with a 802.11n card, I am not sure I have three antenae (I see only two by examining the sides of the screen). How do I check it for sure?

[iatacs19]

I only bought my T60p in December I wish they have offered the 802.11n option back then...

[kulivontot]

So is adding a third antenna possible? anyone know?

[tomh009]

Just to add some information about why the third antenna.....there are two big things that 802.11n (draft) has brought to the table to improve performance.

The first of which is a move to a 40 MHz channel instead of the existing 20 MHz channel in 802.11 b/g. Double the channel size, then roughly double the throughput, assuming that no outside interference exists (a challenge, given that the frequency is license-exempt).

Almost, but not quite. 802.11b/g have a grand total of three non-overlapping channels, each one about 20 MHz wide:
Channel 1: 2412 MHz (2402-2422 MHz)
Channel 6: 2437 MHz (2427-2447 MHz)
Channel 11: 2462 MHz (2452-2472 MHz)

In practice, the signal "leaks" onto the adjacent frequencies from the centre of the channel, and that's why we end up with only three interference-free channels out of the original 11.

Most 802.11n draft devices use a single channel centred on the 802.11b/g channel 6. The trouble is that as soon as there are any 802.11b/g devices around, we end up with gridlock. In practice the 802.11n devices will have to back off to a 20 MHz-wide channel in order to allow 802.11b/g devices to operate on channel 1 and 11. End result? 20 MHz-wide channels in practice, and a throughput about half of what is advertised (isn't that the norm in wireless in general?).

Now, in the 5 GHz "a" band 802.11n won't have this problem, but most "n" implementations are "bgn" only ...

[Patrick B]

Actually, interference has little to do with channel selection by itself, as several outside factors can cause interference (most license exempt products tend to use the 2.4 GHz bands, such as most cordless phones, Bluetooth devices, wireless keyboards, mice etc).

And actually, it's INACCURATE to state that there are 3 non-overlapping channels in the 2.4 GHz band (1, 6, 11). In reality, the channel width of 802.11 b/g technology isn't hard filter specific, but rather limited by a specified spectral mask of the technology. It requires at least 30 dB of attenuation 11 MHz +/- from the centre frequency, and 50 dB of attenuation 22 MHz +/- from the centre frequency. So in theory, this should result in near non-overlapping operation. But it's not guaranteed.

As to 802.11n operation, it's important to note that the performance for 802.11n is based both upon the use of either 20 MHz or 40 MHz channels (802.11 G Turbo was effectively a 40+ MHz channel as well, albeit proprietary), AND the use of MIMO.

As I noted, a 40 MHz channel is only theoretically able to double throughput, assuming modulation rate and SNR are the same. So if one gets 54 Mbps (coded, over the air data rate) in a 20MHz channel, then you would at best get 108 Mbps in a 40 MHz channel under 802.11n.

MIMO is what accounts for more of the throughput than channel size, as I noted, because it allows bitstreams to be divided up using spatial multi-plexing. Option 1 as I noted in my previous post was using spatial diversity (multiple identical transmissions sent to multiple receivers....with the intent of increasing range).

Option 2 is spatial multi-plexing, which means different bitstreams sent to different transceivers and different receivers, and reassembled at the destination MAC. This allows one to increase throughput, dependent upon the number of antennas used (802.11n supports up to 4x4 MIMO).

Option 3 is a combination thereof....some antennas dedicated to spatial multiplexing and some to spatial diversity, with the intent to get the best of both worlds (increased range / increase capacity).

However, even in the 5GHz band, interference is still a possiblity....just significantly reduced, due to the fact that it's a wider license-exempt band. Poor channel selection / spectrum analysis won't result in any difference in performance, it's just that far fewer people use the 5 GHz band.


Patrick

[tomh009]

However, even in the 5GHz band, interference is still a possiblity....just significantly reduced, due to the fact that it's a wider license-exempt band. Poor channel selection / spectrum analysis won't result in any difference in performance, it's just that far fewer people use the 5 GHz band.

Exactly. Often one can see 5-10-15 networks advertising SSIDs, and then if you check on them, you find that they are all 2.4 GHz networks, with not one of them on 5 GHz. That's a good enough reason for me to run 802.11a, even at home -- the real-world throughput is much closer to the theoretical bandwith on 5 GHz.

As for the dreaded cordless phones and microwaves, I have yet to see them cause any significant interference, but, on the other hand, have seen frequent cases of lost connections caused by access point channel conflicts. 2.4 GHz is way overcongested -- it's the Los Angeles highway system of wireless networking.

[Patrick B]

However, even in the 5GHz band, interference is still a possiblity....just significantly reduced, due to the fact that it's a wider license-exempt band. Poor channel selection / spectrum analysis won't result in any difference in performance, it's just that far fewer people use the 5 GHz band.

Exactly. Often one can see 5-10-15 networks advertising SSIDs, and then if you check on them, you find that they are all 2.4 GHz networks, with not one of them on 5 GHz. That's a good enough reason for me to run 802.11a, even at home -- the real-world throughput is much closer to the theoretical bandwith on 5 GHz.

As for the dreaded cordless phones and microwaves, I have yet to see them cause any significant interference, but, on the other hand, have seen frequent cases of lost connections caused by access point channel conflicts. 2.4 GHz is way overcongested -- it's the Los Angeles highway system of wireless networking.

Simply to provide a counter point, I've seen several instances of other 2.4 GHz gear causing interference / dropped connections in a WiFi environment. It's really dependent upon the Client and AP as to how good they are at interference rejection.

Phones, Bluetooth etc all run significantly smaller channel sizes than WiFi however, which is why they *shouldn't* interfere as much.

Even in the case of interference, most users would never realize the impact, since cable / xDSL data rates are significantly lower than low to mid modulation rates on 802.11g. This is especially true in an OFDM environment.

For those same reasons, I run an 802.11a network at home....no interference, and precious few people attempting to hack in (since the vast majority of clients are 2.4 GHz....).

In an enterprise environment, the MAC is more often the limiting factor for throughput / performance of an 802.11 system, rather than interference. This is completely spectrum independent.



Patrick

[tomh009]

In an enterprise environment, the MAC is more often the limiting factor for throughput / performance of an 802.11 system, rather than interference. This is completely spectrum independent.

In a true enterprise, yes, for the most part. But many, many businesses share buildings or are located next door to other buildings (unless you are in a campus environment), and then there will be other access points beyond your control -- and frequently also interference, if you are using 802.11b/g.

It all comes down to understanding the network topology and the particular environmental constraints and choosing the most appropriate wireless technology, not just the one that has the largest Mb/s number on the box.

[Patrick B]

In an enterprise environment, the MAC is more often the limiting factor for throughput / performance of an 802.11 system, rather than interference. This is completely spectrum independent.

In a true enterprise, yes, for the most part. But many, many businesses share buildings or are located next door to other buildings (unless you are in a campus environment), and then there will be other access points beyond your control -- and frequently also interference, if you are using 802.11b/g.

It all comes down to understanding the network topology and the particular environmental constraints and choosing the most appropriate wireless technology, not just the one that has the largest Mb/s number on the box.

I would dispute that. Within a single building, interference can become a factor, but punching through exterior walls is generally a 12-15 dB drop; two building walls, and you're usually well below the 30 dB adjacent channel rejection issue.

Within a building it's more of an issue, as it takes less power to punch through interior walls / ceilings / floors. Proper antenna selection and AP location planning can minimize this. It's another reason why I wouldn't suggest that a company use gear from the local Future Shop as their WiFi APs; interference rejection is generally very poor on Consumer quality equipment.

Even in a single company building a poorly designed WiFi network can lead to interference being an issue.

Ultimately, good RF planning is necessary to work around this, and can often work around most issues. Most Enterprise / Carrier (har, har) gear can deal with interference *reasonably* well.

In ANY 802.11 network. the MAC remains the single biggest limiting factor, especially once you service more than 5-6 clients (and QoS is right out the window).


Patrick