D-Link DIR-655 Xtreme N Gigabit Router Review:Draft 2.0 arrives

[geogecko]

Thanks, I will give that a try!

 

[geogecko]

Hmm...nothing gets along on channel 1. I remember that being the problem trying to setup my WRT54G on channel 1, and not being able to connect anything to it. It is strange though, I do a channel scan using the WRT54G (with Tomato), and find no APs using that channel, at least, in 2.4GHz. Something is blasting it though. After I changed to channel 1, I could no longer connect to the router, even though I was only 10 feet away from it. I had to do a hard reset, and restore the original configuartion file... Unfortuantely, there are a lot of WiFi radios in my neigborhood.

Nothing like running Cat5/6 through your house, I guess. I wonder if adding the 5GHz radio (DAP-1522) would help at all?

 

[geogecko]

Is there any type of application (of course, free is best) that can test a wireless connection between a client and the AP?

 

[thiggins]

Could be that someone close by on Chan 1 (or 2 or 3) is blocking SSID broadcast.

Try Netstumbler.

 

[geogecko]

Ok. I was finally able to move the DAP-1555 into the same room as the router. Now my DWA-160, DIR-655 and DAP-1555 are in the same room, each about 10 feet from each other.

I repeated the throughput testing, and I am still around 17Mbps, between the DWA-160 computer, through the router, to the DAP-1555 connected computer.

Signal strength on the DAP-1555 when in the desired room was between 86% and 100% all the time, now it shows 100%.

So it does not appear to be a range related issue, from what I can tell. Even if I kill the WRT54G, the range is still down about the same level.

If I connect the computer that is normally connected to the DAP-1555, directly to a wired port on the DIR-655, then the throughput goes up to about 30Mbps.

Any other ideas? It seems like part of the problem is with the DAP-1555, but there may also be issues with the DWA-160. Unfortunately, I do not have a way to hardwire the computer with the DWA-160 directly to the router (the on-board Ethernet port does not work).

 

[gunnar-eee]

Accomplishing a good DIR-855 Wifi setup ?

Hello I have a similar question. I am blessed with a 100 Mbit/s downlink at home. I just connected a DIR-855 to this 100 Mbit/s WAN. The router does not choke the performance over LAN, I get pretty much the same throughput on a LAN swith port as if I connect my PC directly to the incoming WAN port, appr 97 Mbit/s measured from an internet performance test server. Pretty nice.
However I have a problem achieving good link speeds between the 855 and Wifi adapters on a laptop, both internal miniPCI-E and USB dongles. However when I connect my PC's over cable to one of the switch ports I get the 97 Mbit/s. So the problem is Wifi related.

Best Wifi result is with my Asus Eee 1000H laptop that contains a miniPCI-E Ralink based card (Azurewave), 2.4 GHz Draft-n. It can download at appr 45 Mbit/s peak, average appr 35 Mbit/s when meassured using as Broadband test site. While downloading the Wifi connection link never seems to indicate over 130 Mbit/s, most often says 65 Mbit/s. The speed is very sensitive to distance and if I go out-of-sight of the router.

Worst experience is using a DWA-160 which is supposed to be a recommended DIR-855 combo ?
Speeds are then decent up to a few metres from the router then drops quickly, however not past an 120 Mbit/s link speed no matter how close I am to the router. At distances over 4 metres the speed is below 802.11g standards. Latest driver is no improvement. Router also runs latest f/w.
I then tried a Netgear WNDA3100 dual chanel Draft-N USB that is supposed to be quite identical to the DWA-160. It is a more smooth experience to install and has better reception but the speeds achievable on any of those USB adapters isn't more than 30 Mbit/s or so when I tested this. What is the attainable speed one can expect from these USB devices if everything is configured optimally ?
Any hints on what settings to look at ?

Prior to getting all this draft-N gear I depended on my beloved Linksys WRT54GL with stock f/w. It delivers data fluidly at some 23 Mbit/s both up/downlink when connected to this WAN. And that is 802.11g ! What should I expect from the N-revolution ? At least marketing buzz about 600 Mbit/s total data flow over Wifi isn't showing up as real in my home. With current results I could just as well stay with the good ol WRT54GL which delivers data over a superior range and through walls only a tad slower. In fact it is a bit quicker on the uplink. There must be more to 802.11n just that I have some tweaking to do ?

A scan in this apartment complex lists a dozen of Wifi routers in nearby places. I try to avoid using the same channels, but probably most do automatic channel selection,should I do that as well ? My tests so far have mostly been 2.4 GHz. I tried 5 GHz briefly but it seemed even more range sensitive and seemed to run slower on the dual band USB sticks than 2.4 GHz. I do not have other computers on same or different speeds connected to the router when I do the speed tests.

The above probably looks like too many unknown params for giving precise advice...any hints on how I best troubleshoot ?

I think If I could accomplish at least a 50 Mbit/s download speed over Wifi at a distance of 6 metres or so with the router in sight that would be enough to motivate why I invest in this. Else maybe I'm better off running an Ethernet cable to the sofa instead of trying to transport stuff wirelessly, twice the speed and very low cost, just a bit less cool

 

[thiggins]

Unfortunately, the DIR-855 is very expensive product with a long, troubled past. If you check the open-air tests in the review, you'll see that I got only around 50 Mbps best case, same room, 3M between router and client and no other networks in range.

Dual-band draft 11n routers just aren't very good right now and there are not many good clients.

Yes, the DWA-160 is a lousy adapter. The WNDA3100 is somewhat better. But sounds like you should stick with your internal adapter.

BTW, the speeds you see reported by Windows (130Mbps, 65 Mbps) are link rates, not actual throughput (comparable to 10/100/1000 Mbps of Ethernet).

To sum up, given your environment and equipment, your experience isn't unusual. If your notebook can take a CardBus card, you might try a D-Link DWA-652 for single band or a Linksys WPC600N for dual-band. But that may be throwing good money after bad.

What did you want more wireless speed for anyway?

 

[gunnar-eee]

Hello I was able to sort things out now. The DWA-160/DIR-855 combo is great on the 5GHz band as long as I'm within a 5 metre range.
The sensitivity of the DWA-160 is very poor still, singnal strength falling quickly as range increases. With direct line in sight of router it seems up to some 6 metres range is ok to get good speed. And I saw speeds temporarily exceeding 100 Mbit/s during the download test, it seems this is bursts that exceed the 100 Mbit/s WAN speed briefly, else it was around 80-90 Mbit/s with the router in sight.

My setup wasn't properly done. I found there are two settings that are vital to getting decent speeds;
(1) Encryption *can* be used without major performance drops as long as
WPA2 with AES is chosen as shown in the review on this site:
http://www.smallnetbuilder.com/content/view/30475/96/1/3/

(2) Channel Width : Auto 20/40 MHz

It seems on my router 20 MHz was a default setting there thus implicitly preventing the router from accomplishing the full potential. But I cannot swear I remember it right. However, I found some discrepancy between two User Manual versions;
The manual available on the U.S site is:
ftp://ftp.dlink.com/Gateway/dir855/Manual/dir855_manual_110.zip

and in that manual it is stated that Auto 20/40 is the default.

But, on the Australian D-Link site I found this User Manual version:

ftp://files.dlink.com.au/products/DIR-855/Manuals/DIR-855_A2_Manual_1.20.pdf

where it is stated that 20 MHz is the default Channel Width setting.

Which of the manuals is the most recent ? Which of them reflect the true state of things on A2 version routers ?

Here are my 5GHz settings:

================
Wireless Band : 5GHz Band

802.11 Mode : 802.11n only

Enable Auto Channel Scan : Yes (checked)

Transmission Rate : Best (automatic)

Channel Width : Auto 20/40 MHz

Security Mode : WPA-Personal

WPA Mode: WPA2

Cipher Type: AES



Sorry if it becomes off-topic; I can try the same router config against the Netgear WNDA3100 and report tomorrow.

 

[thiggins]

For Wi-Fi Certification, draft 11n routers must default to 20MHz bandwidth mode in 2.4GHz to prevent interference with 802.11b/g networks.

Yes, you must use no encryption or WPA2/AES or you will get significant throughput reduction. This is true with all draft 11n routers based on Atheros and Broadcom chipsets. Routers using Ralink chipsets maintain good throughput using WEP or WPA-TKIP.

 

[Q-BZ]

For Wi-Fi Certification, draft 11n routers must default to 20MHz bandwidth mode in 2.4GHz to prevent interference with 802.11b/g networks.

Yes, you must use no encryption or WPA2/AES or you will get significant throughput reduction. This is true with all draft 11n routers based on Atheros and Broadcom chipsets. Routers using Ralink chipsets maintain good throughput using WEP or WPA-TKIP.

I'm in a simillar boat with the DGL 4500 and the DWA 160 two floors away. Latest firmwares and hardware revisions across the boards on both products.

I'm still fiddling with the best settings for this and having fun experimenting.

Right now, I'm running 2.4GHZ N only and the DWA 160 across the house and two floors away from the DGL 4500 is performing like this:

802.11n (2.4GHz) Rate (Mbps):108 Signal: 58 percent

I've tried 5GHz and I liked what I saw when it was on but I was getting a signal of maybe 19-20 percent. The Mbps was somewhat less as well. I suspect that's the reduced range rearing its ugly head at me.

What was interesting about that, though, is I'd have occasional drop outs but when it was "on" even at that weak signal, it was very fast and crisp.

I used speedtest.net and the upload and download speeds were excellent across the boards, much better than my five year old g setup...on both 2.4 and 5 but my pings are ghastly. Well into high 100's or so. Is that "thoroughput?"


My current settings in the 4500 are:

Enable Wireless : Yes

802.11 Band : 2.4GHz

802.11 Mode :N Only

Enable Auto Channel Scan : Yes

Wireless Channel : (whatever it chooses in range per auto setting)

Transmission Rate : (Mbit/s) Best Automatic

Channel Width : 40 MHz

Security:

WPA Personal

WPA mdoe: Auto: (WPA or WPA2)

Ciper: TKP and AES

Group Key Update Interval: 3600

Any settings there I should consider changing? I wish I could move that second computer up a floor and/or closer but I don't think we'll be able to.

Even now, there's occasional drop outs happening on the DWA 160 end of things...

 

[tipstir]

You want the signal at 90% to 100% you might need to install WAPN where it's weaker to improve your signal. Link speed would increase. Are you seeing drops in the signal or how many bars you have in Vista or XP OS?

 

[Q-BZ]

You want the signal at 90% to 100% you might need to install WAPN where it's weaker to improve your signal.

What's this?

Link speed would increase. Are you seeing drops in the signal or how many bars you have in Vista or XP OS?

Both, but it's better since I went to 2.4GHZ 40 MHz. I'm seeing about 2/4 bars or so.

The best signal I'm seeing is in the low 60 percentile range.

 

[Q-BZ]

I'm halfway wondering if I should pack up the DWA 160 and send it back to Amazon in exchange for something with some more kick and some antennas.

I wonder if the DWA 552, 556 would be better, especially with the antenna extensions, or if I could even use something like the DAP 1522 instead in access point mode on that second PC, if that's even how something like that can work. That doesn't have any antennas on it, though. Doesn't seem like that would the approriate application.

Hopefully Tim will look over my post at the bottom of the previous page and read through this and see if I could configure myself any better than I have. As best as I can tell I haven't missed anything major.

I love the DGL4500. I'm definitely keeping that. I may just need some kind of signal booster, repeater, extender or some such that I could tuck away on the first floor in between the DGL 4500 PC and the DWA 160 PC two floors down and away.


I could always try to beef up the antennas on the DGL4500...maybe something like three of these?: http://www.hyperlinktech.com/item.aspx?id=2324

 

[jdabbs]

I could always try to beef up the antennas on the DGL4500...maybe something like three of these?: http://www.hyperlinktech.com/item.aspx?id=2324

The increase in gain comes with an alteration of the radiation pattern. In the case of rubber duck omnis, you lose vertical coverage. This isn't a problem if your clients are all on the same axis, but it seems counterintuitive at first that a "better" antenna can result in worse range/performance in the situation you described. Parts 3 and 4 of the "How To Fix Your Wireless Network" series are a good read.

There are a lot of potential problems with your setup: interference from other devices on the 2.4 GHz band, antennas not aligned optimally, inadequate client performance, etc. Since doing a site survey followed by equipment repositioning is cheapest, I suggest starting there. Eliminating just one wall from the transit path can double the received signal.

 

[thiggins]

The DWA-160 is not a very good adapter, at least not the Rev A1. I had better luck with a NETGEAR WNDA3100, or if you can use a CardBus card, the Linksys WPC600N.

Trying different placements of router and client and higher gain antennas are always better than using repeating solutions. Each repeater you go through will reduce throughput by at least 50%. As jdabbs noted, eliminating even one wall (or floor) can greatly improve performance.

If you are going to try a different antenna, you need a gain of at least 6dB to see any improvement. Standard dipoles that routers come with have a gain of around 2 dBi. So you need an antenna with at least 8 dBi of gain, which the antenna that you are looking at doesn't have.

 

[Q-BZ]

The increase in gain comes with an alteration of the radiation pattern. In the case of rubber duck omnis, you lose vertical coverage. This isn't a problem if your clients are all on the same axis, but it seems counterintuitive at first that a "better" antenna can result in worse range/performance in the situation you described. Parts 3 and 4 of the "How To Fix Your Wireless Network" series are a good read.

Fantastic stuff. I'm digging into it...

There are a lot of potential problems with your setup: interference from other devices on the 2.4 GHz band,

Ideally, I would like to do 5GHz 40MHz N all the way but the range simply is not there.

... antennas not aligned optimally,

Right now I have the router in its stand with the antennas facing straight out towards the opposite side of the house where the DWA 160 is...two floors down.

Is that a mistake? Would I be better off having the 4500's antennas straight up in the air instead?

... inadequate client performance, etc. Since doing a site survey followed by equipment repositioning is cheapest, I suggest starting there. Eliminating just one wall from the transit path can double the received signal.

I'll definitely get into this...

The DWA-160 is not a very good adapter, at least not the Rev A1. I had better luck with a NETGEAR WNDA3100, or if you can use a CardBus card, the Linksys WPC600N.

I'll look into it although this is revision B with the latest firmware.

The biggest problem we have is an occasional drop off down there and I almost wonder if it's some automatic function happening.

The actual connection is strong when in use and we're getting speeds very close to the optimal that our ISP provides. Overall, this is a CLEAR improvement over the 5 year old G setup we had before.

If I could simply find out what's causing those drop outs and fix it, I'd basically be happy enough.

Best guess: If the router is set to auto tune channels vs. me manually setting and assigning a channel...would that setting and auto tuning cause drop outs?

Trying different placements of router and client and higher gain antennas are always better than using repeating solutions. Each repeater you go through will reduce throughput by at least 50%. As jdabbs noted, eliminating even one wall (or floor) can greatly improve performance.

Probably don't need a repeater in this instance. I'll keep it as simple as I can.

If you are going to try a different antenna, you need a gain of at least 6dB to see any improvement. Standard dipoles that routers come with have a gain of around 2 dBi. So you need an antenna with at least 8 dBi of gain, which the antenna that you are looking at doesn't have.

Agreed. Any recommendations?

 

[thiggins]

All I can tell you is to try different antenna orientations. Just keep in mind the "doughnut" shape of the dipole radiation pattern.

Your occasional dropouts are probably a ranging algorithm periodically kicking in. You might try setting a channel (although I think the automatic channel selection occurs only during power-up), but there is probably little you can do. You might check the adapter Advanced properties and play with settings there.

One other thing you can try if your setup lets you do it is to directly browse to the file you are trying to play vs. "streaming it". See this article.

 

[Q-BZ]

All I can tell you is to try different antenna orientations. Just keep in mind the "doughnut" shape of the dipole radiation pattern.

Your occasional dropouts are probably a ranging algorithm periodically kicking in.

It seems to happen once per hour or so. It's lightly suspicious.

You might try setting a channel (although I think the automatic channel selection occurs only during power-up), but there is probably little you can do. You might check the adapter Advanced properties and play with settings there.

One other thing you can try if your setup lets you do it is to directly browse to the file you are trying to play vs. "streaming it". See this article.

Good deal.

I guess in post 30 that I wrote in this thread you don't see any settings there that I posted that stand out to you as being "wrong"?

 

[thiggins]

I guess in post 30 that I wrote in this thread you don't see any settings there that I posted that stand out to you as being "wrong"?

I would not force to 40MHz bandwidth mode. That will actually make performance worse under low signal conditions. I also would not force to N only mode. And if you have security enabled, you should use WPA2/AES.

 

[Q-BZ]

I would not force to 40MHz bandwidth mode. That will actually make performance worse under low signal conditions.

Go to auto 20/40? EDIT: Set it to 20/40.

I also would not force to N only mode.

I figured since this was an all N network and nothing else that was the way to go. What should I change it to?

EDIT: I've kept it on N only for now since this network is nothing but N.

And if you have security enabled, you should use WPA2/AES.

Current security mode is WPA Personal.

The WPA mode is currently at automatic for "WPA or WPA2" You want me to switch it to "WPA2 only"?

EDIT: Set to WPA2 only.




Past that, the cipher is set "TKIP and AES"...set this to "AES" only?

EDIT: Set to AES only.

I appreciate the help and the pointers. I've shilled the heck out of this place the last few months because I've really learned a lot and this is another case in point here.


EDIT: I did all of your recommendations and kept it at N only 2.4GHz and the basement is basically getting near hard wire like speeds up and down and went from ghastly 150 some odd pings down to 44! Wow!