MIMO, Antennas, and DIR-825

[freddyspam]

Long time reader. First time poster. After searching the internet, I'm still confused, and I thought someone might be able to help me.

If I have a MIMO client and one or more of the antennas are disconnected/not present, does the device automatically change modes to function using the remaining antennas or does it not work at all? Does this occur on the hardware or software level? Is there a set standard for this or is it device specific?

Same question but now at the router end.

I am getting the DIR-825. You can guess where I am going with this. In the review for the 825:

You might notice that D-Link has used the 855's case, with an empty middle hole for a third antenna. The photo also shows that the mini-PCI radios even have all three miniature connectors loaded. But if you want to try to turn an 825 into an 855, forget it. All the components required to combine the 2.4 and 5 GHz radio outputs aren't loaded. And you'd have to scare up an additional dual-band antenna, too.

Doesn't this mean that if you work purely in the 5GHz band, then adding a corresponding 5GHz antenna is all you need? or is the device not using the antenna on a firmware level? Is router trying to use the third antenna and just defaulting to two or does it need to be programed into the firmware?

Lastly, even if it all worked out, would I notice a difference?

I know that is a lot of questions. And yes I know I could just buy [Fill in the blank] router instead. But I'm more curious as to how this technology works. Thank you in advance.

 

[jdabbs]

There are at least two versions of the DIR-825. The B1 revision uses actual 2T2R radios, so you wouldn't be able to upgrade them.

If you do have an A1 model, you can check for the presence of U.FL connectors on the radio cards. FCC photos (which SNB often uses) are of pre-production equipment, and manufacturers make minor component changes, so you'll need to open the unit to be certain of what you have.

If you have an A1 with a 3rd U.FL connector, then you could add an antenna by connecting a pigtail directly to either the 2.4 or 5 GHz radio. The quote was probably referring to the lack of a combiner, which you'd need for both radios to share an antenna. For use of one band, this obviously isn't a problem.

Could the 825 be crippled in software as well? It's possible. You could find out by flashing a 855 with a sufficiently tweaked DIR-825 firmware, but for only $20 in hardware (if U.FL connectors are present) it'd be easier for someone to do the mod and report the results.

Whether or not you'd see a benefit depends on the client. If it works, I wouldn't expect wireless performance much different than the DIR-855 A2, since the radios are the same.

 

[freddyspam]

Thank you jdabbs. Hopefully this won't get to the point of modifying firmware.

I was hopping to figure out how this works without doing a "mod and see" approach. Hopefully it will be a B1 model (I think you got the models switched or maybe the review is misleading) but for $25 I can't complain.

One problem I have with just modding it is how do I even tell it is using the third antenna? My client has three antennas, but how do you tell what mode it is in?

I was hoping that antenna management was controlled on the hardware level by default. In which case it would be safe to assume this mod would work. I'm surprised more people aren't talking about this. This should at least be easier than the time I swapped the ram in my router.

If anyone else has any insight, I would appreciate it. Otherwise, I have the antenna, and I'll just wait for the router to show up. I'll report what happens.

 

[thiggins]

All current draft 11n products are at most dual-stream. MIMO "streams" are different than the number of antennas and/or transmit / receive chains used in a product.
Basically, the # of streams determines the link rates that can be used. With two streams, the maximum link rate is 270 or 300 Mbps, which is very indirectly related to the actual maximum throughput you will see.

Bottom line is that worrying about the # of antennas used in a draft 11n product is just spinning your wheels. There are far too many factors that determine draft 11n performance, with antennas only one of them.

 

[jdabbs]

I was hopping to figure out how this works without doing a "mod and see" approach. Hopefully it will be a B1 model (I think you got the models switched or maybe the review is misleading) but for $25 I can't complain.

Both the 855 and 825 have multiple versions, so things can get complicated.

DIR-855 A1: 3T3R Atheros 2133/5133
DIR-855 A2: 3T3R Atheros 9103/9106
DIR-825 A1: 3T3R Atheros 9103/9106
DIR-825 B1: 2T2R Atheros 9223/9220

Getting an unmoddable B1 wouldn't be so bad; it uses the same radios as the WRT400N, which performed very well in benchmarks.

All current draft 11n products are at most dual-stream. MIMO "streams" are different than the number of antennas and/or transmit / receive chains used in a product.

You may be interested in knowing that the CCNA Wireless exam guide explicitly states that number of emitting antennas=number of streams. They then followed up with a reference to special multiplexing.

 

[freddyspam]

Thanks jdabbs for clearing that up.

Getting an unmoddable B1 wouldn't be so bad; it uses the same radios as the WRT400N, which performed very well in benchmarks.

I noticed that too, but where is the fun in that? Getting a device that works great out of the box. I probably could have kept my 802.11g network, but I had reached the limit as far as how much I could mess with it. My home network is a jumble of wireless bridges, wireless cameras, and multipurpose servers.

Honestly, I'm just trying to figure out how mimo is implemented in general. If I get a noticeable increase in stability/connectivity, that will be a plus.

I realized an easy way to see if the third U.FL has been disabled if I do get the A1. I just move one of the existing antennas from an occupied U.FL to the third free one (and then disconnect the other antenna?). Then see if I can connect to the router. I'm still not sure if these devices can function on just one antenna. Some of my other routers dedicate one to transmit and one to receive but that is with custom firmware.

Is there any way to know whether you are utilizing 2T2R or 3T3R or 2T3R or 3T2R or R2D2...? Normally, I can see how this would be a stupid question as you would know the capabilities of your client and router.

At the very least, if I'm not using the dual band or the dual SSID's, it would be fun to remove one of the network cards, put it in a laptop, and turn that into an draft n wireless bridge or repeater.

I hope I find a job soon.

 

[thiggins]

You may be interested in knowing that the CCNA Wireless exam guide explicitly states that number of emitting antennas=number of streams. They then followed up with a reference to special multiplexing.

Since there are no shipping three-stream products, that means that the third antenna, if present, is used for receive-only.

Was that "special" multiplexing or spatial multiplexing?

 

[jdabbs]

I realized an easy way to see if the third U.FL has been disabled if I do get the A1. I just move one of the existing antennas from an occupied U.FL to the third free one (and then disconnect the other antenna?).

Is there any way to know whether you are utilizing 2T2R or 3T3R or 2T3R or 3T2R or R2D2...? Normally, I can see how this would be a stupid question as you would know the capabilities of your client and router.

At the very least, if I'm not using the dual band or the dual SSID's, it would be fun to remove one of the network cards, put it in a laptop, and turn that into an draft n wireless bridge or repeater.

Migrating the pigtail would test whether a particular connector was disabled, but wouldn't determine that the radio is limited to using two antennas.

Stream usage is directly related to link rate, and determined by the least capable device.

For some time now, laptops come with mini PCIe slots only, while almost all AP/router radio cards are mini-PCI.

Since there are no shipping three-stream products, that means that the third antenna, if present, is used for receive-only.

Was that "special" multiplexing or spatial multiplexing?

The word "special" was used, which is why I italicized it. They did get it correct on the previous page though.

How does Atheros get away with calling a two-stream product 3T3R? Is it implemented by choosing the two best quality transmitters to use, or is it creating a two-stream link from three antenna outputs?

 

[thiggins]

How does Atheros get away with calling a two-stream product 3T3R? Is it implemented by choosing the two best quality transmitters to use, or is it creating a two-stream link from three antenna outputs?

As you seem to know, the # of transmit and receive channels isn't directly related to the number of streams. It's very confusing and not easily explained unless you get into the details of the draft 11n MCS.

I had a long discussion with Atheros when they introduced their Align single-stream product about this very topic. This discussion prompted Atheros' Align: Does Single Stream 802.11n Really Help?.

There is no simple answer that I can see. Note also that single-stream product announcements were not made at CES in January.

 

[jdabbs]

I suspect that if there is a succinct explanation, it hinges upon the understanding of other concepts. I'll do some research and if I feel I have a decent grasp of the material, I'll attempt an explanation for you to validate.

 

[freddyspam]

Update:

The router came today and it is an A1 model. I tested it and everything works on it by default b, g, n ( 2.5 & 5.0). I then removed one of the two antennas from both of the cards, and I could still connect to the router. I didn't try any throughput tests. I then removed the second antenna from the wireless cards (meaning now nothing is connected to them) and surprise, no networks were detected. I then connected one of the antennas to the third previously unused U.FL connectors on both cards leaving the second antenna still unattached and.......nothing. It looks like the third U.FL's are disabled at the firmware level.

This was interesting though. Unless someone sees a flaw in my logic/approach, looks like this router is left at 2t2r. It still preforms damn well even in a noisy environment.

 

[toolshed]

Other Mods Possible?

I was reading your posts on the DIR-825
How did you open the case to the device?

Do you think separating the 2 and 5 ghz frequencies
would offer any better performance by
using 4 antennas instead of combining
the 2 as is currently.

Basically I want to open to the device drill some
holes and add two antennas and separate the two
frequency bands.

Thanks

 

[freddyspam]

I was reading your posts on the DIR-825
How did you open the case to the device?

Do you think separating the 2 and 5 ghz frequencies
would offer any better performance by
using 4 antennas instead of combining
the 2 as is currently.

Basically I want to open to the device drill some
holes and add two antennas and separate the two
frequency bands.

Thanks

To open the unit, remove the rubber pads on the bottom to reveal two screws. Once the screws are removed, the unit will pop apart with a little force.

As for separating the two bands, I don't see any benefit. They inherently don't interfere with one another. The only time I would imagine an issue could arise is if both bands are being simultaneously used heavily, and even in that case, your bottle neck will be the processing and routing power of the unit.

But if you decide to do it, please let me know how it goes.

I've come to the conclusion that dual band is kind of pointless. You are always better off operating in one band or the other, and if you need 802.11 b and g capabilities, just get an inexpensive second router to handle that traffic while your draft n takes the 5.0 band. Otherwise, operating in mixed modes, you are crippling your draft n throughput even before you factor in interference. Also, you hardly ever need to jump between the two bands while using draft n. Obviously, I'm speaking in broad terms. Your mileage might vary.

 

[toolshed]

Re: Other Mods Possible?

freddyspam

Thanks for the response.

I pulled off the wrong rubber foot I guess. I was expecting 4 screws
so when I pulled one off and saw no screw I was a little confused about
how the device was held together. I had a 50/50 chance oh well!

I was thinking about separating the bands because I seem to get better signal levels when only band is switched on but maybe I am wrong. I'll check again. The only easy mod here I think would be to just go around the combiners and separate the antennas.

Thanks again

 

[toolshed]

Follow Up

When I turn off the 2.4 Ghz channel, signal power on the 5Ghz goes from 16%
to 30% on the wireless section of the status screen.

Does anybody else see the same thing?

If this is real is this because of the combiner
or firmware power limitations when both frequencies are on?

I guess I will have to open it up and find out unless someone
has done the work for me.

 

[stevech]

When I turn off the 2.4 Ghz channel, signal power on the 5Ghz goes from 16%
to 30% on the wireless section of the status screen.

Does anybody else see the same thing?

If this is real is this because of the combiner
or firmware power limitations when both frequencies are on?

I guess I will have to open it up and find out unless someone
has done the work for me.

I suggest doing dual band by using an 11g or 11n consumer WiFi router, then add an 11a or 11a/n router configured to be an access point.

It may be RF switching related to a combiner. But a small change like that isn't usually significant. Wish it stated dBm rather than % since a % is a dimensionless number. A 3dB change is double/half power, but that 3dB is small compared to the typical WiFi path loss for common distances/walls - where this is often 80dB or more.

RF systems use a log scale because of the inverse square law in RF propagation path loss. So a value in % is more of a marketing thing. The "bars" on a cell phone more or less depict a log scale for a good reason.

 

[jalyst]

Anyone know if this ever ended-up with excellent 3rd party fw support, or is the wndr3700 also better in that regard?
https://dev.openwrt.org/wiki/WantedHardware